UNDERSTANDING THE SUICIDAL MIND:

AN ECOLOGICAL INVESTIGATION OF THE

DIFFERENTIAL ACTIVATION HYPOTHESIS
OF SUICIDAL RELAPSE IN FIRST EPISODE PSYCHOSIS

by

DONNA BELLA BACK

A thesis submitted to the University of Birmingham for the degree of

DOCTOR OF PHILOSOPHY

School of Psychology
University of Birmingham

February 2013
 University of Birmingham Research Archive
e-theses repository

This unpublished thesis/dissertation is copyright of the author and/or third

parties. The intellectual property rights of the author or third parties in respect
of this work are as defined by The Copyright Designs and Patents Act 1988 or
as modified by any successor legislation.

Any use made of information contained in this thesis/dissertation must be in

accordance with that legislation and must be properly acknowledged. Further
distribution or reproduction in any format is prohibited without the permission
of the copyright holder.
 University of Birmingham Research Archive
e-theses repository

This unpublished thesis/dissertation is copyright of the author and/or third parties. The
intellectual property rights of the author or third parties in respect of this work are as defined
by The Copyright Designs and Patents Act 1988 or as modified by any successor legislation.
Any use made of information contained in this thesis/dissertation must be in accordance with
that legislation and must be properly acknowledged. Further distribution or reproduction in
any format is prohibited without the permission of the copyright holder.
 Abstract
The incidence of completed and attempted suicide among those with first episode

psychosis (FEP) is high. Studies have shown that history of self-harming behaviour remains

as the strongest predictor of both attempted and completed suicide in psychosis. Due to the

lack of understanding about the suicidal thinking mechanism in psychosis, the development of

effective treatment interventions continues to be a major gap for clinicians and patients. More

importantly, the rate of suicidal relapse remains tragically frequent. In view of the fact that

hopelessness is one of the most significant cognitive risk factors for suicidal behaviour in

psychosis, the Differential Activation Hypothesis (DAH) of suicidal relapse may serve as a

potential framework for understanding suicidality in psychosis. According to the DAH model,

low mood triggers the recurrence of hopeless/suicidal thoughts in individuals who have

previously felt suicidal during their early episodes of depression. This thesis sets out to

investigate the suicidal thinking process in FEP, by comparing those with a history of suicidal

attempt or deliberate self-harm vs. those without using the DAH of suicidal relapse as the

main theoretical framework.

First, the Experience Sampling Method (ESM) was employed in order to examine the

relationship between hopelessness and mood in the day to day life of people with psychosis.

The ESM is a diary keeping procedure that systematically samples real-life data over a 6-day

period. The ESM data showed that attenuated hopelessness was found to be more

differentially active in response to negative affectivity in the suicidal history group (N = 35)

than the non-suicidal group (N = 40).

Second, borrowing from the DAH methodology, the sad mood induction procedure

(MIP) was employed. The purpose of the sad MIP was to induce feelings of sadness necessary

to create a context that was suitable for reactivating hopeless thoughts. In conjunction with the
sad MIP, the Means-Ends Problem Solving (MEPS) task was employed in order to measure

the individual’s problem solving skills. It has been previously established that lack of problem

solving skills is an important characteristic of hopelessness. In order to test if the dampening

of mood will impair the individual’s problem solving ability, the MEPS task was performed

before and after the sad MIP. The results were as predicted by the DAH: the reduction in

problem solving ability following the mood challenge was significantly greater in the suicidal

history group (N = 48) than the non-suicidal group (N = 49).

The Future Thinking (FT) task was also employed conjunction with the sad MIP.

Similar to the MEPS task, the purpose of the mood challenge was to test if fluency for

considering positive events, another important characteristic of hopelessness, will also

respond to the changes in mood. The results indicated that the observed reduction in fluency

for positive events following the mood challenge was more evident in the suicidal history

group (N = 49) than the non-suicidal group (N = 50).

Together, these studies support the validity of the DAH of suicidal relapse as a

framework for understanding the suicidal thinking mechanism in psychosis. More

importantly, the consistent pattern of results shared between the ecological (ESM) and

experimental (sad MIP) studies validates the application of the DAH in the real-life, day to

day experiences of those at risk of suicidal relapse. The evidence suggesting the applicability

of the DAH in FEP will help establish the importance of the interaction between distal and

proximal risk factors for suicidality, which will be of great clinical value in improving the

existing risk assessment procedures.

 Dedication

To my father

Marino Rodil (1947 – 2011)

“In your last breath, you uttered my name,

Broken and lost, my coldest winter came.
In my dreams, your face I always see,
You are dearly missed today,
tomorrow, and forever will be.”

(Donna ‘13)
 Acknowledgement
I am profoundly and sincerely grateful to a number of people who have helped me in various ways in
order to enable to me to complete this research. First, I would like to thank my supervisor Prof. Max Birchwood
who persistently supported me throughout my PhD with his wisdom, and patience. Thank you for believing in me
- your enthusiasm and guidance made the completion of this thesis possible. I also would like to thank Dr Chris
Jackson who helped me keep my sanity when things were falling apart! Thank you for your advice, insight, and for
meticulously checking my drafts. Thanks to Prof. Inez Myin-Germeys for your expertise in ESM, and for being
such an excellent host during my visits in Maastricht. Thanks also to Dr. Tineke Lataster & Margreet Oorschot for
your assistance during my stay in Maastricht and for making Stata a lot less complicated! I also want to thank
Prof Andy MacLeod and Prof Matthias Schwannauer for their constructive comments and corrections to the
initial version of this thesis.

Second, I would like to thank my family for your constant support and confidence in me. Despite the
distance, your support means a lot to me. I also would like to thank Philip for your kindness and thoughtfulness.
Thank you for being always there for me, especially for cooking so I never had to go to that place with the golden
arches everyday! Thanks also to my cats (Lily & Maisy) who kept me company when I was writing this thesis and
for getting me out of bed in the morning. I also wanted to thank my friends, especially Vagelis, Romony, Jason, Si,
Giggs, & Aimee whose support and banters made my life as a PhD student less stressful. Thanks to Matt - a true
friend through good and bad times. You and my cats (Frank & Molly) have been my family for years, thank you
for standing by me.

Third, I would like to thank Cathy, Asha, Nita, Priti, & Brett for their assistance in the data coding and
data entry. Thanks for your hard work and commitment! I am also grateful to the clinicians, managers, and staff
of the Early Intervention Service, especially Afshan, Marva, Yvette, Nicky, Caroline, Jai, Becks, John, Pritty, Ruth,
and Joel who made my idle moments in between recruitment fun. Thank you all for helping me out massively in
my recruitment! Linda, thanks for being a supportive manager. I especially would like to thank the Birmingham
and Solihull Mental Health Trust and the University of Birmingham who supported me financially throughout
my PhD. And most importantly, I would like to thank the Birmingham Early Intervention service users who took
part in my studies. Especially those who took part in all the three studies, thank you so much for making this
thesis possible!
 Table of Contents

Chapter 1 Suicidality in Psychosis

1.0. Introduction .................................................................................................................... 1
1.1. Diagnosis and Prevalence of Psychosis ......................................................................... 2
1.2. Phase of Psychosis .......................................................................................................... 3
1.3. Prevalence of Suicide in First Episode Psychosis ......................................................... 5
1.4. Risk Factors of Suicide in FEP ....................................................................................... 7
1.4.1. Demographic Risk Factors .................................................................................. 7
1.4.2. Clinical and Psychosocial Risk factors ................................................................ 8
1.4.3. Behavioural Risk Factors..................................................................................... 9
1.5. Hopelessness as a Risk Factor in FEP ............................................................................ 9
1.5.1. Hopelessness: Studies that Link Suicidality in FEP and Hopelessness............. 10
1.5.2. Hopelessness: The Need for a Theoretical Framework ..................................... 12
1.6. The Differential Activation Hypothesis of Suicidal Relapse ....................................... 14
1.6.2. Generalised Hopelessness vs. Cognitive Reactivity to Hopelessness ............... 15
1.6.3. Studies on Cognitive Reactivity to Hopelessness ............................................. 15
Chapter 2 The Experience Sampling Method
2.0. Introduction .................................................................................................................. 22
2.1. The ESM: An Overview .............................................................................................. 22
2.2. The Use of ESM in Psychosis Research ....................................................................... 23
2.2.1. The Contemporary ESM in Psychosis Research ............................................... 24
2.2.2. Feasibility and Compliance ............................................................................... 25
2.2.3. Limitations and Strengths .................................................................................. 26
2.2.4. Validity and Reliability ..................................................................................... 27
2.3. The ESM Studies in Psychosis ..................................................................................... 28
2.4. The ESM: A potential Tool to Test the Differential Activation Hypothesis ................ 35
2.5. Overview of Thesis ....................................................................................................... 36
2.6. Note on Collaboration .................................................................................................. 37
Chapter 3 The Mechanims of Hopelessness Linked to the Mood
Fluctuations in Everyday Life: An ESM Study
3.0. Introduction .................................................................................................................. 38
3.1. Hypotheses ................................................................................................................... 43
3.2. Method .......................................................................................................................... 44
3.2.1. Sampling ............................................................................................................ 44
3.2.2. Measures ............................................................................................................ 45
3.2.3. Procedure ........................................................................................................... 51
3.2.3.a. Case Identification ................................................................................ 51
3.2.3.b. Pilot Feasibility Study of the ESM ....................................................... 53
3.2.3.c. The Experience Sampling Method ....................................................... 54
3.3.4. Analysis Strategy ............................................................................................... 55
3.3. Results .......................................................................................................................... 56
3.3.1. Factor Analyses on the ESM questionnaire items ............................................. 56
3.3.2. Sample Characteristics ...................................................................................... 61
3.3.3. Descriptive Statistics and T-tests ....................................................................... 61
3.3.4. Hypotheses Testing ........................................................................................... 61
3.4. Discussion..................................................................................................................... 78
3.4.1. Strengths and Limitations .................................................................................. 86
Chapter 4 Assessing the Link between Low Mood and Lack of Problem
Solving Skills as a Behavioural Feature of Hopelessness:
A Mood Priming Study
4.0. Introduction .................................................................................................................. 89
4.1. Hypotheses ................................................................................................................... 94
4.2. Methods ........................................................................................................................ 95
4.2.1. Sampling ............................................................................................................ 95
4.2.2. Measures ............................................................................................................ 95
4.2.3. Procedure ........................................................................................................... 96
4.2.3.a. Case Identification ................................................................................ 96
4.2.3.b. Sad Mood Induction Procedure ............................................................ 98
4.2.3.c. Means-End Problem Solving Task ....................................................... 99
4.2.4. Analysis Strategy ............................................................................................. 100
4.3. Results ........................................................................................................................ 101
 4.3.1. Sample Characteristics .................................................................................... 101
4.3.2. Descriptive Statistics and T-tests ..................................................................... 102
4.3.3. Hypotheses Testing ......................................................................................... 103
4.4. Discussion................................................................................................................... 113
4.4.1. Strengths and Limitations ................................................................................ 120
Chapter 5 Assessing the Link between Low Mood and Lack of Positive
Future Fluency as a Behavioural Feature of Hopelessness:
A Mood Priming Study
5.0. Introduction ................................................................................................................ 123
5.1. Hypotheses ................................................................................................................. 127
5.2. Methods ...................................................................................................................... 128
5.2.1. Sampling .......................................................................................................... 128
5.2.2. Measures .......................................................................................................... 129
5.2.3. Procedure ......................................................................................................... 130
5.2.3.a. Case Identification .............................................................................. 130
5.2.3.b. Sad Mood Induction Procedure .......................................................... 131
5.2.3.c. Future Thinking Task.......................................................................... 131
5.2.4. Analysis Strategy ............................................................................................. 134
5.3. Results ........................................................................................................................ 134
5.3.1. Sample Characteristics .................................................................................... 134
5.3.2. Descriptive Statistics and T-tests ..................................................................... 135
5.3.3. Hypotheses Testing ......................................................................................... 136
5.4. Discussion................................................................................................................... 154
5.4.1. Strengths and Limitations ................................................................................ 162
Chapter 6 General Discussion
6.0. Introduction ................................................................................................................ 165
6.1. Summary of Findings ................................................................................................. 165
6.2. Limitations .................................................................................................................. 169
6.3. Observations From Research: Recommendations for Future Studies on Suicidality
in Psychosis ................................................................................................................ 170
6.4. Clinical Implications .................................................................................................. 172
6.5. Conclusion .................................................................................................................. 175
References ......................................................................................................................... 176
Appendices ........................................................................................................................ 204
Appendix 1. Participant Information Sheet ....................................................................... 205
Appendix 2. Letter of Invitation to the Participants .......................................................... 210
Appendix 3. Letter to the Participant's GP ........................................................................ 211
Appendix 4. Participant Consent Form (Mood Priming Study) ........................................ 212
Appendix 5: Participant Consent Form (ESM Study) ....................................................... 213
Appendix 6. Columbia Suicide History Form ................................................................... 214
Appendix 7. Leiden Index of Depression Scale - Revised ................................................ 218
Appendix 8. Calgary Depression Scale for Schizophrenia ............................................... 220
Appendix 9. Beck Hopelessness Scale .............................................................................. 223
Appendix 10. InterSePT Scale for Suicidal Thinking ....................................................... 224
Appendix 11. The ESM Time Sampling Schedule ........................................................... 226
Appendix 12. The ESM Debriefing Form ......................................................................... 227
Appendix 13. Visual Analogue Scale ................................................................................ 229
Appendix 14. The MEPS Problem Scenarios .................................................................... 230
Appendix 15. Velten's Negative Statements ..................................................................... 231

Total word count = 48,690

 List of Tables
Table 1. Summary of the ESM Studies in Psychosis .......................................................... 30
Table 2. A Factor Analysis of the ESM Affective Items .................................................... 58
Table 3. A Factor Analysis of the ESM Hopelessness Items .............................................. 59
Table 4. A Factor Analysis of the ESM Activity Appraisal Items ...................................... 60
Table 5. Summary of Descriptive Statistics for the key ESM factors ................................. 60
Table 6. Descriptive and t-statistics for Age and Key Clinical Symptoms ......................... 62
Table 7. Summary of Multiple Regression Analyses on Affectivity and Suicidality
as Predictors of Momentary Hopelessness ............................................................ 66
Table 8. Summary of Multiple Regression Analyses on Daily Hassles and Suicidality
as Predictors of Momentary Hopelessness ............................................................ 70
Table 9. Summary of Multiple Regression Analyses on Daily Hassles and Suicidality
as Predictors of Changes in Affectivity ................................................................. 74
Table 10. Summary of Multiple Regression Analyses on LEIDS' Hopelessness Subscale
Scores and Negative Affectivity as predictors of Momentary Hopelessness ........ 76
Table 11. Summary of Multiple Regression Analyses on LEIDS' Hopelessness Subscale
Scores and Daily Hassles as predictors of Momentary Hopelessness ................... 76
Table 12. MEPS: Means and Standard Deviations of Age and Key Clinical Symptoms
For the Suicidal History Group and Non-Suicidal Group ................................. 101
Table 13. Means and Standard Deviations of the Pre- and Post- Mood Induction Number
of Problem Solving Solutions ............................................................................. 145
Table 14. Means and Standard Deviations of the Pre- and Post- Mood Induction
Effectiveness Ratings of Problem Solving Solutions ........................................ 107
Table 15. Means and Standard Deviations of the Pre- and Post- Mood Induction Happiness
and Despondence Ratings ................................................................................... 110
Table 16. Means and Standard Deviations of Age and Key Clinical Symptoms for the
For the Suicidal History Group and Non-Suicidal Group ................................. 135
Table 17. Means and Standard Deviations of the Pre- and Post- Mood Induction Number
of Positive and Negative Events ........................................................................ 139
Table 18. Means and Standard Deviations of the Pre- and Post- Mood Induction Valence
Ratings on Positive and Negative Events .......................................................... 143
Table 19. Means and Standard Deviations of the Pre- and Post- Mood Induction Likelihood
Ratings on Positive and Negative Events .......................................................... 147
Table 20. Means and Standard Deviations of the Pre- and Post- Mood Induction Happiness
and Despondence Ratings .................................................................................. 151
 List of Figures
Figure 1. The ESM Questionnaire on Affectivity, Momentary Hopelessness,
Hopelessness, Activity, and Event ....................................................................... 52
Figure 2. Average Pre- and Post-induction Number of Problem Solving Solutions for
the Suicidal History Group & Non-Suicidal Group ........................................... 106
Figure 3. Average Pre- and Post-induction Effectiveness Ratings of Problem Solving
Solutions for the Suicidal History Group & Non-Suicidal Group ..................... 108
Figure 4. Average VAS Despondence and Happiness Ratings on Pre- and Post-Induction
Tasks in the the Suicidal History Group and Non-Suicidal Group .................... 112
Figure 5. Average Pre- and Post-induction Number of Positive Events and Negative
Events for the Suicidal History Group & Non-Suicidal Group .......................... 141
Figure 6. Average Pre- and Post-induction Positive Valence Ratings of Good Events
and Negative Valence Rating of Bad Events for the Suicidal History Group
and Non-Suicidal Group ..................................................................................... 145
Figure 7. Average Pre- and Post-induction Likelihood Ratings of Good Events and
Bad Events for the Suicidal History Group and Non-Suicidal Group ................ 149
Figure 8. Average VAS Despondence and Happiness Ratings on the Pre- and Post-Induction
Tasks in the the Suicidal History Group & Non-Suicidal Group ....................... 153
 List of Common Abbreviations

ANOVA Analysis of Variance

ANCOVA Analysis of Covariance

BHS Beck Hopelessness Scale

CDSS Calgary Depression Scale

CR to Hopelessness Cognitive Reactivity to Hopelessness

CSHF Columbia Suicide History Form

DAH Differential Activation Hypothesis

DSH Deliberate Self-harm

ESM Experience Sampling Method

FEP First Episode Psychosis

FT Task Future Thinking Task

ISST InterSept Scale for Suicidal Thinking

LEIDS Leiden Index of Depression Scale

MEPS Means-Ends Problem Solving

MIP Mood Induction Procedure

NA Negative Affectivity

PA Positive Affectivity

VAS Visual Analogue Scale

 CHAPTER 1
SUICIDALITY IN PSYCHOSIS

1.0. Introduction

The main goal of this thesis is to investigate the suicidal thinking process of

individuals who recently suffered an initial episode of psychosis. Suicide is a major health

issue worldwide with significant economic implications. According to the World Health

Organisation (2012), the worldwide prevalence rate of suicide is about a million a year,

which is approximately one complete suicide every 32 seconds. In the UK alone, the Office

of National Statistics (2012) reported an incidence of 6,045 completed suicide in 2011, of

which 4,552 are men and 1,493 women. Contrary to the popular belief, not everyone who

attempts suicide is mentally ill. Whereas many previous studies have indicated a strong link

between suicidal behaviour and mental illness, a previous study suggests that only 1 out of 4

suicide attempters have been in contact with the mental health services a year prior to their

death (Pirkis & Burgess, 1998).

Understanding suicide and suicidal attempts remains difficult for clinicians and

researchers alike. Suicidal behaviour is a complex phenomenon to predict given the

enormous amount of potential risk factors, which can be intertwined in a number of ways.

The purpose of this chapter is to review the scientific literature on hopelessness as a

significant risk factor for suicidal behaviour in early psychosis. Preceding the review is a

brief discussion of the general aspects of psychosis, its definition, diagnosis, and associated

features. Following this is an overview of the prevalence and risk factors of suicidal

behaviour in psychosis, with a particular emphasis on the role of hopelessness as an

associated feature of suicidal vulnerability. The concluding discussion will point at the

1
application of the Differential Activation Hypothesis as a potential model for understanding

hopelessness and managing suicidal vulnerabilities in early psychosis.

1.1. The Diagnosis and Prevalence of Psychosis

According to the definition of the APA's Diagnostic and Statistical Manual of Mental

Disorders (DSM IV; 2000), psychosis is a symptom of a distortion in rational thinking that is

often characterised by the person’s inability to recognise reality from that of the imaginary.

Also commonly described in the literature as a “loss of contact with reality”, psychosis

typically manifests itself in the form of imaginary experiences (e.g. hallucinations) or

fictitious beliefs (e.g. delusions or paranoia). Other forms of psychosis also include

incoherent speech (e.g. word salad) and muddled thoughts (e.g. flight of ideas) along with a

lack of awareness of the psychotic experience (APA, 2000). Whereas psychosis occurs as a

symptom of other mental health conditions, the experience of psychosis alone does not

warrant a diagnosis of mental illness. The initial episode of psychosis is often referred to as

“early psychosis” or “first-episode psychosis” (Kirch, Lieberman, & Matthews, 1992). In

general, psychotic episodes can range from briefly losing touch with reality due to the effects

of sensory-altering drugs, to perpetually experiencing lapses from reality due to the presence

of a long-term and severe psychiatric condition. Both the length and the causal factor of the

psychotic experience will help determine the diagnosis of a psychotic illness. A diagnosis is

particularly difficult to make during the initial psychotic episode due to the lack of

information on the causal factors that triggered the symptom on the first place. In order to

formulate a diagnosis, a clinician carries out a mental health examination in the form of a

clinical interview. A diagnosis of Schizophrenia is usually given when a psychotic episode

lasts for 6 months or more. A diagnosis of Bipolar is typically given when the psychotic

symptoms are accompanied by cycles of polar opposite mood swings, from extreme highs

2
(mania) to lows (depression; APA, 2000). Other types of diagnostic labels for psychosis are:

drug-induced psychosis, organic psychosis (psychotic episode due to a particular physical

condition), brief reactive psychosis (a brief psychotic experience due to a traumatic life

event), psychotic depression (depression with psychotic features), schizophreniform disorder

(psychotic episode of less than 6 months), and schizoaffective disorder (psychotic symptoms

are neither that of schizophrenia nor a mood disorder; APA, 2000).

According to the National Institute of Clinical Excellence (2009), the prevalence rate

of psychotic illness in the UK across age is about 7 per 1000 of the population. Mangalore

and Knapp (2006) indicated that about 37 – 40% of the incidence of psychotic episodes in the

country satisfy the diagnostic criteria for Schizophrenia. In a survey conducted by the Office

of National Statistics (2000), schizophrenia alone has a yearly prevalence rate of 5 per 1000

of the population in the UK. Although schizophrenia affects men and women equally, the

onset of schizophrenia in men occurs at an earlier point in their lives (15 – 24 years of age)

compared to the onset in women (24 – 35 year of age; Hafner, Maurer, Loffler, & Riecher-

Rossler, 1993; Hafner, Riecher-Rossler, Maurer, et al., 1992). The incidence of schizophrenia

has also been reported to be particularly higher in the black and minority ethnic or BMI

groups (Bresnahan et al., 2007; Sharpley, Hutchinson, McKenzie, & Murray, 2001).

1.2. Phases of Psychosis

There are three stages to a psychotic episode (APA, 2000). The length of each stage,

however, varies greatly from person to person. Stage 1 is known as the prodromal phase and

is characterised by changes in the person’s behaviour and perception of things, along with

his/her feelings and thoughts. These changes may be too subtle for some people to and to

some extent, completely undetectable (Jackson, McGorry, & McKenzie, 1994; Heinrichs, &

Carpenter, 1985; Malla & Norman, 1994). The length of this phase varies but may last up to

3
several months in some people. Stage 2 is known as the acute phase and is characterised by

severe, observable psychotic symptoms. This is typically the phase when the person gets

referred for diagnosis and treatment. Finally, stage 3 is known as the recovery phase. This is

the point when the person’s psychotic symptoms start to recede with the help of an

appropriate treatment. Although recovery has been strongly linked with the delays in

treatment (Marshall, Lewis, Lockwood, Drake, Jones, & Croudace, 2005; Wunderink,

Sytema, Nienhuis, & Wiersma, 2009), recovery is also linked with a number of individual

and social factors. Despite the subjective nature of recovery, the prognosis of psychosis

following its first episode is generally good with approximately between a quarter and a third

of the people never re-experiencing any form of psychosis again after the initial episode

(Wunderink et al., 2009; Verma, Subramaniam, Abdin et al., 2012 ).

When a person loses touch with reality, that person also loses touch with people who

are important to them (e.g. family & friends) and his/her surrounding environment (e.g.

school or work). The prodromal phase can be complicated for both the sufferers and their

family, as the subtle, peculiar changes in the sufferers’ behaviour can sometimes cause

misunderstanding, or even a break down in relationships. The acute phase, on the other hand,

can be a very frightening and traumatic experience not just for the sufferers but also for the

people closest to them (Jackson, Knott, Skeate, et al., 2004; Barton & Jackson, 2008). A lack

of awareness on the part of a family who is caring for an acutely psychotic individual can

make a difficult situation distressing for both parties. Unfortunately, identification of

psychosis can also be problematic for some general practitioners and many non-mental health

nurses (Lamph, 2010). The formation of a specialist service such as the Early Intervention

Teams meant that specialist intervention is now available at the earliest sign of psychosis.

Intervening at the earliest possible stage not only reduces the trauma associated with the acute

psychotic phase, but also ensures a much better prognosis and recovery (Birchwood &

4
McMillan, 1993a; Birchwood, McGorry, & Jackson, 1997; Birchwood, Fowler & Jackson,

2001; Craig, Garety, & Power, 2004; McGorry & Jackson, 1999; NICE, 2009).

Given that psychosis typically occurs between late adolescence and the early years of

adulthood (18 – 25), a particularly crucial period for identity formation and psycho-social

development, its disruptive effect often prompts secondary problems such as lack of self-

confidence or self-esteem (Birchwood, Fowler, & Jackson, 2001; Gumley, O'Grady, Power,

& Schwannauer, 2004; Gumley, Karatzias, Power, et al., 2006). A number of studies

conducted by Birchwood and his colleagues indicated that individuals experiencing an FEP

were also more prone to depression and suicidal ideation (Birchwood, Smith, McMillan et

al., 1989; Birchwood, Mason, McMillan, & Healy, 1993b; Rooke & Birchwood, 1998; Iqbal,

Birchwood, Chadwick, & Trower, 2000).

1.3. The Prevalence of Suicide in First Episode Psychosis

The reported prevalence rate of suicidal attempt from the moment psychosis starts

until the onset of treatment (also known as the duration of untreated psychosis) is between

6.5 and 9.6% (Clarke, Whitty, Browne et al., 2006; Foley, Jackson, McWilliams et al., 2008).

In studies whose samples were recruited from the initial presentation to psychiatric service,

rather than the actual psychosis onset, the rate of suicidal attempt prior to starting the initial

treatment is between 14% and 28% (Bertelsen, Jeppesen, & Petersen, 2007; Robinson,

Harris, Harrigan et al., 2009; Barret, Sundet, Faerden et al., 2010). Despite the timing

discrepancy upon which the rates of suicidal attempts were measured, the incidence of

suicidal attempts following treatment remains high across the FEP spectrum. Short-term and

long-term follow-up studies indicated a variety of results. One-year follow up studies

reported a prevalence rate of attempted suicide between 2.9 and 25.4% (Addington, Williams,

Young, & Addington, 2004; Bakst, Rabinowitz, & Bromet, 2010; Nordentoft, Jeppesen,

5
Kassow et al., 2002; Robinson et al., 2009), while 2 to 7-year follow-up studies reported a

prevalence rate of attempted suicide between 6 and 29.4% (Bakst et al., 2010; Foley et al.,

2008; Levine, Bakst, & Rabinowitz, 2010; Melle, Johannesen, Friis, et al., 2006; Robinson et

al., 2010; Walsh, Harvey, White et al., 2001). In schizophrenia alone, the lifetime prevalence

rate of completed suicide is about 5% (Palmer, Pankratz, & Botswick, 2005; Hor & Taylor,

2010), with the highest suicide risk during the early stages of the illness (Brown, 1997; Harris

& Barraclough, 1997; Palmer et al., 2005). Studies on the FEP spectrum over a 4 to 5-year

follow-up period, on the other hand, have estimated the rate of completed suicide in early

psychosis between 1 and 3% (Bertelsen et al., 2007; Clarke et al., 2006; Crumlish, Whitty,

Kamali et al., 2005)

The variability in the prevalence rates of both completed and attempted suicide within

the FEP spectrum is probably due to two things. One, the “diagnostic instability” during the

early phase of psychosis (Haahr, Friis, Larsen et al., 2008) makes the identification of eligible

research volunteers complicated for many researchers. Two, the timing discrepancy due to

the psychosis being inconsistently detectable during its prodromal phase also contributes to

the variability of the study time scales. So far, only a few studies have looked into the

incidence of suicidal attempts and deliberate self-harm within the duration of untreated

psychosis, or that time between the psychosis onset and start of treatment (Harvey, Dean,

Morgan et al., 2008; Upthegrove, Birchwood, Brunnet, McCollum, & Jones, 2010). The

majority of studies have focused only on the time between the initial presentation and

treatment onset, or the treatment onset and follow up period.

6
1.4. Risk Factors of Suicide in FEP

In a systematic review conducted by Hawton and his colleagues in 2005, it was found

that a number of risk factors for suicide in schizophrenia were comparable to that of the non-

psychiatric population (e.g. previous suicidal attempts, depression, recent loss, & drug

misuse). Similar risk factors were found when Hor and Taylor (2010) conducted a systematic

review on the studies published after June 2004, the cut- off date for the studies included in

the previous review. Hor and Taylor (2010) have also found that in addition to those factors

that were shared by the non-psychiatric population, being young, male, and well educated

emerged to be the strongest risk factors. Just like Hawton et al. (2005), Hor and Taylor

(2010) found other risk factors that were illness-specific. Previous studies have shown that

individuals suffering from psychotic illnesses are not only at high risks of attempted suicide

(Harris & Barraclough, 1998; Harkavy-Friedman, 2006), but also completed suicide (Brown,

1997; Saha, Chant, & McGrath, 2007). According to Limosin et al. (2007), the risk of

completed suicide for individuals with psychosis is about 16 times greater than that of the

non-psychiatric population. It is for this reason why a more precise identification of the risk

factors in this particular clinical group is of great clinical importance. The risk factors

identified below were extracted from studies that investigated suicidal behaviour within the

FEP spectrum.

1.4.1. Demographic Risk Factors

The most commonly cited demographic risk factors that were found to be

significantly associated with completed suicides are young age (Ceskova et al., 2011; Walsh

et al., 2001) and male gender (Ceskova, Prikryl, & Kasparek, 2011; De Hert, McKenzie, &

Peuskens, 2001). Alternatively, the risk factors associated with attempted suicide are female

gender (Cotton, Lambert, Schimmelmman et al., 2009; Hawton, 1997; Melle et al., 2006;

7
Nordentoft et al., 2002; Zahl & Hawton, 2004) and younger age at illness onset (Barret et al.,

2010). Substance abuse issues, non-compliance to treatment, and more impaired cognitive

functioning are some of the key reasons why the incidence of completed suicide is higher in

males than females (Cotton et al., 2009).

1.4.2. Clinical and Psychosocial Risk Factors

The most common clinical symptoms that predicted suicidal attempts in both short-

term and long-term follow-up studies were depression (Barret et al., 2010; Bertelsen et al.,

2007; Cohen, Lavelle, Rich, & Bromet, 1994; Cotton et al., 2009; Crumlish et al., 2005;

Flanagan & Compton, 2012; Fialko et al., 2006; Hawton et al., 2005; Melle et al., 2006;

Robinson et al., 2010) and hopelessness (Cohen et al., 1994; Klonksy, Kotov, Bakst,

Rabinowitz, & Bromet, 2012; Nordentoft et al., 2002; Robinson et al., 2009), along with

greater insight (Barret et al., 2010; Crumlish et al., 2005; Flanagan & Compton; 2012; Foley

et al., 2008). Illness-specific risk factors that were found to be associated with subsequent

suicidal attempts include the early stages of the illness (Brown, 1997; Harris & Barraclough,

1997; Palmer et al., 2005), less positive symptoms (Verdoux, Liraud, Gonzales et al., 2001),

hallucination (Bertelsen et al., 2007; Fialko, Freeman, Bebbington et al., 2006; Nordentoft et

al., 2002), negative beliefs (Barret et al., 2010; Fialko et al., 2006), and anxiety (Fialko et al.,

2006). Other illness-related factors that were also linked to the recurrence of suicidal

behaviour are duration of untreated psychosis (Clarke et al., 2006; Melle et al., 2006), higher

premorbid functioning (De Hert et al., 2001), and prolonged initial admission (Verdoux et al.,

2001).

8
1.4.3. Behavioural Risk Factors

Overall, the most prevalent risk factor for suicidal behaviour across the FEP spectrum

is the history of self-harm and/or suicidal attempt (De Hert et al., 2001; Hawton et al., 2005;

Flanagan & Compton, 2012; Nordentoft et al., 2002; Robinson et al., 2010; Verdoux et al.,

2001). A number of studies have indicated that problems with alcohol also increased the risk

of attempted suicide in early psychosis (Fialko et al., 2006; Robinson et al., 2010). According

to Verdoux (2001), the risk of engaging in suicidal behaviour over a 2-year follow-up was

seven-fold in substance abusers. According to the systematic review conducted by Hawton

and colleagues (2005), drug misuse was also found to significantly increase the suicidal risks

in FEP patients. On the other hand, Tiihonen, Wahlbeck, and Lonnqvist (2006) have

indicated that recently discharged first-episode schizophrenia patients were about 37 times

more likely to die by suicide than those who are at a later stage of the psychotic illness. This

was especially true for those who have an irregular compliance to their anti-psychotic

medication.

1.5. Hopelessness as a risk factor in FEP

There is an enormous amount of literature on the role of hopelessness as a risk factor

of attempted and completed suicide in both non-psychiatric and psychiatric populations.

However, to date, there are only three prospective studies that looked into the relationship

between hopelessness and subsequent suicide attempts within the FEP spectrum (Klonksy,

Kotov, Bakst, Rabinowitz, & Bromet, 2012; Nordentoft et al., 2002; Robinson et al., 2009).

The other published studies were either retrospective or cross-sectional, with samples that

were not exclusive to FEP (Borgeois, Swendsen, Young et al., 2004; Cohen et al., 1994;

Kim, Jayathilake, & Meltzer, 2002; Montross, Kasckow, Golshan et al., 2008).

9
1.5.1. Hopelessness: Studies that Link Suicidality in FEP and Hopelessness

Of the three prospective studies that examined the link between hopelessness and

suicidal risks in FEP, only two studies were able to demonstrate the predictive value of

hopelessness in determining the recurrence of suicidal behaviour in this particular sample. In

a randomised controlled trial of first-episode schizophrenia-spectrum patients, Nordentoft and

colleagues (2002) found that baseline hopelessness was significantly associated with the

attempted suicides during the 1-year follow-up. However, hopelessness did not emerge as a

predictor of subsequent suicidal attempts after controlling for the effects of the other clinical

variables in the multivariate analysis (e.g. positive & negative symptoms, depression, etc.).

On the contrary, Robinson et al.’s (2010) 7.4 year follow-up study on the prevalence and risk

factors of suicide in FEP indicated that along with self-harm, suicidal tendencies, and

depression, hopelessness emerged as one of the key predictors of subsequent suicidal

attempts during the follow-up period. More importantly, Robinson and colleagues (2010)

confirmed that the predictive value of hopelessness holds true after covarying out the effects

of age at psychosis onset, gender, DUP and previous self-harm. A similar pattern of results

were found in Klonsky et al.’s (2012) 10 year cohort study of first admission patients with

psychosis. To this point, this is the only study that specifically set out to examine

hopelessness as a predictor of future suicidal behaviour exclusively within the FEP spectrum.

Results from this authoritative study revealed that baseline hopelessness significantly

predicted subsequent suicidal attempts within the 10-year follow-up period, and this holds

true after controlling for depression. Intriguingly, further analyses indicated that the

predictive power of hopelessness was strongest over short-term intervals (2 years or less).

Klonsky and colleagues (2010) indicated the predictive power of hopelessness as a risk factor

of a suicide attempt remained strong only until the subsequent 2 years. Following the 2-year

period after the hopelessness was assessed, the predictive power of hopelessness declines to

10
the minimum level. For example, results showed that baseline hopelessness predicted suicide

attempts up until the 24-month follow-up period, but not the later follow-ups (i.e. 48 month –

10 years). A similar pattern of results was found when hopelessness was measured at 24th

month and predicted suicide attempts over the subsequent 2 years (i.e. between 24 and 48th

month follow up), but not the follow-ups after that.

Despite the congruence of outcome of Robinson et al.’s (2010) and Klonsky’s et al.’s

(2012) studies, the overall findings about the predictive role of hopelessness on future

suicidal attempts in FEP are subject to a number of limitations, thus, making it hard to

interpret. First, the studies used different scales to measure the construct of hopelessness.

Klonsky et al. (2012) employed the Beck Hopelessness Scale (BHS; Beck, Weissman, Lester,

& Trexler ,1974). The BHS is a tool that is renowned for being the “gold-standard” measure

of hopelessness. On the other hand, Robinson et al., (2010) employed The Royal Park

Multidiagnostic Instrument for Psychosis or the RPMIP (McGorry, Singh, Copolov et al.,

1990). Although the RPMIP has a respectable reliability (mean kappa for all items = .70) and

validity (RPMIP vs. DSM-III-R: kappa = .65, 74% agreement) overall, to date, there is no

published information about the validity and reliability of the hopelessness items in this

measure. Second, due to the fact that Robinson et al.’s study (2010) was only on a part of an

overarching research programme, the study was not originally designed to examine the risk

factors of suicidal behaviour. Due to this, hopelessness was only measured at baseline and

unlike Klonsky et al. (2012), Robinson and colleagues (2010) was not able to demonstrate the

trajectory of hopelessness as a predictor of attempted suicide across the different follow-up

periods. Third, although both studies were able to demonstrate the relationship between

hopelessness and recurrence of suicidal behaviour in early psychosis, both studies were not

able to capture the social contexts (e.g. social support) and other clinical or behavioural

11
factors (e.g. depression, problem-solving skills, & others) that might have facilitated

hopelessness, and more importantly the recurrence of suicidal behaviour.

1.5.2. Hopelessness: The Need for a Theoretical Model in Psychosis

While there is an increasing amount of information on the risk factors of suicidal

behaviour in early psychosis, to date, there is no model that explains the mechanism of

suicidal thinking in either the FEP spectrum or general psychotic disorders. With the

mortality rate by suicide in schizophrenia alone being 10 times greater than the non-

psychiatric population (Nordentoft, Laursen, Agerbo et al., 2004) and the first-episode

patients being at higher risks of killing themselves than those who are at a later stage of the

illness (Bertelsen et al., 2007), it is crucial to have a model of suicidality that takes into

account the experience of psychosis. Although previous studies have shown that the risks of

attempted suicide in FEP are strongly linked with depression (Barret et al., 2010; Bertelsen et

al., 2007; Cohen et al., 1994; Cotton et al., 2009; Crumlish et al., 2005; Flanagan &

Compton, 2012; Fialko et al., 2006; Hawton et al., 2005; Melle et al., 2006; Robinson et al.,

2010) and hopelessness (Cohen et al., 1994; Klonksy et al., 2012; Nordentoft et al., 2002;

Robinson et al., 2009), along with higher insight (Barret et al., 2010; Crumlish et al., 2005;

Flanagan & Compton; 2012; Foley et al., 2008), there are risk factors that are specific to the

experience of the illness itself. For example, the early stage of the illness (Brown, 1997;

Harris & Barraclough, 1997; Palmer et al., 2005) has been linked to subsequent suicidal

attempts. At this point in time, it is a fact that the presence of a psychotic disorder (Cohen et

al., 1994; Nordentoft et al., 2004; Verdoux et al., 2001), especially those with significant

depressive symptoms (Barret et al., 2010; Bertelsen et al., 2007; Cohen et al., 1994; Cotton et

al., 2009; Crumlish et al., 2005; Flanagan & Compton, 2012; Fialko et al., 2006; Hawton et

al., 2005; Melle et al., 2006; Robinson et al., 2010), are at particular high risks for both

12
attempted and completed suicide. However, it is not exactly clear if the experience of

psychosis per se, more specifically the early phase of the illness, has any impact on the

relationship between hopelessness and risks for suicidal behaviour.

So far, the literature on suicidal behaviour more broadly is predominantly limited by

two things: First, theoretical models of suicidality were narrowly grounded on either

biosocial [i.e. Schotte & Clum’s stress-diathesis model (1987)] or the cognitive [i.e.

Baumeister’s Escape theory (1999) & Carver & Scheier’s Self-regulation or Goal-

Disengagement model (1998)] aspects of suicidal behaviour. And although the stress-

diathesis model (Schotte & Clum, 1987) paved the way to the conception of two of the most

promising theoretical models of suicidal behaviour to date [i.e. Williams & Pollock’s Cry of

Pain Model (2001), which led to the formation of the Differential Activation Hypothesis of

suicidal relapse (Lau, Segal, & Williams, 2004)], the earlier stress-diathesis model (Schotte

& Clum, 1987) was simply restricted to the importance of certain risk factors and the

relationship between them (Mann, Waternaux, Haas, & Malone, 1999). Second, samples used

to test these models were limited to either the non-psychiatric, healthy population, or

currently and previously recovered depressed individuals. In view of these two current

limitations in the literature, the present study looks into the application of the DAH of

suicidal relapse (Lau, Segal, & Williams, 2004) as a potential model for understanding

suicidality in psychosis. With the assumptions of the DAH framework focusing on the

underlying mechanism of suicidal thinking, the applicability of this model in psychosis will

help clinicians manage and prevent suicidal relapse better. This is especially crucial in FEP as

the suicidal thinking process often happens too quickly and the incidence of completed

suicide often happens unexpectedly.

13
1.6. The Differential Activation Hypothesis of Suicidal Relapse

The DAH of suicidal relapse (Lau et al., 2004) was an extension of Teasdale’s DAH

of depressive relapse (1988), which in brief suggests that due to the formation of a link

between the depressed mood and certain negative thinking patterns during the early

depressive episodes, reoccurrences of low mood will trigger these patterns of negative

thinking (Teasdale & Barnard, 1993). The ease and the extent to which these negative

thinking patterns are triggered by the depressed mood is what Teasdale referred to as the

“cognitive reactivity” to depression (Teasdale & Barnard, 1993). Lau and colleagues (2004)

extend Teasdale’s DAH of depressive relapse (1988) by employing the assumptions of his

model to explain the mechanism of suicidal thinking. By adopting the assumptions of the

original DAH, the differential activation model of suicidal relapse suggests that during the

early episodes of depression, a link is formed between a depressed mood and a pattern of

negative and maladaptive thoughts. Hopelessness, as a form of an intensely negative, self-

referential thinking, occurs as part of these negative and maladaptive thinking patterns that

becomes associated with the depressed mood. The link that is formed between the depressed

mood and hopelessness is then reinforced through repeated episodes of depression. The

stronger the link between the depressed mood and hopelessness, the more easy and likely

hopeless thoughts will get reactivated in the event that low mood reoccurs. In keeping with

Teasdale’s idea of “cognitive reactivity”, the ease and extent to which the depressed mood

can trigger hopelessness is what characterises the individual’s cognitive reactivity to

hopelessness (Lau et al., 2004; Williams et al., 2008). In other words, the greater the

reactivity to hopelessness is, the more vulnerable the individual is to a suicidal relapse. An

elevated CR to hopelessness would simply mean that even minor negative shifts in mood will

easily reactivate hopeless/suicidal cognition.

14
1.6.1. Generalised hopelessness vs. Cognitive reactivity to hopelessness

Historically, the word hopeless originate from the 16th century and was a combination

of the old English words “hopa”, which means “to place trust in, or to rely in”, and “leas”,

which means “without” (dictionary.com unabridged). By literally combining the meaning of

these two old English words together, the definition of hopelessness then becomes without

having anyone or anything to put your trust in. Linehan and colleagues (1983) described

hopelessness as the lack of reasons for living, while Beck and colleagues (1999) characterised

it as a negative outlook for the future. In 1975, Beck, Kovac, and Weissman first linked

hopelessness with suicidality and since then, numerous attempts have been made to uncover

the role of this multifaceted construct in suicidal behaviour.

To date, the literature on suicidal behaviour in early psychosis has only looked into

hopelessness as a generalised pessimistic view of the future, which is typically measured by

using the Beck Hopelessness Scale (Beck et al., 1974). In 2004, Lau and colleagues

introduced the concept of “cognitive reactivity to hopelessness” as the core idea of their DAH

of suicidal relapse. The term “cognitive reactivity” to hopelessness literally translates as the

vulnerability to hopeless thoughts. Unlike the concept of generalised hopelessness which

characterises how negative the individual perceives the future on the whole, cognitive

reactivity to hopelessness characterises the individual’s tendency to pessimistic thinking

given a negative situation. In summary, generalised hopelessness describes the overall

response of the individual to a difficult situation, while CR to hopelessness describes the

more immediate response should the individual encounter a difficult situation.

1.6.2. Studies on cognitive reactivity to hopelessness

As the DAH of suicidal relapse (Lau et al., 2004) is still in its early stages, there is

only a limited number of studies that currently supports its assumptions. So far, the

15
application of the DAH of suicidal relapse has only been tested on a sample or previously

depressed individuals, and that was largely due to the fact that suicidal ideation is a one of the

most crucial symptoms of depression. In 2005, Williams, Barnhofer, Crane, and Beck

conducted a study to test the hypothesis. The main objective of the study was to investigate

the effects of mood on the individual’s problem-solving ability. Previous studies have

identified problem-solving deficit as a behavioural marker of hopelessness (Schotte & Clum,

1982; Orbach, Bar-Joseph, & Dror, 1990; Sadowsky & Kelly, 1993). The sample consisted of

3 groups: (1) 15 previously depressed individuals without the history of suicidal ideation, (2)

19 previously depressed individuals with a history of suicidal ideation, and (3) 22 never

depressed individuals. According to Williams and colleagues (2005), the lack of coping

options is exacerbated by impaired problem solving ability, which then facilitates the

escalation of the depressed mood into suicidal thoughts. A mood priming technique was

employed in order to test if a downward shift in mood will significantly impair the problem

solving ability of the previously depressed group with a history of suicidal ideation.

Consistent with the authors’ assumption, results showed that only the previously depressed

group with a history of suicidal ideation exhibited impaired problem solving performance

following a sad mood induction. However, the impairment was only evident in the

effectiveness, but not in the number of problem solving means. In other words, although there

was a significant decrease in the effectiveness of the problem solving means following the

mood challenge, quantity of the problem solving means generated did not differ between

groups. As cited by authors themselves, this study has a number of limitations. First, the

sample size is relatively small. Second, the autobiographical memory data are in conflict with

previous studies. The autobiographical memory task was also employed in this study as the

authors also speculated that impaired performance is associated with the lack of specificity in

autobiographical memory. Unexpectedly, scores of individuals with & without histories of

16
depression did not differ in autobiographical memory tests. Finally, the absence of a neutral

mood induction control group that could have helped identify the effects of any undesirable

variables on the problem solving performance.

Following this, Hepburn, Barnhofer, and Williams (2006) investigated the effects of

mood on future thinking on a sample of 52 non-depressed individuals. A number of studies

have evidenced that the lack of fluency in positive future thinking is a significant feature of

hopelessness (MacLeod, Rose, & Williams, 1993; MacLeod, Pankhania, Lee, & Mitchell,

1997; MacLeod & Byrne, 1996). The sample was randomly allocated to the two mood

priming conditions (positive vs. negative). The results were in agreement with the authors’

hypothesis as the negative mood induction reduced the fluency for good events while the

positive mood induction reduced the fluency for bad events. Intriguingly, however, the

negative mood induction did not increase the fluency for bad events just as the positive mood

induction did not increase the fluency for good events. The authors suggested that such a

pattern of results may be due to the possibility that future fluency was more sensitive to

diminution than increase when subjected to subtle mood changes. In addition to investigating

the effect of mood on future thinking, it was also predicted that future fluency is due to

mood-related changes in the evaluation process (i.e. positive vs. negative categorisation of

events). Although evidence showed that mood influenced the perceived valence of events

(e.g. good events were rated as more negative in a sad than recovered mood), the change in

pre- to post-induction future fluency did not correlate with the change in pre- to post-

induction valence ratings. However, the authors suggested the lack of association between

perceived valence and future fluency might be due to the small sample size and should

therefore not be ruled out on this occasion. In addition to the sample size, there are further

limitations to this study. The sample consisted of non-depressed students whose

characteristics are different to that of a clinical sample, who are at higher risks of suicidal

17
behaviours. Although the mood challenge in general altered future fluency as predicted, the

lack of neutral mood induction made it slightly difficult to isolate the effects of the positive

and negative mood induction on future fluency. However, overall, this study was an

important step in the literature of the DAH of suicidality. The confirmation that even subtle

shifts in mood altered future fluency was a novel and important finding, which served as a

platform for studies that aim to explore the mechanism of hopeless or suicidal cognition.

Williams, Crane, Barnhofer, Van der Does, and Segal (2006) also published a study,

which prospectively examined the recurrence of suicidal ideation across depressive episodes.

The aim of this study was twofold: (1) to examine the extent of association between suicidal

ideation and other symptoms of major depression across depressive episodes, and (2) to

investigate the nature of inconsistencies in suicidality across episodes, and when they arise. A

total of 69 individuals with a history of Major Depression (MDD) were allocated to treatment

as usual and prospectively studied over a 12-month period. Follow-up data revealed that a

total of 38 individuals (56%) had a recurrence of depression. Results suggest that suicidal

ideation is the only symptom that appears consistent across depressive episodes. The authors,

however, have pointed out that there was a decrease in severity of suicidality from previous

episode to recurrence. One of the reasons is the possibility that patients might have

underreported current suicidal ideation in fear of intervention. Overall, this study provided

initial evidence on the recurrence of suicidal behaviour across depressive episodes in line

with the assumptions of DAH of suicidal relapse. According to the DAH, once suicidal

ideation has occurred during a depressive episode, it is more likely to reoccur along with the

re-emergence of another depressive episode. The two main limitations of this study include

the small sample size, and that the reoccurrence of suicidal ideation was measured in the

absence of suicidal attempts.

18
 Enthused by the results of the previous study, the same authors (Williams, Van der

Does, Barnhofer, Crane, & Segal, 2008) conducted a study to investigate if the reoccurrence

of suicidal or hopeless thoughts over time can be illustrated using the DAH of suicidality.

There were 3 parts to this study. The aim of studies 1 and 2 was to examine if cognitive

reactivity as measured by the hopelessness/suicidality subscale of the Leiden Index of

Depression Scale – revised version (Van der Does & Williams, 2003) will be associated with

previous suicidal ideation. On the other hand, the aim of study 3 was to examine if higher CR

as measured by the LEIDS’ hopelessness subscale will be associated with impairment in

future fluency when in a sad mood state. Participants in the study 1 consisted of 36

previously depressed and 80 never depressed first year undergraduate psychology students,

while participants in study 2 consisted of 63 previously depressed and 57 never depressed

middle-aged adults. Participants who have been identified as previously depressed met the

criteria for the previous Major Depression using Major Depression Questionnaire (Van der

Does, Barnhofer, & Williams, 2003). As predicted, results of studies 1 and 2 indicated that

individuals who had higher scores on LEIDS’ hopelessness/suicidal ideation subscale also

admitted to having suicidal ideations in the past when feeling depressed. Such pattern of data

was consistent with the previous study (Williams et al., 2006), which illustrated that suicidal

ideation was the only symptom that was consistent across depressive episodes. On the other

hand, participants in study 3 were a subgroup of individuals who took part in study 2. Of the

32 individuals who met the inclusion criteria, 13 had been previously depressed without

suicidal ideation, 5 had been previously depressed with suicidal ideation, and 14 had never

been depressed at all. Results for study 3 also confirmed the assumption that CR, as measured

by the LEIDS’ hopelessness/suicidality subscale, was predictive of the changes in positive

future fluency (as measured using the Future Fluency Task) following the sad mood

induction. Similar results were found in an earlier study conducted by Hepburn and

19
colleagues (2006), which indicated a decrease in positive future fluency following a sad

mood induction. Despite the promising results, careful considerations must be observed when

drawing conclusions from this study. The authors have identified a number of limitations.

Firstly, the sample size is relatively small and the rate of depression in both genders did not

differ. Secondly, the specificity of the relationship between history of suicidal ideation and

CR to hopelessness was based on two things: (1) with the exception of guilt, all other

depressive symptoms did not predict CR to hopelessness, and (2) history of suicidal ideation

did not predict the other subscale of LEIDS. Despite the association between history of

suicidal ideation and CR to hopelessness holding true after controlling for current depression

and severity of past depression, the authors suggested that the sample size might be lacking in

power to detect the differences between the other subscales of LEIDS. More importantly, the

lack of distinction between the magnitude of group differences in the hopelessness subscale,

and the other LEIDS’ subscales across the two groups (with suicidal ideation vs. without

suicidal ideation) seemed to suggest that history of suicidal ideation may not only be specific

to higher CR to hopelessness. Further analyses showed that both the ruminative and avoidant

tendency were significant covariates to CR to hopelessness. Finally, the authors pointed out

that the LEIDS’ hopelessness subscale was devised to measure CR to hopelessness or

suicidal ideations and not to suicidal attempts. As the majority of the participants in this study

only had histories of ideation, the interpretation of results with regards to the use of this

subscale is therefore limited only within this type of sample. The authors recognised that this

measure needs to be validated in a clinical sample with higher rates of suicidal attempts.

In summary, the results of the previous studies support the assumptions of the DAH

of suicidal relapse by demonstrating that a subtle downward shift in mood impairs problem

solving (Williams et al., 2005) and fluency for positive events (Hepburn et al., 2006), two of

the most widely recognised characteristics of hopelessness. The observed recurrence of

20
suicidal behaviour across depressive episodes also renders support to the DAH of suicidal

relapse, which suggest that once suicidal ideation occurs as a feature of an early depressive

episode, it is more likely to reoccur in another depressive episode (Williams et al., 2006 &

2007). Together, these results illustrate that vulnerability to suicidal thinking can be measured

via quantifiable behavioural features of hopelessness (e.g. problem solving & future fluency).

Identification of the most relevant risk factors for suicide is crucial for a successful

prevention and treatment of suicidal behaviour. While the DAH framework is only in its

infancy, evidence that supports its concept on cognitive reactivity represents a good starting

point for further investigation of the suicidal thinking mechanism.

21
 CHAPTER 2

THE EXPERIENCE SAMPLING METHOD

2.0. Introduction

One of the major difficulties in investigating suicidal thinking is the lack of context

that is relevant to the occurrence of this pernicious thinking process. To date, the suicidal

thinking process has only been studied within the confines of the laboratory. The purpose of

this chapter is to review the scientific literature on the Experience Sampling Method

(Delespaul, 1995; de Vries, 1992) in order to assess its potential as a research tool for

investigating the suicidal thinking process in early psychosis. Preceding the review is a brief

discussion of the general aspects of the ESM, its definition and use in psychosis research,

limitations and strengths, and validity and reliability as a research method. Following this is a

review of the previous ESM studies in psychosis. The concluding discussion will point at the

application of the ESM as a potential tool for investigating the suicidal thinking mechanism

in psychosis using the DAH framework.

2.1. The Experience Sampling Method (ESM): An Overview

The ESM was originally defined as a process of collecting data about a person’s daily

life experiences (Hektner, Schmidt, & Csikszentmihayli, 2007). The use of ESM was first

initiated by Hektner, Schmidt, and Csikszentmihayli during the early 1970’s to study the

‘flow’ (Hektner et al., 2007) of daily life experiences. It all started with the use of pagers

activated at random times from a central radio station, prompting people to write in their

diaries about the things they have done and enjoyable moments of their day. The last 4

decades of research has transformed the ESM into using a more structured diary method,

22
making it a widely popular tool in investigating an extensive range of human behaviour and

activities, in a variety of disciplines (e.g. psychology, sociology, & anthropology).

Researchers of contemporary ESM studies characterise the ESM as a systematic diary

keeping technique, which requires individuals to fill in a self-report questionnaire at

predetermined times of the day within his/her real life environment (Delespaul, 1995; de

Vries, 1992). The term “diary keeping” was central to the description of the ESM for two

reasons: (1) the questionnaires are compiled in a form of a small diary or booklet, and (2) just

like the traditional diary; the ESM booklet keeps a record of daily events and activities over a

specific period of time. The self-report questionnaires in an ESM diary usually consists of

open- and close-ended, Likert formatted questions, which were formulated to assess for

topics that are of key interest in the study. Depending on the study, each questionnaire will

take about 1 to 1.5 minutes to complete and each ESM diary consists of at least 3 to 10

identical questionnaires that are to filled in one questionnaire at a time, as and when

prompted by a programmable device (e.g. a digital wristwatch, mobile phone, personal digital

assistant/PDA, or beeper), at predetermined times of the day. Sampling time schedules are

always semi-randomised in order to avoid clustering of the sampling time points (Delespaul,

1995; de Vries, 1992). The duration of ESM studies vary from a day to several years

(Csikszentmihayli & Schneider, 2001).

2.2. The Use of ESM in Psychosis Research

In clinical research, the ESM is also referred to as the “Ecological Momentary

Assessment” (Stone & Shiffman, 1994). However, for the purpose of this review, only the

term ESM will be used. The use of ESM as a research tool in psychosis has come a long way

since it was initially used in 1987 (Hurlburt & Melancon) when a patient with schizophrenia

was asked to write a narrative description of her daily hallucinatory experiences. Over the last

23
25 years, this purely qualitative, freestyle-written diary method has evolved into what is now

the present-day ESM, a diary keeping technique that is more systematic and structured. The

advancement of ESM as a research tool in psychopathology is largely due to a group of

clinicians and researchers from the University of Maastricht in The Netherlands (e.g.

Delespaul, de Vries, Myin-Germeys, Van Os, & others). At the same time Hurlburt and

Melancon (1987) first used ESM on a single case study, Delespaul and de Vries (1987)

devised an ESM diary with open- and close-ended questions. The reformulated diary

questionnaire was used to capture the daily life experiences of 11 non-psychiatric volunteers

and 11 patients with chronic mental illness. Through the use of both open- and close-ended

questions, Delespaul and de Vries (1987) were able to qualitatively and quantitatively

measure the day to day activities of both groups, and more importantly, the illness-related

experiences of the patient group. Since then, the use of ESM has been dramatically

transformed from being a mere qualitative measure to a dual-function research tool that is

capable of sampling qualitative and quantitative data all together.

2.2.1. The Contemporary ESM in Psychosis Research

Nowadays, the ESM questionnaire or the experience sampling form (ESF; Delespaul,

1995; Delespaul & de Vries, 1987) in psychosis research generally consist of questions about

the individual’s thoughts, mood, somatic and psychotic symptoms, context (e.g. place &

people), activities, and events. Questions on thoughts, context, activities, and events are a

combination of open-ended (e.g. “What are you thinking?” or “Who are you with?”) and

close-ended, follow-up questions (e.g. “My thoughts are pleasant.” or “My thoughts are

clear.”) with a Likert-type response scale (1 = not at all and 7 = almost always). On the other

hand, questions on mood, somatic, and psychotic symptoms are entirely close-ended (e.g. “I

feel tired.” or “I feel secure.”) with the identical 7-point Likert-type response scale. The

24
questions on the ESF are based on the standard mental health examination procedure in

psychiatry while the coding of the open-ended questions is based on the ESM instruction

manual formulated by Delespaul and de Vries (1987). Although many researchers still use

some of the components of the original Delespaul and de Vries’ ESF (1987), the entire

content of the ESF can be tailored according to the purpose of the study provided that: (a)

newly formulated questions must be piloted to establish its reliability and validity, (b)

completion time of the entire ESF must be between 2 to 3 minutes to retain good compliance

(Delespaul, 1995; Delespaul & deVries, 1994).

Until recently, the use of ESM in psychosis research has always been conducted using

a paper and watch procedure. Paper based diaries (typically A6 in size) along with a

signalling device (e.g. digital wristwatch) were considered to be the most economical,

convenient, easy, and efficient way of conducting ESM studies (Palmier-Claus, Taylor,

Gooding, Dunn, & Lewis, 2011). With the recent advances in handheld computing

technology, two studies were able to demonstrate the use of electronic devices (i.e. PDA’s) to

conduct the ESM in a sample of patients with psychotic disorders (Kimhy, Delespaul,

Corcoran et al., 2006; Granholm, Loh, & Swendsen, 2008). Findings from both studies

indicated a high compliance rate (87% - Granholm et al., 2008 & 80% - Kimhy et al., 2006),

that did not significantly differ from that of the non-psychiatric control group (81% - Kimhy

et al., 2006). Whereas the participants in Granholm and colleagues’ study (2008) reported

positive feedback for their overall electronic diary keeping experience, participants in Kimhy

et al.’s (2006) study found the use of electronic devices quite challenging.

2.2.2. Feasibility and Compliance

The feasibility of employing the ESM in a sample of individuals, with a spectrum of

psychotic disorders has already been demonstrated in previous studies (Delespaul, de Vries,

25
& Van Os, 2002; Delespaul & de Vries, 1987; Hurlburt & Melancon, 1987; Myin-Germeys,

Delespaul, & de Vries, 2000; Myin-Germeys, Nicolson, & Delespaul, 2001; Myin-Germeys,

Krabbendam, Jolles, Delespaul, & Van Os; 2002; Myin-Germeys, Krabbendam, Delespaul, &

Van Os, 2003; Lardinois, Myin-Germeys, Bak, Mengelers, Van Os, & Delespaul, 2003, &

many others). Despite the relatively high drop-out rates in patients with more severe and

chronic psychotic symptoms (Oorschot et al., 2009), previous studies have illustrated that a

respectable number of valid diary reports can be achieved in this particular clinical sample.

Compliance rates in ESM studies are calculated by dividing the total number of valid

diary reports (also called ESF) completed with the total number of expected reports. For

instance, 10 diary reports over 6 days is equivalent to 60 expected reports. If a participant

manages to complete 30 valid reports (completed within 15 minutes after the signal;

Delespaul, 1995) out of the 60 expected reports, then the calculated compliance rate is

equivalent to 50%. Oorschot and colleagues (2009) indicated that the compliance rate in

schizophrenia sample was around 66%, although higher rates were reported from other

studies (79% - Kimhy et al., 2010; 87% - Graholm et al., 2008).

2.2.3. Limitations and Strengths

The key strength of the ESM is that it measures real life experiences as they occur in

their natural context (Myin-Germeys et al., 2009). Due to this, the ecological validity is high

and the chances of selective memory or recall bias is minimal (Kiviniemi & Rothman, 2006;

Kikuchi, Yoshiuchi, Mikasaka, Ohashi et al., 2006; Myin-Germeys et al., 2009). As the ESM

is designed to repeatedly collect multiple data at different time points over a prolonged period

of time and more importantly, within the natural everyday life context of the participant, the

ESM data offers: (a) an opportunity to examine the role of contextual factors and its

interaction with thoughts, feelings, and behaviour of an individual, (b) a chance to explore

26
other potentially important underlying mechanisms or processes, and (c) a better

understanding of how the variables under study function over time (Myin-Germeys et al.,

2009).

The ESM also has a number of limitations that need to be borne in mind. As a self-

report assessment, the ESM is prone to subjective personal biases (Christensen et al., 2003).

However, it is important to note that the ESM was purposely devised to assess the

individual’s subjective account in order for researchers to understand the nature of his/her

personal experiences in everyday life. Hektner and colleagues (2007) pointed out that being

able to measure the individuals’ subjective experiences may bring a more practical insight

about the reality of some of the mental illnesses.

Another limitation is the relatively high financial cost of running an ESM study. Due

to its prolonged data sampling, time consumption is also high in ESM studies. Some of the

participants also find the repeated assessments and the overall length of the study quite

intense and challenging. Such issues sometimes cause participants to skip or miss a

significant amount of sampling times, “back fill” or “forward fill” their diaries (Granholm et

al., 2007), and misreport the time of reports in their diary (Hektner et al., 2007).

2.2.4. Validity and Reliability

As discussed briefly in the previous section, the fact that ESM relies on self-reports

poses questions on the validity of its procedure. However, the fact that a good number of

widely used psychometric measures in many clinical and research settings also rely on self-

reports, does not necessarily make the ESM a valid measure. As the completion of this self-

report measure depends on the prompts of a signalling device, the ESM reports are generally

less prone to selective memory biases or “recall biases”. Kimhy and colleagues (2006),

however, pointed out that due to the highly subjective nature of the ESM questions (i.e.

27
questions on thoughts & mood) and the natural tendency of these variables (e.g. mood &

thoughts) to vary over time, the ESM data may not necessarily capture a valid behavioural

outcome. Previous studies have illustrated that affective variability or instability is

characteristic of individuals who are at risk of developing psychosis (Delespaul & de Vries,

1987; Myin-Germeys et al., 2000; Palmier-Claus et al., 2011). Such variability in affect,

along with the fluctuations of the psychotic symptoms over time (Delespaul, 1995; de Vries,

1992), explains why the conventional reliability testing is not applicable for the ESM

questionnaire. Instead, Delespaul (1995) has suggested employing a “multiple indicator”

approach by looking at the reliability of correlated constructs (e.g. negative mood and

psychotic symptoms).

2.3. The ESM Studies in Psychosis

The use of the Experience Sampling Method (ESM) in psychosis research started 26

years ago with a single case study of a patient with schizophrenia (Hurlburt & Melancon,

1987). Prior to that, clinicians and researchers knew very little about the day to day

experiences of those who suffer from psychosis. Since ESM was first employed in psychosis

studies, researchers began to uncover some of the important aspects of the illness; from the

frequency of hallucinatory and delusional experiences to the momentary fluctuations of

mood, the incidence of substance misuse, the individual’s reactivity to minor stresses in

everyday life, along with the people and places that provide a suitable context for the

worsening or improvement of certain psychotic symptoms. However, to date, the ESM has

not yet been utilised to explore the occurrence and fluctuation of hopeless or suicidal

thoughts in a sample of individuals with psychosis. For the purpose of illustrating the

reliability of the ESM as a valid research tool for assessing momentary experiences in

psychosis, a summary of ESM studies published between 1987 and 2011 is presented on

28
Table 1. For the sake of brevity, only studies that are relevant to the ESM study in this thesis

will be discussed.

Of the identified studies, 9 investigated affective variability and/or stress reactivity

while the rest of the studies examined hallucinatory experiences (7), cognition and genetics

(4), substance misuse (2), and anticipatory pleasure towards daily activities (1). Of the 9

relevant studies, 3 were specifically focused on affective variability, 5 on stress reactivity,

and 1 on the relationship between life events and stress reactivity on a day to day basis. All

three studies on affective variability confirmed that affective variability is characteristic of a

psychotic illness (Delespaul & deVries, 1987; Myin-Germeys et al., 2000; Palmier-Claus et

al., 2011). Specifically, according to Myin-Germeys and colleagues (2000), patients with

schizophrenia had less variability and intensity in their positive affective responses but

greater variability and intensity in their negative affective responses. Contrary to the results

of previous laboratory-based experiments (Gaebel & Woelwer, 1992; Kring, Kerr, Smith, &

Neale, 1993; Kring & Neale, 1996), the ESM data presented by Myin-Germeys et al. (2000)

gave emphasis to the importance of contextual factors in the psychopathology.

Of the three studies on affective variability, Palmier-Claus et al.’s, (2011) was the

only one who investigated the association between ESM-measured affectivity, and baseline

severity and frequency of suicidal behaviour (i.e. ideation, suicidal attempt, or self-harm).

Although the study illustrated a link between affective variability and suicidal behaviour, the

results were limited by a number of important issues: (1) the sample size is relatively small

(N = 27), (2) the definition of suicidal behaviour is too broad (i.e. suicidal ideation was

included), (3) the assessment of suicidal behaviour was based on a retrospective interview,

and (4) the number of ‘suicidal’ individuals were not reported. More importantly, as affective

29
Table 1. Summary of ESM Studies in Psychosis
Authors Sample size and characteristics Sampling Results
Method
Delespaul & deVries (1987) N = 11 patients with schizophrenia 10 samplings/day Patients with schizophrenia displayed more variability in
& 11 non-psychiatric controls over 6-day period their thoughts, mood, & activity motivation

Hurlburt & Melancon (1987) Single case study of a 23 year old At least 10 Patient reported 71 narrative descriptions of distorted
patient with schizophrenia samplings/day images (e.g. blue glass appearing as yellow, patient seeing
over a 2-week things in a crooked or tilted angles/shapes) including
period visualisation of her voice in the form of hand-printed,
colourful displays

Myin-Germeys, Delespaul, & N = 58 schizophrenia patients with 10 samplings/day Schizophrenic patients had less intensity & deviations in
deVries (2000) blunted or non-blunted affect & 65 over 6-day period positive emotions but greater intensity & variability in
non-clinical controls negative emotions compared to the control group.
Blunted & non-blunted sub-groups did not differ in their
patterns of emotional experience.

Myin-Germeys, Nicolson, & N = 34 individuals with 10 samplings/day Increases in negative emotion & inactivity were associated
Delespaul (2001) schizophrenia spectrum disorder over 6-day period with delusional moments.
Delusional moments intensified auditory hallucinations.

Delespaul, deVries, & Van Os (2002) N = 57 individuals with 10 samplings/day Intensity of auditory hallucination increased with
schizophrenia spectrum disorder over 6-day period engagement in leisure activities and decreased with social
withdrawal.
Higher baseline anxiety was associated with subsequent
auditory hallucinations.

30
Table 1. Summary of ESM Studies in Psychosis
Authors Sample size and characteristics Sampling Results
Method
Myin-Germeys, Krabbendam, Jolles, N = 42 patients with schizophrenia 10 samplings/day Overall cognitive functioning did not influence emotional
Delespaul, & Van Os (2002) in remission over 6-day period sensitivity to stress, although some data illustrated that
higher cognitive functioning facilitated greater emotional
sensitivity to stress.

Myin-Germeys, Krabbendam, N = 42 patients with schizophrenia 10 samplings/day Life events (LE) were not associated with subjective
Delespaul, & Van Os (2003) in remission over 6-day period appraisal of stress (activity or event-related stress).
LE was associated with emotional reactivity (increased
NA & decreased PA).

Lardinois, Myin-Germeys, Bak, N = 35 individuals with psychosis 10 samplings/day Voice-hearing patients with more effective coping
Mengelers, Van Os, & Delespaul spectrum disorder over 6-day period strategies (e.g. not following the voices) experienced more
(2003) distress.

Myin-Germeys, Krabbendam, N = 42 patients with schizophrenia 10 samplings/day Female participants exhibited greater emotional reactivity
Delespaul, & Van Os (2004) in remission over 6-day period (increased NA & decreased PA) to every daily life stresses
compared to male participants.

Myin-Germeys, Delespaul, & Van N = 42 psychosis spectrum patients 10 samplings/day An increase in subjective stress (activity & event-related
Os (2005) in remission, 47 first degree over 6-day period stress) was associated with an increase in the intensity of
relatives, & 49 non-psychiatric psychotic experiences in the patient group.
controls

Kimhy, Delespaul, Corcoran, Ahn, N = 10 patients with schizophrenia 10 samplings/day The patient group and healthy control group did not differ
Yale, & Malaspina (2006) and 10 healthy controls over 6-day period in their ratings of stress.

31
Table 1. Summary of ESM Studies in Psychosis
Authors Sample size and characteristics Sampling Results
Method
Gard, King, Gard, Horan, & Green N = 15 patients with schizophrenia 7 samplings/day Patient group exhibited a more reduced anticipatory
(2007) and 12 healthy controls over 7-day period pleasure towards goal-directed activities (e.g. work &
studying) compared to the healthy controls

Morrens, Krabbendam, Bak, N = 25 patients with psychosis 10 samplings/day In some instances cognitive functioning was not associated
Delespaul, Mengelers, Sabbe, spectrum disorder over 6-day period with stress sensitivity while in other instances, the former
Hulstijn, Van Os, & Myin-Germeys was inversely related to the latter.
(2007)

Henquet, Rosa, Delespaul, Papiol, N = 31 patients with psychosis 10 samplings/day COMT Val(158)Met genotype moderates the association
Fananas, Van Os, & Myin-Germeys spectrum disorder & 25 healthy over 6-day period between cannabis use and psychotic experiences in
(2009) controls everyday life.

Lataster, Collip, Lardinois, Van Os, N = 40 patients with psychosis 10 samplings/day Stress reactivity in patient group and healthy controls was
& Myin-Germeys (2010) spectrum disorder & 47 healthy over 6-day period significantly associated.
controls (siblings of the patient
Positive psychotic symptoms and stress reactivity in
group)
healthy controls were also significantly associated.

Kimhy, Delespaul, Ahn, Cai, N = 20 patients with psychosis 10 samplings/day Momentary increases in stress had a negative correlation
Shikhman, Lieberman, Malaspina, & spectrum disorder over 6-day period with concurrent parasympathetic activity and positive
Sloan (2010) correlation with sympathovagal balance.

Ben-Zeev, Morris, Swendsen, & N = 113 patients with Unknown Negative self-esteem predicted delusional experiences
Graholm (2010) schizophrenia and schizoaffective sampling rate. while hallucination predicted delusions of control.
disorder
7-day period. Frequency of delusions of control was associated with
reduced ability to gather information.

32
Table 1. Summary of ESM Studies in Psychosis
Authors Sample size and characteristics Sampling Results
Method
Henquet, Van Os, Kuepper, N = 42 patients with psychosis 10 samplings/day Daily cannabis intake in the patient group predicted
Delespual, Smits, Campo, & Myin- spectrum disorder & 38 healthy over 6-day period increases in positive affect and decreases in negative
Germeys (2010) controls affect.

Thewissen, Bentall, Oorschot, N = 82 patients with schizophrenia 10 samplings/day An increase in anxiety and a decrease in self-self esteem
Campo, Van Lierop, Van Os, & and schizoaffective disorder & 37 over 6-day period predicted the onset of paranoid experiences.
Myin-Germeys (2011) healthy controls

Swendsen, Ben-Zeev, & Graholm N = 145 patients with Unknown A bi-directional relationship was found between substance
(2011) schizophrenia & schizoaffective sampling rate. use and psychotic symptoms.
disorder
7-day period

Palmier-Claus, Taylor, Gooding, N = 27 individuals at ultra high risk 10 samplings/day Ultra high risk individuals who previously reported
Dunn, & Lewis (2011) of developing psychosis over 6-day period suicidal ideation exhibited greater affective variability.

Varese, Udachina, Myin-Germeys, N = 42 patients with schizophrenia 10 samplings/day Dissociation was associated with auditory hallucinations
Oorschot, & Bentall (2011) spectrum disorder & 23 healthy over 6-day period during highly stressful situations.
controls

Lardinois, Lataster, Mengelers, Van N = 50 non-affective psychosis 10 samplings/day Childhood trauma was associated with greater affective
Os, & Myin-Germeys (2011) individuals over 6-day period and psychotic reactivity to the stresses of everyday life.

33
variability has already been found in individuals with psychosis (Delespaul & de Vries, 1987;

Myin-Germeys et al., 2000), the lack of a control group in Palmier-Claus et al.’s study (2011)

made the interpretation of results quite difficult.

The ESM studies on stress reactivity, on the hand, demonstrated that the healthy

controls were just as sensitive to the minor stresses in everyday life as the individuals

affected by psychosis (Kimhy et al., 2006; Lataster et al., 2006). Such findings were

unexpected as a previous study has indicated that the increase in the intensity of psychotic

symptoms in this particular group was associated with the increase in the subjective stress

caused by the minor strains in everyday life (Myin-Germeys, 2005). A year prior to this,

Myin-Germeys and colleagues (2004) also found that across the psychosis spectrum disorder,

sensitivity to stress was greater in females than males. Intriguingly, no link was found

between stress sensitivity and the incidence of recent life events in patients with a diagnosis

of schizophrenia (Myin-Germeys et al., 2003). Instead, life events were found to be

associated with greater affective variability in this particular group (Myin-Germeys et al.,

2003). However, Myin-Germeys and colleagues (2003) pointed out that the incidence of life

events did moderate the effect of minor stresses in everyday life on mood.

Overall, results of the previous ESM studies on affective and stress reactivity have

important implications for the concept of cognitive reactivity to hopeless and suicidal

thoughts as proposed by the Differential Activation Hypothesis of suicidal relapse (DAH;

Lau et al., 2004). First, the unstable affectivity of individuals affected by psychosis,

particularly the higher instability and intensity in positive affectivity (PA) than negative

affectivity (NA), might suggest that the individuals with psychosis are at a greater risk for

suicidal relapse. The more unstable and intense PA is than NA, the more likely the

reactivation of hopeless or suicidal thoughts may occur amongst those with histories of

suicidal attempt or deliberate self-harm. Second, the mediating effect of traumatic life events

34
on the individual’s affective responses to minor stresses in everyday life, sustains the idea

that early psychosis individuals will be more vulnerable to suicidal relapses. It is now

established that life following the initial episode of psychosis can be distressing for many

individuals. The traumatic experience of the illness itself, along with the other life events

associated with the illness (e.g. hospitalisation, leaving work or school due to psychosis,

stigma, & many others), will render this particular group of individuals more vulnerable to

affective variability when confronted with the everyday life stresses. Again, following the

assumption of the DAH, the more unstable PA, the more likely it may lead to NA reactivating

hopeless thoughts in previously suicidal individuals.

2.4. The ESM: A potential tool to test the DAH

The last 26 years has demonstrated the competence of the ESM to capture data that

were otherwise impossible to obtain from any laboratory setting. The ability of the ESM to

assess momentary fluctuations in mood, along with the changes in contextual factors, makes

this research technique an ideal tool to test the assumptions of the DAH for suicidal relapse.

Since the core idea of the DAH rests on the interactive relationship between mood and

hopeless/suicidal thoughts, it is vital to examine this relationship over a prolonged period of

time and in its most natural context. Whereas a number of laboratory procedures nowadays

can induce the appropriate emotional context necessary to elicit certain responses (whether it

is affective, cognitive, or behavioural), these procedures are by no means comparable to what

happens in real life. Although it can be argued that there are certain behaviours and

psychological processes that can be successfully studied within the premises of a laboratory,

there are also behaviours and processes which can only be meaningfully studied in their

natural context. Suicidal thinking has already been established to occur in the context of a

depressed mood and hopeless cognition (Lau et al., 2004). Such contexts are not only

35
unethical and precarious to replicate, these are also complicated emotional states that involve

a number of other contextual factors. For example, persistent social isolation and lack of

structured activities can both trigger feelings of despair. These are some of the contextual

factors that only exist in the context of the individual’s “natural habitat”. This is when

research techniques like the ESM is most needed. In testing the assumptions of the DAH, the

use of ESM not only makes it possible to understand that interaction between affectivity and

hopeless/suicidal cognitions, but also the interaction between the individual and his/her

natural, day to day environment. The repeated sampling over a period of time (typically 6

days) will help establish the pattern of fluctuations in mood and hopelessness across a range

of contexts (e.g. people, places, activities, & events).

2.5. Overview of Thesis

The overarching aim of this thesis is to investigate the mechanism of suicidal thinking

in early psychosis. The central aim is to examine if the reoccurrence of suicidal or hopeless

thoughts over time can be explained within the framework of the DAH of suicidal relapse

(Lau et al., 2004). To do this, two contrasting methodological approaches were employed.

First, in Chapter 3, an Experience Sampling Method (Delespaul, 1995; de Vries, 1992) is

employed to assess the momentary fluctuations in hopelessness in response to the changes in

mood over time. Second, in Chapters 4 and 5, a mood induction procedure is conducted in

order to induce feelings of sadness, a context that is necessary to illustrate the mechanism of

suicidal thinking as proposed by the DAH of suicidal relapse. In chapter 4, the Means-End

Problem Solving (Platt & Spivack, 1975) task is carried out before and after the mood

challenge in order to test if the change in mood altered the interpersonal problem ability, a

behavioural marker that is closely linked with hopelessness. In Chapter 5, the Future

Thinking Task (MacLeod et al., 1993) is also carried out before and after the mood challenge

36
(alongside the task in chapter 4), in order to test if the change in mood will reduce fluency for

positive events, a signature characteristic of suicidality or hopelessness.

2.6. Note on Collaboration

The author completed the research presented in this thesis in collaboration with a

number of other individuals. The author’s supervisors, Professor Max Birchwood and Dr.

Chris Jackson provided input on research development, design and write-up, and are

therefore recognised as co-authors. Dr. Inez Myin-Germeys and Margreet Oorschot are

recognised as co-authors on Chapters 3, where they contributed to the design of the ESM

questionnaire and the analyses of data. Recruitment was solely carried out by the author of

this thesis in collaboration with the team managers and care coordinators of the Early

Intervention Service in Birmingham. All of the analyses were undertaken solely by the author

of this thesis, with information and advice provided by the author’s supervisors. All write-ups

were solely the work of the author of this thesis, with the author’s supervisors providing input

in terms of feedback on drafts and ideas.

37
 CHAPTER 3

The Mechanism of Hopelessness Linked to the Mood Fluctuations

in Everyday Life: An ESM Study

3.0. Introduction

Empirical research into the underlying mechanisms of the suicidal thinking process in

the first episode psychosis sample is limited. Over the last decade, studies on suicidality in

psychosis have been mainly focused on the incidence and risk factors of suicidal behaviour.

So far, we know “what” makes these individuals want to end their own lives, and to a certain

extent, we understand “why” they have come to feel this way. And yet, we know very little

about the ‘how’ and the ‘when’ of this complex phenomenon. How does one acquire a

suicidal mind? When and how does it start? The answers to these questions are especially

crucial in the FEP sample as the stage of post-psychotic recovery is often characterised by a

rapid increase in suicidal attempts (Power, 2010). The aim of this study is to address this gap

in the literature by investigating the suicidal thinking process using the Differential

Activation Hypothesis of suicidal relapse (Lau et al., 2004) framework. Specifically, the core

idea was to uncover the relationship between positive and negative affectivity, and

hopelessness by employing the Experience Sampling Method (Delespaul & de Vries, 1987).

As discussed in Chapter 1, the DAH of suicidal relapse (Lau et al., 2004) suggests

that hopeless or suicidal thoughts occur as a feature of the maladaptive and dysfunctional

thinking process during a severe episode of depression. The key idea of the hypothesis is that

repeated episodes of depression will strengthen the link between the suicidal/hopeless

thoughts and the depressed mood. The stronger the link, the easier it will be for a subsequent

depressed mood to reactivate these suicidal/hopeless thoughts. Such ease in the reactivation

38
process is referred to as the “cognitive reactivity” to hopelessness (Lau et al., 2004). To date,

CR to hopelessness as a proximal risk to suicidal thinking has only been validated in a sample

of healthy and previously depressed individuals. So far, previous studies have only examined

CR to hopelessness using a laboratory-based, experimental method called the “mood

priming” or “mood induction procedure”. The MIP as its name suggests, is a procedure where

a certain type of mood is induced or primed in an individual in order to examine occurrences

(e.g. behavioural or cognitive) that can only be studied under a certain mood state. While the

MIP has been established as an effective way to alter mood in healthy and previously

depressed individuals (Hepburn et al., 2006; Williams et al., 2005; Williams et al., 2006 &

2007; Hepburn et al., 2009), the extent to which it can mimic the natural mechanism of mood

in real life is subject to speculations. Data from mood priming studies are especially difficult

to interpret if the behavioural or cognitive occurrences under study have a known functional

relationship with real life contexts. For example, suicidal ideation is a cognitive phenomenon

that has been established to interact with the constantly dynamic individual and

circumstantial or contextual factors. This is especially true in the case of the FEP sample as

simple day to day activities (i.e. reading, socialising, & others) can be a struggle due to the

persistence of cognitive impairments following the psychotic episode (Power & McGowan,

2011). Along with the lack of activity, social isolation/alienation, stigmatisation, and

discrimination also characterise the everyday life of this particular clinical group. These

circumstantial factors, together with the appropriate combination of distal and proximal risks,

have been recognised to lead to a suicidal state (Power & Robinson, 2009). It is for this

reason why the data collected via experimental methods such as the MIP become

problematic. The lack of ecological validity in laboratory-based experimental procedures

draws attention to the fact that naturally occurring phenomenon such as the suicidal thinking

39
cannot be effectively measured via artificial means, within the realms of a simulated

environment.

To complement the methodological limitations of the MIP (see studies on chapter 4 &

5), a naturalistic yet highly systematic approach was employed in the current study. The

Experience Sampling Method (Delespaul, 1995; de Vries, 1992), or also known as the

Ecological Momentary Assessment (Stone & Shiffman, 1994), is a “structured diary keeping”

technique that collects data on the individual’s real-time experiences in real-life contexts. In

brief, the ESM entails keeping a record of the momentary changes in thoughts, mood, and

contexts (e.g. places, people, events, & activities) whenever prompted by a signalling device

(usually 10 times a day), over a period of time (e.g. 6 days). Unlike the traditional diary

keeping method, the ESM is not merely a record of events but more importantly, it is a

structured assessment of the individual’s everyday life experiences. Due to the highly

subjective nature of the ESM data and its reliance on self-report measures, the ESM has been

mainly criticised for its validity and reliability. As the purpose of the ESM is to measure how

certain individuals perceive their experiences in everyday life, the validity of the ESM mainly

depends on how correlated variables interact. For example, it is conceptually (and intuitively)

logical to observe greater positive affectivity when the individuals are confronted with events

that are more pleasant in nature. If similar studies replicate such a pattern of results, then the

reliability of the ESM questionnaire is assumed. The validity and reliability of the ESM as a

research tool in psychosis have already been demonstrated in a number of studies (Delespaul

et al. 2002; Delespaul & de Vries, 1987; Hurlburt & Melancon, 1987; Myin-Germeys et al.,

2000; Myin-Germeys et al., 2001; Myin-Germeys et al., 2002; Myin-Germeys et al., 2003;

Lardinois et al., 2003, & many others). Previous studies have shown that individuals with

psychosis are characterised by unstable affectivity (Delespaul & de Vries, 1987; Myin-

Germeys et al., 2000; Palmier-Claus et al., 2011). Myin-Germeys and colleagues (2000) have

40
indicated that patients with chronic schizophrenia are characterised by a less variable and a

less intense positive affectivity and a more variable and a more intense negative affectivity.

The affective variability of patients with chronic schizophrenia has also been found to be

associated with the incidence of recent traumatic life events (Myin-Germeys et al., 2003).

In order to measure the relationship between mood and hopelessness as proposed by

the DAH of suicidal relapse (Lau et al., 2004), the ESM questionnaire in this study was

specifically devised to measure positive and negative affectivity, hopelessness, and the

corresponding contexts of the captured experiences (e.g. places, people, activities, & events

at the time of sampling). The key intention is to examine the ease to which negative

affectivity will trigger hopeless thoughts, or also known as the CR to hopelessness. Williams

and colleagues (2006) first attempted to measure CR to hopelessness using the newly added

subscale in the Leiden Index of Depression Scale (i.e. hopelessness/suicidality subscale),

which was specifically devised to measure the individual’s susceptibility to hopeless thoughts

when in a sad mood. The results of the study showed that individuals who had higher scores

on the LEIDS’ hopelessness/suicidal ideation subscale also admitted to having suicidal

ideations in the past when feeling depressed. Further, the results also indicated that the

LEIDS’ hopelessness/suicidality subscale was predictive of the changes in positive future

fluency, a behavioural outcome that is associated with hopelessness. By employing the

LEIDS’ hopelessness subscale in the present study, mood-linked hopelessness data from the

ESM can confirm if the CR to hopelessness, as measured by the LEIDS, is predictive of the

individual’s vulnerability to hopeless thoughts in everyday life.

As suggested at the beginning of this chapter, until now, there is a lack of

understanding on the suicidal thinking process in individuals with early psychosis. Not

knowing when and how suicidal thinking starts makes it difficult for clinicians to understand

and manage suicidal behaviour in this particular group of individuals who are at a greater risk

41
of killing or hurting themselves. Although the DAH of suicidal relapse (Lau et al., 2004)

provides a potentially valuable framework for the mechanism of suicidal thinking, previous

investigations have only illustrated the application of this framework in a sample of healthy

and previously depressed individuals by employing a laboratory-based mood priming

procedure. The lack of real-life contextual factors in the previous experimental studies on the

DAH, along with the lack of literature on suicidal thinking process within the FEP sample,

prompted the use of the ESM in the present study. By employing the ESM, the present study

will be able to investigate if the assumption of the DAH on hopelessness as a mood-

dependent cognition holds true for the FEP sample with a history of suicidal attempt or

deliberate self-harm. To the best of the author’s knowledge, this is the first study to

investigate the suicidal thinking process in FEP. This is also the first to apply the DAH

framework and the ESM to explore the suicidal thinking mechanism in psychosis.

To ensure clarity of the terminologies used in this chapter, the term momentary

hopelessness is used to refer to the ESM-measured hopelessness, while generalised

hopelessness is used to refer to the global hopelessness as measured by the Beck

Hopelessness Scale (Beck & Steer, 1988). The term cognitive reactivity or CR to

hopelessness, on the other hand, is used to refer to the propensity of the individual to hopeless

thoughts when in a sad mood.

The first aim of this study is to examine the differences between the suicidal history

group (those with a lifetime history of suicidal attempt & deliberate self-harm) and non-

suicidal group (those without a lifetime history of suicidal attempt & deliberate self-harm) by

looking into their level of momentary hopelessness as measured by the hopelessness items in

the ESM diary. Specifically, the present study intends to determine the effect of affectivity

(positive & negative) and daily hassles/minor irritations (activity- & event-related) on the

individual’s momentary hopelessness.

42
 The second aim of the study is to investigate the effects of the minor stresses in

everyday life on the individual’s positive and negative affectivity. Further, the incidence of

recent life events will also be compared between groups.

The final aim of this study is to assess the validity of the LEIDS’ hopelessness or

suicidality subscale as a measure of CR to hopelessness. In particular, the present study

examines if the CR to hopelessness as measured by the LEIDS will be predictive of the

individual’s vulnerability to momentary hopelessness in everyday life when faced with

unpleasant events and challenging activities.

3.1. Hypotheses

Momentary Experiences in Everyday Life as Measured by the ESM

In keeping with the assumptions of the DAH –

1. The suicidal history group will exhibit significantly higher levels of momentary

hopelessness than the non-suicidal group.

Compared to the non-suicidal group, the suicidal history group will -

2. Demonstrate greater momentary hopelessness linked to negative affectivity, and less

momentary hopelessness linked to positive affectivity.

3. Display greater momentary hopelessness when dealing with unpleasant events and

challenging activities

4. Show greater negative affectivity and less positive affectivity when confronted with

unpleasant events and challenging activities.

43
The Validity of LEIDS’ Hopelessness Subscale as a Measure of CR to hopelessness

5. Compared to the non-suicidal group, the suicidal history group will exhibit higher levels of

cognitive reactivity to hopelessness as measured by the LEIDS’ hopelessness subscale.

6. The individual’s cognitive reactivity to hopelessness, as measured by the LEIDS’, will be

predictive of his/her susceptibility to momentary hopelessness when affectivity is negative.

Finally,

7. The individual’s cognitive reactivity to hopelessness, as measured by the LEIDS’, will be

predictive of his/her susceptibility to momentary hopelessness when faced with unpleasant

events and challenging activities.

3.2. Method

3.2.1. Sampling

The inclusion criteria for the study were: (a) able to give fully informed consent as

judged by their care coordinator or other appropriate healthcare professional, (b) fluent in

English, (c) have had their first episode of psychosis and fulfilled the ICD10 criteria for

schizophrenia and schizophrenia related disorder (F20 F21 F22 F23), and (c) have a lifetime

history of deliberate self- harm (DSH) or suicide attempt (as verified from historical risk

assessments and as assessed using the Columbia Suicide History Form or CSHF).

Participants were excluded if: (a) their diagnosis of psychosis was due to an organic disease,

(b) have moderate to severe learning disability, and (c) they were severely suicidal (as

assessed using the InterSept for Suicidal Thinking Scale) at the time of assessment.

Two groups of participants were identified within this sample: those with a lifetime

history of suicidality and those without. The suicidal history group consisted of individuals

44
who had a lifetime history of deliberate self-harm (DSH) or suicidal attempt whilst the non-

suicidal group consisted of those who neither have a history of suicidal attempt nor DSH. In

keeping with the criteria of the Columbia Suicide History Form (Oquendo, Halbestam, &

Mann, 2003), an act was identified as a suicidal attempt if it was carried out with the intent to

die, or the severity of the act itself posed a lethal threat to the individual’s life (e.g. severe

physical damage or prolonged hospitalisation due to the act). Alternatively, an act was

identified as a DSH if the individual deliberately engaged in a “self-poisoning or self-

injurious” behaviour without the intent to die, or the severity of the act itself was by no means

life-threatening (Kreitman, 1977).

3.2.2. Measures

Columbia Suicide History Form (CSHF; Oquendo, Halbestam, & Mann, 2003; Appendix 6)

The CSHF is a semi-structured interview, which accounts the number of lifetime

suicide attempts and instances of the incident (e.g. method, medical lethality, & others). It has

a very good inter-rater reliability correlation of .97 for identifying the history, number, and

fatality of suicide attempts (Oquendo, Bongiovi-Garcia, Galfalvy, et al., 2007). Several

clinical cross-sectional studies that used CSHF to document previous suicide attempts found

that the recorded attempts correlate with more hopelessness, suicidal thinking, and subjective

depression regardless of psychiatric diagnosis (Rush, First, & Blacker, 2008).

The Leiden Index of Depression Sensitivity - Revised version (LEIDS-R; Van der Does &

Williams, 2003; Appendix 7)

The LEIDS is a self-report measure, which was specifically devised to assess for the

individual’s cognitive reactivity to sad mood. In order to measure CR, the conditional

questions in the LEIDS questionnaire require individuals to imagine how they would feel,

45
think, or react if they are feeling sad or low (e.g. “When in a low mood, I take fewer risks”).

Individuals rate their answers on a 5-point Likert scale (1 = not at all & 5 = very strongly).

The revised scale has six subscales including: (1) hopelessness/suicidality, (2)

acceptance/coping, (3) aggression, (4) control/perfectionism, (5) harm avoidance, and (6)

rumination (Van der Does and Williams, 2003). In keeping with the aim of the present study,

only the LEIDS’ hopelessness/suicidality subscale was used in the analyses. The LEIDS’

hopelessness/suicidality subscale has a high internal consistency of .89 Cronbach’s alpha.

Higher scores in this subscale indicate a greater CR to hopeless/suicidal thoughts (Van der

Does & Williams, 2003).

The Calgary Depression Scale for Schizophrenia (CDSS; Addington et al., 1993; Appendix 8)

The CDSS is a 9-item semi-structured interview scale, which was purposely

developed to assess for the severity of depressive symptoms in individuals with

schizophrenia. Compared to the other depression scales (e.g. Hamilton Depression Rating

Scale), the overall CDSS rating has the minimum amount of overlap with the negative

symptoms of schizophrenia (Collins et al., 1996). The superior ability of the CDSS to

discriminate the depressive symptoms from the negative and extrapyramidal symptoms has

made the CDSS a widely used depression scale for schizophrenia amongst many researchers

and clinicians (Collins, Remington, Coulter, & Birkett, 1996; Lancon, Auquiere, Reine, et al.,

1999). The CDSS has a good internal consistency (Cronbach’s alpha=0.79) and a high test-

retest reliability (intraclass correlation coefficient = .90; Addington et al., 1993).

Beck Hopelessness Scale (BHS; Beck & Steer, 1988; Appendix 9)

The BHS is a self-report inventory which was devised to measure three main aspects

of hopelessness: feelings about the future, loss of motivation, and expectations. It consists of

46
20 true/false items, 11 items of which are negatively phrased whilst the remaining nine items

are positively phrased. Overall, the BHS is a well-constructed and validated instrument with

an average reliability coefficient of .92 and test-retest reliability of .69 (Beck & Steer, 1988).

InterSept Scale for Suicidal Thinking (ISST; Lindenmayer, Czobor, Alphs, Nathan, Anand,

Islam, & Chou, 2003; Appendix 10)

The ISST is a 12-item semi-structured interview schedule, which was designed to

assess for suicidal ideation in schizophrenia and schizo-affective disorder. It has a very good

psychometric properties including a high test re-test reliability (intraclass correlation

coefficient = 0.90) and a very good internal consistency (Cronbach’s alpha=0.88;

Lindenmayer, Czobor, Alphs, Nathan, Anand, Islam, & Chou, 2003).

The Experience Sampling Method (ESM; Delespaul, 1995; de Vries, 1992)

The ESM Wristwatch

The ESM is a structured diary keeping procedure that utilises semi-random time

sampling method. The ESM in this study was conducted using a paper and pen diary method

with a programmable digital wristwatch as a signalling device. The ESM wristwatch was set

to emit 10 semi-random signals per day (between 7.30am and 10.30pm) over six consecutive

days. The wristwatches (Timex Ironman USA) used in this study were all password-protected

to ensure that the time sampling schedule were free from any alterations. The author pre-

programmed each watch with randomly allocated (drawn from a hat) time sampling schedule

at least a day before the start of the ESM task. The time sampling schedule or TSS is a list of

pre-determined, semi-random times upon which the watches are set to emit a signal or a

bleep. The researcher adopted the three routinely used TSS’s, which were originally created

by a team of ESM researchers at the University of Maastricht (Appendix 11). The times in

47
the TSS were pre-determined so that it is known exactly how long after the bleep the

participants complete their diary assessments. Diary entry times are vital to assessing

moment-to-moment changes in an individual’s thoughts, moods, events and activities as these

experiences are all transitory in nature. The semi-randomness of the times in the TTS was

equally important to ensure that the captured thoughts, feelings, and events are a part of the

individual’s natural, day to day routine. By setting the bleep times at a schedule that is harder

to predict, the participants were not be able to pre-plan their activities around the ESM

schedule. The semi-random sampling times meant that each time the watch bleeped,

participants were expected to have paused from their activities (only when it was possible and

safe to do so) to complete one ESM questionnaire.

The ESM Diary

The ESM diary consisted of 12 identical sets of self-report ESM questionnaires (1

ESM questionnaire = 2 diary pages) attached together in the form of an A6-size booklet. Each

booklet had two spare sets of questionnaires in case additional assessments were completed

on mistaken bleeps. Each participant received a total of 7 diary booklets, one diary per day

over the 6-day ESM period and an extra diary in case of loss or accidental damage.

All of the ESM items (affectivity, activity & events) used in the present study, with

the exception of the hopelessness items, were adopted from the ESM questionnaire that was

developed by Delespaul (1995). This questionnaire has been validated in a sample of

psychiatric patients in numerous studies (e.g. Delespaul et al., 2002; Lataster et al., 2010;

Myin-Germeys et al., 2000; 2003; 2005; 2009, & many others).

48
Affectivity/Mood

As the word affectivity suggests, items under this section of the ESM diary

questionnaire consisted of words that describe how positive or negative the individual’s mood

is at the time of his/her diary entry. To help create a mindset that was relevant to that specific

moment of time when the diary entry was made, this section was prefaced with “Right now, I

feel…”. Questionnaire items measuring positive affectivity included four positively worded

adjectives (“cheerful”, “content”, “energetic”, & “enthusiastic”) whilst items measuring

negative affectivity (NA) included six negatively worded adjectives (“lonely”, “anxious”,

“insecure”, “low”, “irritated”, & “guilty”). All of the items were rated using the 7-point

Likert scale (1 = not & 7 = very), which is in keeping with the original, standardized ESM

questionnaire used in previous studies (Delespaul et al., 2003; Lataster et al., 2010; Myin-

Germeys et al., 2000, 2001, 2002, 2003, 2004 & many others).

Hopelessness

Given that this is the first ESM study that investigated the concept of hopelessness

vulnerability, items for this construct were initially tested in a pilot study conducted by Luke

Brown in 2008 as a part of his Master of Science dissertation. Brown’s pilot study collected

data from a sample of both healthy and first-episode of psychosis individuals with the aim of:

(a) testing the feasibility of the ESM in a clinical sample in the UK, (b) validate the link

between mood and hopelessness proposed by the Differential Activation Hypothesis (DAH)

for suicidal relapse, and (c) pilot the newly added hopelessness items on the ESM

questionnaire.

Items under the ‘hopelessness’ section of the ESM diary questionnaire was

formulated to mimic the concept of positive future thinking, which is a feature of

hopelessness (see future thinking study in Chapter 5). The questionnaire items were a

49
combination of words and phrases that describe the individual’s feelings and expectations

about the future. This section was divided into two sets. The first set of questions was

prefaced with “Right now, I feel the future is…” followed by items on future expectations

(“bright” & “hopeful”). The second set of questions was prefaced by the phrase “I feel…”

followed by items on feelings about the future (“supported” & “the future has possibilities”).

Similar to the affectivity section, all hopelessness items were rated using the 7-point Likert

scale. In keeping with the term hopelessness, all of the ratings were reverse coded (1=7, 2=6,

3=5, 4=4, 5=3, 6=2, & 7=1) as the questionnaire items were all originally positively worded.

Daily Hassles or Minor Everyday Stresses: Challenging Activities vs. Unpleasant events

Challenging Activities

The activity section of the ESM diary questionnaire was split into two parts. The first

part is the activity type, which asks individuals to describe the activity that they were

involved in prior to being prompted by the watch (“What am I doing just before the bleep

went off?”). The second part is the activity appraisal, which asks the individuals to rate the

degree of difficulty of their activity (“I prefer doing something else”, “I am active”, “This

activity requires a lot of effort”, “This activity is challenging”, & “I’m good at this

activity”). The open-ended question was coded using the ESM coding manual developed by a

group of researchers at the University of Maastricht, whilst the activity appraisal items were

rated using the 7-point Likert scale employed in the earlier sections of the ESM

questionnaire. All of the activity items were adopted from the standardized ESM

questionnaire used in previous studies (Delespaul et al., 2003; Lataster et al., 2010; Myin-

Germeys et al., 2000, 2001, 2002, 2003, 2004 & many others). In keeping with the

hypotheses, only the data from the activity appraisal section were used in the analyses.

50
Unpleasant Events

Similar to the activity section, items under the event section of the ESM diary

questionnaire were also split into two parts. The first part is the event type, which consists of

an open-ended question asking individuals to describe the most significant event that

occurred to them since their previous diary entry (“Since the last bleep, the most important

event that happened to me was…”). The second part is the event appraisal, which asks the

individuals to rate the pleasantness/unpleasantness of the event (“It was…”) using a 7-point

Likert scale (-3 = very unpleasant, 0 = neutral, 3 = very pleasant). Similar to the activity

items, the event items were also adopted from the standardized ESM questionnaire used in

previous studies (Delespaul et al., 2003; Lataster et al., 2010; Myin-Germeys et al., 2000,

2001, 2002, 2003, 2004 & many others). In line with the hypotheses, only the data from the

event appraisal was used. For the sake of clarity, all of the positive ratings were recoded as

“0” whilst all of the negative ratings were re-coded as positive values. Recoding was applied

so that higher ratings would signify more unpleasant events.

Please refer to figure 1 for an illustration of the ESM questionnaire.

3.2.3. Procedure

3.2.3. a. Case Identification

The participants in this study were recruited from the Early Intervention Service (EIS)

in Birmingham from March 2009 to March 2011. The author of this study approached every

care coordinator within EIS to identify service users who conformed to the inclusion criteria.

As established in the earlier section, two groups of participants were identified: (1) suicidal

history group, and (2) non-suicidal group. In order to ensure that all of the participants fulfil

both the inclusion and exclusion criteria, care coordinators were provided with a leaflet that

51
Figure 1. The ESM questionnaire on Affectivity (A), Momentary Hopelessness (B),

Activity (C), and Event (D).

A. Affectivity appraisal items

Right now, I feel... Not Moderately Very
 cheerful 1 2 3 4 5 6 7
 content 1 2 3 4 5 6 7
 lonely 1 2 3 4 5 6 7
 energetic 1 2 3 4 5 6 7
 enthusiastic 1 2 3 4 5 6 7
 anxious 1 2 3 4 5 6 7
 insecure 1 2 3 4 5 6 7
 low 1 2 3 4 5 6 7
 irritated 1 2 3 4 5 6 7
 guilty 1 2 3 4 5 6 7

B. Hopelessness appraisal items

Right now, I feel the future is... Not Moderately Very
 bright 1 2 3 4 5 6 7
 hopeful 1 2 3 4 5 6 7

I feel...
 supported 1 2 3 4 5 6 7
 the future has possibilities 1 2 3 4 5 6 7

C. Activity appraisal items

What am I doing just before the bleep went off?_________________________________
________________________________________________________________________
________________________________________________________________________

Not Moderately Very

 I prefer doing something else 1 2 3 4 5 6 7
 I’m active 1 2 3 4 5 6 7
 I’m good at this activity 1 2 3 4 5 6 7
 This activity requires a lot of effort 1 2 3 4 5 6 7
 This activity is challenging 1 2 3 4 5 6 7

C. Event- appraisal items

Since the last beep, the most important event that happened to me was_______________
_______________________________________________________________________
_______________________________________________________________________

Very unpleasant Neutral Very pleasant

It was... -3 -2 -1 0 1 2 3

52
briefly explained the study and its recruitment criteria. Following referral, participants were

approached over the phone or in person, depending on their preference. During the initial

meeting, the research was presented a three-part study [ESM, Problem-solving (MEPS), &

Future Thinking (FT)], with each study investigating the mechanism of hopeless thinking in

contrasting methodologies – the ecological and experimental approach. In order to

counterbalance the order to which the two sets of methodologies were conducted, the three

studies were split into two sets. Set A consisted of the ecological methodology (Study 1: the

ESM) and set B consisted of the experimental methodology (Studies 2 & 3: MEPS & FT

studies). Those participants who agreed to take part in all the three studies were randomly

allocated to sets AB or BA. Following written consent, the Columbia Suicidal History Form

was conducted in order to determine lifetime histories of suicide attempt or deliberate self-

harm. An audit on the participant’s clinical case notes at EIS was also carried out in order to

check for any historical records of suicidal behaviour.

3.2.3. b. Pilot Feasibility Study of the ESM

Prior to conducting this ESM study, a feasibility pilot research was conducted by

Luke Brown, a Master’s student from the University of Birmingham who was also supervised

by two of the co-authors of this PhD study (Prof Max Birchwood & Dr Chris Jackson). One

of the main aims of the pilot study was to investigate the feasibility of the ESM in a UK-

based clinical sample of FEP patients with a history of suicidal behaviour. The standard 10

bleeps/day sampling frequency was employed for over a period a 6 consecutive days.

Following completion of the data collection, a focus group discussion was held to discuss the

practicality of the ESM. “Irritation” due to the frequency of the prompts/bleeps and the

overall diary keeping task being relatively “demanding/challenging” were amongst the

prominent themes of the discussion. However, on the whole, everyone agreed that the ESM

53
was not an exceedingly difficult task to do because of the very little amount of time it takes to

complete each diary entry. Overall results indicated that the ESM was a valid and feasible

research tool for a UK-based FEP sample.

3.2.3. c. The Experience Sampling Method

The initial session involved completing all of the clinical measures and briefing the

participant about the diary keeping procedure. Prior to consenting to take part in this study,

all of the participants were provided with an information sheet (Appendix 1) that was

reviewed and approved by the National Research Ethics Committee. Pre-ESM clinical

measures consisted of the Calgary Depression Scale for Schizophrenia, Beck’s Hopelessness

Scale, and InterSePT Scale for Suicidal Thinking. Upon completion of all of the measures, a

20-minute briefing session was carried out to explain the details of the ESM procedure. In

keeping with the ESM protocol, participants were only informed of the “general” aim of the

study, which was to examine the nature of their everyday life experiences. The specific aims

of the study were only revealed in the debriefing session in order to avoid potential

measurement biases.

During the briefing session, participants were asked to complete an ESM

questionnaire as a form of practice to confirm that they have understood all of the items in the

questionnaire and the overall ESM procedure. They were advised not to back-fill their diaries

for signals or bleeps that they missed or falsify their diary entry times during the 6-day ESM

period. More importantly, the researcher gave emphasis on the importance of completing

their diary questionnaires immediately after the watch bleeped, without disclosing the time

frame upon which entries must be made. A valid diary entry in this study must be completed

less than 5minutes before but no more than 15 minutes after the bleep. This time frame was

54
adopted from previous ESM studies conducted in a similar clinical sample (Delespaul, 1995;

Myin-Germeys et al., 2005).

During the six-day ESM period, the researcher telephoned the participants on three

separate occasions (end of the 1st, 3rd and 6th day) to help keep up their motivation, and also to

check how they had been getting on so far with the diary keeping task. Throughout the 6-day

ESM period, participants were also free to contact the researcher between the hours of 9am to

5pm (Monday to Sunday). Upon completion of the ESM task, participants met with the

researcher for a 20-minute debriefing session. The purpose of the debriefing was to: (a) check

the completed diaries for any unintelligible entries, (b) count the total number of valid diary

questionnaires (valid data must be ≥ 20 valid entries; Delespaul, 1995), (c) explain the

specific aims of the study, (d) give participants an opportunity to ask questions, and (e)

complete the ESM debriefing questionnaire (Appendix 12). Participants who had 20 or more

valid diary questionnaires received a payment of £30 as an appreciation of their time and

effort. Those who dropped out in the middle of the study or failed to meet the minimum

number of valid entries required were paid according to the amount of time they have spent

doing the study.

3.2.4. Analysis Strategy

The ESM data consisted of two levels: (1) participant level and, (2) day level, which

simply means that there are multiple observations nested within each participant. The nesting

of the ESM data meant that it violated the assumption of independent observations. To satisfy

this assumption, a multi-level linear regression analysis was employed using Stata version

11.0 (Stata Corp, USA). The main variables of interest were analysed using the stata

xtreg command with mle (maximum likelihood estimation) option. The interpretation of

results in multi-level regression is similar to that of a simple linear regression model. Both

55
models assume that the effect of each independent variable is always the same. However,

both also recognise that the effect of one variable may depend on another (interaction effect).

The interpretation of the β coefficients in multi-level linear regression analysis is also

identical to that of standard linear regression, where beta (β) quantifies the degree and

direction of the relationship between the independent (predictor) variables and the dependent

(response) variables. An alpha level of .05 was used for all statistical tests.

The data was analysed in collaboration with Professor Myin-Germeys, one of the

leading ESM researchers in psychosis at the University of Maastricht, the Netherlands.

3.3. Results

3.3.1. Factor Analyses on the ESM Questionnaire Items

As discussed in the earlier section (The ESM Diary), all of the items used in the ESM

questionnaire for this study, with the exception of the hopelessness items, were adopted from

the ESM questionnaire that was developed and validated by Delespaul in 1995 and employed

in many other ESM studies since then (Myin-Germeys et al., 2000, 2001, 2002, 2003, 2004,

& 2005). However, in order to ensure a more robust hypotheses testing, factor analyses were

carried out on the entire questionnaire items that were later used in the multi-level regression

(MLR) analyses. Specifically, a Principal Component Analysis (PCA) with orthogonal

varimax rotations was conducted to determine how strongly each ESM questionnaire item

(variable) was associated with the constructs (factors) that this study intended to measure.

Three independent PCA’s were conducted for each of the variable groups: affect/mood,

hopelessness, and challenging activities. The Kaiser’s eigenvalue-greater-than-one or K1 rule

was employed in determining which factors to retain. In other words, only factors with

eigenvalues greater than one were retained for the MLR analysis.

56
Positive and negative affectivity/mood

The PCA analysis yielded two factors, which accounted for 93.24% of the total

variance in mood. The first factor, which accounted for 55.63% of the variance, was labelled

as positive affectivity (PA) due to the high loadings of variables which altogether strongly

characterise a positive and bright mood. The variables under this construct include: Right

now, I feel “cheerful”, “content”, “energetic”, and “enthusiastic”. The second factor, which

accounted for 37.62% of the variance, was labelled as negative affectivity (NA) due to the

high loadings of variables that characterise a type of mood that is negative and depressed.

The variables under this construct include: Right now, I feel “insecure”, “low”, and

“irritated”. On the other hand, variables such as: Right now, I feel “guilty”, “anxious”, and

“lonely” did not correlate well with the construct of NA and were therefore removed from

the factor and excluded from the multi-regression analysis. Table 2 displays the results of the

factor analysis on affectivity items.

Hopelessness

In agreement with the results of the pilot study (Brown, 2008), the PCA analysis in

the present study yielded only one factor, which accounted for 94.4% of the total variance in

the data. This factor was labelled as hopelessness due to the high loadings of variables, which

when reverse-coded, define pessimistic thoughts about the future. The variables under this

construct include: Right now, I feel the future is “bright”, “hopeful”, and I feel “the future has

possibilities”. Alternatively, the variable I feel “supported” did not correlate well with the

hopelessness construct and was therefore removed from the factor and excluded from the

multi-regression analysis. Table 3 displays the results of the factor analysis on hopelessness

items.

57
 Table 2. Results of the Factor Analysis on the Affective Items

Affective Items Factor 1** Factor 2** Factor 3 Uniqueness

Positive Affectivity Negative Affectivity Guilt/Anxiety
Right now I feel...
Cheerful 0.7251 0.1684
Content 0.6907 0.2195
Energetic 0.8761 0.2092
Enthusiastic 0.8904 0.1779

Right now I feel...

insecure 0.6583 0.4442
low 0.7217 0.2569
irritated 0.5945 0.5236

Guilty -0.0566 0.6525

anxious -0.1781 0.7573
lonely -0.0439 0.6586
**Factors with eigenvalues that are greater than 1. Items under these factors constitute the components of cheerful and dysphoric
variables used in the multi-level regression analyses.

58
Table 3. Results of the Factor Analysis on the Hopelessness Items

Hopelessness Items Factor 1 Uniqueness

Right now, I feel the future is...

Bright 0.932* 0.110

Hopeful 0.930* 0.116

I feel...

Supported 0.550 0.651

the future has possibilities 0.861* 0.243

*Items that constitute the components of the hopelessness variable used in the multi-level regression analyses
(hopefulness items were reverse coded to describe hopelessness).

Daily Hassles: Challenging Activities

The results of the PCA analysis identified one factor, which accounted for 107.3% of

the variance in the data. This factor was labelled as “challenging activities” due to the high

loadings of variables that characterise a difficult activity. The variables under this construct

consist of: “this activity requires a lot of effort” and “this activity is challenging”. The

variables “I prefer doing something else”, “I’m active”, and “I’m good at this activity”, on the

other hand, did not correlate with the challenging activities construct and was therefore

removed from the factor and excluded from the multi-regression analysis. Table 4 displays

results of the factor analysis.

In keeping with K1 rule mentioned earlier, only factors with eigenvalues greater than

one were retained. In order to determine how well the variables that reflect the same

construct/factor yield similar results, the cronbach alpha was calculated for all of the

extracted factors. With the exception of negative affectivity, which has a good internal

reliability, the rest of the factors (positive affectivity, hopelessness, & challenging activities)

59
 have an excellent level of internal consistency. The data on descriptive statistics are

summarised in Table 5.

Table 4. Results of the Factor Analysis on the Activity Appraisal Items

Activity Appraisal Items Factor 1** Factor 2 Uniqueness

Activity-related stress Undefined

Preface: What I am doing just before

the bleep went off?....
I prefer doing something else -0.2325 0.9459

I’m active 0.2878 0.8591

I’m good at this activity 0.1828 0.9664

This activity requires a lot of effort 0.8572 0.2633

This activity is challenging 0.8474 0.2819

**Factor with eigenvalues that are greater than 1. Items under factor 1 constitute the
components of the challenging activities variable used in the multi-level regression analyses

Table 5. Summary of Descriptive Statistics for the Key ESM Factors

Factors No. of items Eigenvalue Alpha M (SD)

Positive Affectivity 4 2.984 0.92 4.17 (1.45)

Negative Affectivity 3 2.018 0.82 1.82 (1.12)

Hopelessness 3 2.475 0.94 3.35 (1.63)

Challenging Activities 2 1.511 0.90 4.74 (3.25)

60
3.3.2. Sample Characteristics

Of the 105 individuals who were approached, only 5 individuals responded with an

outright refusal. Out of the 100 recruited participants, a subsample of 4 (4%) changed their

mind about participating (those who did the pre-ESM assessments but did not start the ESM

diary task), 5 (5%) dropped out within the 6-day ESM assessment period whilst 16 (16%)

failed to meet the minimum number of valid ESM diary entries required (>20; Delespaul,

1995). In total, the final sample consisted of 75 participants (29 females and 46 males) of

which, 35 (46%) were identified to have had a history of suicidal behaviour whilst 40 (54%)

have had no history of suicidal behaviour.

Altogether, the final sample of 75 participants completed a total of 2661 valid ESM

observations (min = 20, max = 58, avg = 35.30), with the suicidal history group significantly

completing more valid ESM observations than the non-suicidal group [mean (SD) = 39.61

(11.1) and 38.33 (11.8) observations, respectively; t (2660) = 2.90, p = 0.003].

3.3.3. Descriptive Statistics and T-tests

Prior to starting the ESM study, all of the participants completed assessments on

suicidal thinking (ISST), hopelessness (BHS), and depression (CDSS). Means, standard

variation (SD), minimum (min) scores, maximum (max) scores, and t-statistics for age and

key clinical symptoms are presented in Table 6.

3.3.4. Hypotheses Testing

To test the hypotheses, independent t-tests on the main variables of interest and a

series of Multilevel Regression (MLR) analyses were carried out using the Stata statistical

software version 11 (Stata Corp, USA). For many social scientists, the MLR is often called as

the multilevel/nested model analysis whilst many statisticians referred to it as mixed model

61
Table 6. Descriptive and T-test Statistics for Age and Key Clinical Symptoms

Measured Variables Groups Min Max Mean SD SE t-statistics

Age Non-suicidal 17 47 23.97 5.40 .86 -.47

Suicidal History 17 37 23.46 4.96 .84
Hopelessness vulnerability -5.95**
(LEIDS – hopelessness subscale) Non-suicidal 1 19 5.97 4.63 0.76
Suicidal History 2 20 12.77 5.04 0.85

Hopelessness -4.34**
(Beck Hopelessness Scale) Non-suicidal 0 19 4.64 4.65 0.74
Suicidal History 1 19 9.71 5.34 0.90

Suicidal thinking -3.10**

(InterSept for Suicidal Thinking) Non-suicidal 0 12 0.56 2.20 0.35
Suicidal History 0 15 3.31 4.82 0.82

Depression -3.63**
(Calgary Depression Scale for Non-suicidal 0 14 1.82 3.06 0.49
Schizophrenia) Suicidal History 0 15 5.34 4.96 0.84
*p<0.05, **p<0.001

62
analysis. For most people in education, the MLR is referred to as the Hierarchical Linear

Model. As discussed earlier, the MLR is best suited to cross-sectional time-series data such as

that of the ESM because it satisfies the assumption of independent observation, which is

violated by the nesting of multiple ESM observations within the subjects or participants.

In Stata, multilevel (xt) regression (reg) was carried out using the “xtreg” command.

The basic syntax for the “xtreg” command using the maximum likelihood estimation (mle)

model is: xtreg y x1, i (varname) mle. Similar to the basic regression formula, the “y”

(dependent/outcome variable) is followed by the “x” (independent/predictor variable).

Following the principle of multiple regression, the number of independent or predictor

variables depends on the variables of interest in the hypothesis [e.g. xtreg y x1 x2 x3, i

(varname) mle]. The “i” (individual) is the identification variable where the multiple

observations are nested, which is the participant level (variable name: subj_no) in the case of

our analyses. The “mle” option (maximum likelihood estimation), as the name suggests, fully

maximizes the likelihood of the random effects model. The random effects model assumes

that the differences across cases are random and not correlated with the predictor variables.

An alpha level of .05 was used for all statistical tests.

Momentary Experiences in Everyday Life as Measured by the ESM

In keeping with the assumptions of the DAH –

1. The suicidal history group will exhibit significantly higher levels of momentary

hopelessness than the non-suicidal group.

To test this hypothesis, an independent t-test was conducted to compare the

magnitude of momentary hopelessness that was experienced by each group (suicidal history

group vs. non-suicidal group) on a day to day basis. As predicted, the suicidal history group

63
(M = 3.56, SD = 1.37) showed significantly higher momentary hopelessness mean score than

the non-suicidal [M = 3.16, SD = 1.86, t (2319.37) = 6.17, p <.001].

2. Compared to the non-suicidal group, the suicidal history group will demonstrate greater

momentary hopelessness linked to negative affectivity, and less momentary hopelessness

linked to positive affectivity.

Prior to testing this hypothesis, preliminary analyses were conducted to ensure that

the Experience Sampling Method was able to detect the fluctuations in momentary

hopelessness linked to both the negative affectivity and positive affectivity as suggested by

the DAH for suicidal relapse. To do this, NA and PA (“x” or predictor variables) were

separately fitted into the model predicting momentary hopelessness (“y” or outcome

variable). To test if momentary hopelessness is linked to NA, multilevel regression was

carried out using the syntax:

xtreg y(momentary hopelessness) x(NA), i(subj_no) mle

To test if momentary hopelessness is linked to PA, the same form of syntax was employed

but using PA as the predictor variable:

xtreg y(momentary hopelessness) x(PA), i(subj_no) mle

The results showed that both NA and PA significantly predicted momentary hopelessness

(statistics are shown in Table 7).

To test the hypothesis, analyses were performed in two stages: First, the group

(suicidal history group & non-suicidal group) variable was added as an independent predictor

in the model predicting momentary hopelessness. The interaction term between group and

affectivity (NA & PA) was also included to check if NA and PA remained as significant

predictors. The syntax employed to carry out this initial stage of the analysis was:

64
 xtreg y (momentary hopelessness) x1(NA/PA) x2(group) x3(NA/PA*group),

i(subj_no) mle

The results showed a significant main effect of NA and PA, and also interaction

effects for both NA x group and PA x group combinations. Second, given the significant

results for both NA and PA from the initial analyses, stratified analyses were conducted to

determine the differences between each group.

xtreg y (momentary hopelessness) x1(NA/PA) if group = non-suicidal group,

i(subj_no) mle

xtreg y (momentary hopelessness) x1(NA/PA) if group = suicidal history group,

i(subj_no) mle

The results revealed that the suicidal history group had a greater increase in

momentary hopelessness linked to NA than the non-suicidal group. The suicidal history

group also had the greater reduction in momentary hopelessness in relation to PA compared

to the non-suicidal group. Table 7 displays summary of results.

To control for the possible effects of the key clinical symptoms, scores from CDSS

(depression), BHS (generalised hopelessness), and ISST (suicidal thinking) were separately

added as covariates. Both NA and PA remained as significant predictors of momentary

hopelessness after controlling for the previously identified key clinical symptoms.

In summary, the results were found to be consistent with the hypothesis as the suicidal

history group exhibit greater momentary hopelessness linked to negative affectivity and

reduced momentary hopelessness linked to positive affectivity compared to the non-suicidal

group.

65
Table 7. Summary of Multiple Regression Analysis on Affectivity and Suicidality as Predictors of Momentary Hopelessness (N = 75)

Predictor Variables χ2 β SE p-value Lower CI Upper CI

Negative affectivity (NA) .000 0.47 0.02 0.000 0.43 0.51
NA*Group .000
NA 0.34 0.04 0.000 0.25 0.42
Group -0.14 0.28 0.622 -0.69 0.41
NA x Group 0.18 0.05 0.000 0.08 0.28
NA if group = non-suicidal .000 0.34 0.42 0.000 0.25 0.42
NA if group = suicidal .000 0.52 0.03 0.000 0.46 0.57

Positive Affectivity (PA) .000 -0.48 0.02 0.000 -0.52 -0.45

PA*Group .000
PA -0.42 0.02 0.000 -0.47 -0.37
Group 0.70 0.25 0.005 0.21 1.20
PA x Group -0.12 0.03 0.000 -0.18 -0.05
PA if group = non-suicidal .000 -0.42 0.02 0.000 -0.46 -0.37
PA if group = suicidal .000 -0.54 0.02 0.000 -0.58 -0.50
χ2 = F-statistic of the regression model

66
3. Compared to the non-suicidal group, the suicidal history group will display greater

momentary hopelessness when dealing with unpleasant events and challenging activities.

Prior to testing this prediction, initial analyses were conducted to verify if unpleasant

events and challenging activities in everyday life are linked to momentary hopelessness. To

do this, unpleasant events and challenging activities (“x” or predictor variables) were

separately fitted into the model predicting momentary hopelessness (“y” or outcome

variable). To test if momentary hopelessness is linked to unpleasant events, multilevel

regression was carried out using the syntax:

xtreg y(momentary hopelessness) x(unpleasant events), i(subj_no) mle

To test if momentary hopelessness is linked to challenging activities, the same form of syntax

was employed but using challenging activities as the predictor variable:

xtreg y(momentary hopelessness) x(challenging activities), i(subj_no) mle

Results from this initial analyses indicated that unpleasant events but not challenging

activities significantly predicted momentary hopelessness (statistics are shown in Table 8).

Similar to the analyses in hypothesis 2, a two-fold analysis was carried out to test the

hypothesis.

Unpleasant events

For first part of the analysis, the group (suicidal & non-suicidal) variable was added

as an independent predictor in the model predicting momentary hopelessness. The interaction

term between group and unpleasant events was also included to determine whether

unpleasant events remain as a significant predictor. The syntax employed to carry out this

initial stage of the analysis is:

xtreg y (momentary hopelessness) x1(unpleasant events) x2(group)

x3(unpleasant events*group), i(subj_no) mle

67
A significant main effect of unpleasant events and an interaction event x group was found. In

the second part of the analysis, a stratified analysis was carried out to identify which group

had greater increase in momentary hopelessness in relation to the unpleasant events.

xtreg y (momentary hopelessness) x1(unpleasant events) if group = non-suicidal

group, i(subj_no) mle

xtreg y (momentary hopelessness) x1(unpleasant events) if group = suicidal

history group, i(subj_no) mle

As predicted, a greater increase in momentary hopelessness was found in the suicidal history

group compared to the non-suicidal group when confronted with unpleasant events. The

results remain unchanged after the key clinical symptoms, scores from CDSS (depression),

BHS (generalised hopelessness), and ISST (suicidal thinking) were separately added as

covariates. Table 8 displays summary of results.

Challenging activities

Whereas the results of the preliminary analyses earlier showed that challenging

activities did not significantly predict momentary hopelessness on the whole, adding the

group variable in the regression model might yield different results. Following the two-step

analysis conducted previously, first, the group (suicidal & non-suicidal) variable was added

in the model predicting momentary hopelessness. Similarly, the interaction term between

group and challenging activities was also included to determine whether challenging

activities remain as a significant predictor. The syntax employed to carry out this initial stage

of the analysis is:

xtreg y (momentary hopelessness) x1(challenging activities) x2(group)

x3(challenging activities*group), i(subj_no) mle

68
No significant main effect of challenging activities and activity x group interaction effect

were found. No further analysis was made as challenging activities did not significantly

predict momentary hopelessness both on the whole and even after the group variable was

added in the model. Table 8 displays summary of results.

In summary, it was found that unpleasant events but not challenging activities

predicted momentary hopelessness. Stratified analyses for each group showed that when

faced with unpleasant events, the suicidal history group had a greater increase in momentary

hopelessness compared to the non-suicidal group. The results hold true after controlling for

depression (CDSS), generalised hopelessness (BHS), and suicidal thinking (ISST).

4. Compared to the non-suicidal group, the suicidal history group will show greater negative

affectivity and less positive affectivity when confronted with unpleasant events and

challenging activities.

Similar to item 3, initial analyses were conducted to verify if the daily hassles

(unpleasant events & challenging activities) are linked to mood or affectivity (NA & PA)

prior to testing the hypothesis. To do this, unpleasant events and challenging activities (“x”

or predictor variables) were separately fitted into the model predicting positive and negative

affectivity (“y” or outcome variable). To test if affectivity (NA/PA) is linked to daily hassles

(challenging activities/unpleasant events), multilevel regression was carried out using the

syntax:

xtreg y(NA/PA) x(unpleasant events/challenging activities), i(subj_no) mle

The results revealed that unpleasant events were a significant predictor for both negative

affectivity and positive affectivity. Unlike the unpleasant events, challenging activities

69
Table 8. Summary of Multiple Regression Analysis on Daily Hassles (unpleasant events & challenging activities) and Suicidality as

Predictors of Momentary Hopelessness (N = 75)

Predictor Variables χ2 β SE p-value Lower CI Upper CI

Unpleasant events (UE) .000 0.48 0.03 0.000 0.42 0.53
UE*Group .000
UE 0.31 0.04 0.000 0.22 0.40
Group 0.18 0.28 0.516 -0.37 0.74
UE x Group 0.24 0.05 0.000 0.14 0.35
UE if group = non-suicidal .000 0.31 0.04 0.000 0.23 0.40
UE if group = suicidal .000 0.56 0.03 0.000 0.49 0.62

Challenging activities (CA) .228 0.01 0.01 0.228 -0.01 0.03

CA*Group .052
CA -0.01 0.01 0.389 -0.04 0.01
Group 0.17 0.31 0.596 -0.45 0.78
CA x Group 0.04 0.02 0.073 0.01 0.07
CA if group = non-suicidala n/a n/a n/a n/a n/a n/a
CA if group = suicidala n/a n/a n/a n/a n/a n/a
a
χ2 = F-statistic of the regression model. = Stratified Analyses were not carried out because main effect of ARS was not significant in the
analysis using 2nd model on the table.

70
significantly predicted NA but not PA (statistics are shown in Table 9).

To test the hypothesis, a two-fold analysis was carried out separately for each of the

daily hassles:

Unpleasant events

First, the group (suicidal history group & non-suicidal group) variable was added as

an independent predictor to the models predicting NA and PA. In order to find out if

unpleasant events will remain as significant predictor of mood, an interaction term between

affectivity and unpleasant events was also added in the model.

xtreg y (NA/PA) x1(unpleasant events) x2(group) x3(unpleasant events *group),

i(subj_no) mle

The results of these further tests showed significant main effects of unpleasant events in

predicting both the NA and PA. It also revealed significant event x group interaction effect in

both models predicting NA and PA.

Second, independent stratified analyses for models predicting NA and PA were

carried out to determine which group was more emotionally sensitive to unpleasant events.

xtreg y (NA/PA) x1(unpleasant events) if group = non-suicidal group/suicidal

history group, i(subj_no) mle

As expected, stratified analysis in the model predicting NA revealed that the suicidal

had a significantly greater increase in NA when confronted with unpleasant events compared

to the non-suicidal group. On the other hand, stratified analysis in the model predicting PA

also confirmed the hypothesis with the suicidal history group showing significantly greater

decrease in PA when confronted with unpleasant events compared to the non-suicidal group.

The results remained unchanged after depression (CDSS), generalised hopelessness (BHS),

and suicidal thinking (ISST) were entered as covariates (statistics are shown in Table 9).

71
Challenging activities

Although the results of the preliminary analyses earlier indicated that challenging

activities were significant predictors of NA, adding the group variable in the regression

model might reveal different results. Following the two-step analyses conducted in the

previous section; first, the group (suicidal & non-suicidal) variable was added an independent

predictor to the models predicting NA and PA. An interaction term between group and

challenging activities was also added in the model in order to validate whether challenging

activities will remain as a significant predictor for NA and PA.

xtreg y (NA/PA) x1(challenging activities) x2(group) x3(challenging activities

*group), i(subj_no) mle

The results of this analysis indicated that there is a significant main effect of challenging

activities and an activity x group interaction effect in both models predicting NA and PA.

Table 9 displays summary of results.

Second, individual stratified analyses were carried out for each models predicting NA

and PA to determine which group is more emotionally sensitive to challenging activities.

xtreg y (NA/PA) x1(challenging activities) if group = non-suicidal group/suicidal

history group, i(subj_no) mle

As expected, the results of stratified analyses in the model predicting NA revealed that the

suicidal history group had a significantly greater increase in NA when faced with challenging

activities compared to the non-suicidal group. These results remained unchanged after

depression (CDSS), generalised hopelessness (BHS), and suicidal thinking (ISST) were

entered as covariates. The results of the stratified analyses in the model predicting PA, on the

other hand, were unable to discriminate the differences between each group. Challenging

activities as a significant predictor of PA was only found in the non-suicidal group but not the

suicidal history group. Such confounding outcome may be due to the fact that challenging

72
activities did not significantly predict PA in the preliminary analysis. Table 9 displays

summary of results.

In summary, the outcome was in keeping with the hypothesis as the suicidal history

group exhibited greater NA and less PA when faced with unpleasant events compared to the

non-suicidal group. In contrast, when faced with challenging activities, the suicidal history

group only exhibited greater NA than the non-suicidal group. Stratified analysis on PA

between groups produced incompatible results, thus making it unfeasible to discriminate the

differences between the suicidal history group and the non-suicidal group. This may be due to

challenging activities significantly predicting NA, but not PA in the preliminary analysis.

Similar results were found after controlling for depression (CDSS), generalised hopelessness

(BHS), and suicidal thinking (ISST).

The Validity of LEIDS’ Hopelessness Subscale as a Measure of CR to hopelessness

5. Compared to the non-suicidal group, the suicidal history group will exhibit higher levels of

cognitive reactivity to hopelessness as measured by the LEIDS’ hopelessness subscale.

In line with our hypothesis, the suicidal history group (M = 12.7, SD = 5.0) scored

significantly higher on the Leiden Index of Depression Scale or LEIDS’ hopelessness

subscale than the non-suicidal group (M = 6.1, SD = 4.6), t(71) = 5.90, p < .001. Similar

results were found after controlling for depression (CDSS), generalised hopelessness (BHS),

and suicidal thinking (ISST).

73
Table 9. Summary of Multiple Regression Analysis on Daily Hassles (unpleasant events & challenging activities) and Suicidality as
Predictors of Changes in Affectivity (N = 75)
Response Variable Predictor Variables χ2 β SE p-value Lower CI Upper CI
Negative Unpleasant events (UE) .000 0.56 0.02 0.000 0.51 0.60
Affectivity UE *Group .000
UE 0.22 0.04 0.000 0.14 0.29
Group 0.19 0.17 0.274 -0.15 0.52
UE x Group 0.51 0.05 0.000 0.42 0.60
UE if group = non-suicidal .000 0.21 0.03 0.000 0.15 0.27
UE if group = suicidal .000 0.72 0.03 0.000 0.66 0.79
Challenging Activities (CA) .000 0.05 0.01 0.000 0.66 0.79
CA*Group .041
CA 0.02 0.01 0.023 0.00 0.05
Group 0.21 0.21 0.330 -0.21 0.63
CA x Group 0.05 0.02 0.002 0.02 0.08
CA if group = non-suicidal .002 0.02 0.01 0.002 0.01 0.04
CA if group = suicidal .000 0.07 0.01 0.000 0.05 0.10

Positive Unpleasant events (UE) .000 -0.53 0.03 0.000 -0.59 -0.47
Affectivity UE*Group .000
UE -0.40 0.06 0.000 -0.51 -0.30
Group -0.17 0.24 0.478 -0.64 0.30
UE x Group -0.19 0.07 0.005 -0.32 -0.06
UE if group = non-suicidal .000 -0.40 0.04 0.000 -0.51 -0.30
UE if group = suicidal .000 -0.59 0.02 0.000 -0.67 -0.51
Challenging Activities (CA) .342 0.01 0.01 0.342 -0.01 0.03
CA*Group .052
CA 0.03 0.02 0.034 -0.70 0.40
Group -0.14 0.28 0.596 -0.70 0.40
CA x Group -0.04 0.02 0.044 -0.12 -0.00
CA if group = non-suicidal .022 0.03 0.01 0.022 0.00 0.06
CA if group = suicidal .536 -0.01 0.02 0.536 -0.04 0.02
χ2 = F-statistic of the regression model

74
6. The individual’s cognitive reactivity to hopelessness, as measured by the LEIDS, will be

predictive of his/her susceptibility to momentary hopelessness when affectivity is negative.

To test the hypothesis, a two-step analysis similar to hypothesis 4 was carried out.

First, NA, scores on the LEIDS’ hopelessness subscale, and an interaction term between these

two were added as independent predictors to the model predicting momentary hopelessness.

xtreg y (momentary hopelessness) x1(LEIDS) x2(NA) x3(LEIDS*NA),

i(subj_no) mle

As expected, the LEIDS predicted momentary hopelessness when affectivity is negative.

Second, the LEIDS variable was dichotomised into upper and lower halves to identify if high

and low scorers will differentially predict momentary hopelessness when affectivity is

negative. Separate analyses were then carried out for the upper half and the lower half.

xtreg y (momentary hopelessness) x(NA) if dichotomised LEIDS = upper

half/lower half, i(subj_no) mle

High LEIDS scorers or those with higher CR to hopelessness had a greater increase in

momentary hopelessness when affectivity is negative compared to the low LEIDS scorers or

those with lower CR to hopelessness. Table 10 displays the summary of results.

In summary, the outcome was in keeping with the hypothesis as those with higher CR

to hopelessness as measured by the LEIDS’ hopelessness subscale exhibited a greater

increase in momentary hopelessness when affectivity is negative compared to the low scorers

or those with lower CR to hopelessness. Similar results were found after controlling for

depression (CDSS), generalised hopelessness (BHS), and suicidal thinking (ISST).

75
Table 10. Summary of Multiple Regression Analysis on LEIDS Score on Hopelessness Subscale and Negative Affectivity as a Predictor of

Momentary Hopelessness (N = 75)

Predictor Variables χ2 β SE p-value Lower CI Upper CI

LEIDS-hopelessness subscale * NA .000 0.03 0.00 0.000 0.03 0.04

NA if LEIDS score = lower half .000 0.40 0.03 0.000 0.04 0.47
NA if LEIDS score = upper half .000 0.51 0.03 0.000 0.45 0.57
χ2 = F-statistic of the regression model

Table 11. Summary of Multiple Regression Analysis on LEIDS score on Hopelessness Subscale and Daily Hassles as a Predictor of

Momentary Hopelessness (N = 75)

Predictor Variables χ2 β SE p-value Lower CI Upper CI

LEIDS-hopelessness subscale * unpleasant events .000 0.02 0.00 0.000 0.02 0.02
Unpleasant events if LEIDS score = lower half .000 0.11 0.02 0.000 0.07 0.07
Unpleasant events if LEIDS score = upper half .000 0.31 0.02 0.000 0.27 0.34

LEIDS-hopelessness subscale *challenging activities .626 0.01 0.02 0.626 -0.06 0.03
Challenging activities if LEIDS score = lower half .297 0.00 0.01 0.297 -0.02 0.02
Challenging activities if LEIDS score = upper half .998 0.01 0.01 0.998 -0.01 0.04
χ2 = F-statistic of the regression model

76
Finally,

7. The individual’s cognitive reactivity to hopelessness, as measured by the LEIDS, will be

predictive of his/her susceptibility to momentary hopelessness when faced with unpleasant

events and challenging activities.

To test the hypothesis, a two-step analysis similar to hypothesis 6 was carried out

separately for each type of daily hassles (unpleasant events & challenging activities).

Unpleasant events

First, unpleasant events, scores on LEIDS’ hopelessness subscale, and the interaction

term between these two variables were added as independent predictors to the model

predicting hopelessness.

xi: xtreg y (momentary hopelessness) x1(unpleasant events) x2(LEIDS)

x3(unpleasant events*LEIDS), i(subj_no) mle

A significant LEIDS x unpleasant events interaction effect was found, which indicated that

scores on LEIDS’ hopelessness subscale predicted momentary hopelessness when dealing

with unpleasant events.

Second, the LEIDS variable was dichotomised into upper and lower halves to identify

if high and low scorers will differentially predict momentary hopelessness when faced with

unpleasant events. Separate analyses were then carried out for the upper half and the lower

half.

xi: xtreg y (momentary hopelessness) x(unpleasant events) if dichotomised LEIDS =

upper half/lower half, i(subj_no) mle

In keeping with the hypothesis, high LEIDS’ hopelessness subscale scorers or those with

higher CR to hopelessness had a greater increase in momentary hopelessness when faced with

77
unpleasant events compared to the low scorers or those with lower CR to hopelessness. Table

11 displays the summary of results (please see page 76).

Challenging activities

Following the two-step analyses-- first, challenging activities, scores on LEIDS’

hopelessness subscale, and the interaction term between these two variables were added as

independent predictors to the model predicting hopelessness.

xi: xtreg y (momentary hopelessness) x1(challenging activities) x2(LEIDS)

x3(challenging activities *LEIDS), i(subj_no) mle

Contrary to the hypothesis, CR to hopelessness as measured by the LEIDS’ hopelessness

subscale did not predict momentary hopelessness during challenging activities. Due to this

non-significant result, no further analyses were conducted. Table 11 displays the summary of

results (please see page 76).

In summary, the individual’s CR to hopelessness as measured by the LEIDS’

hopelessness subscale was found to be predictive of his/her susceptibility to momentary

hopelessness when faced with unpleasant events but not with challenging activities. This

pattern of results was unaffected after controlling for depression (CDSS), suicidal thinking

(ISST), and generalised hopelessness (BHS).

3.4. Discussion

This study set out to test the Differential Activation Hypothesis of suicidal relapse in

early psychosis through the use of the Experience Sampling Method, a systematised diary

keeping method, which semi-randomly samples affective, cognitive, and behavioural data as

they occur in an individual’s everyday environment. Specifically, this study aimed to explore

the link between momentary hopelessness and affectivity (positive vs. negative) in

78
individuals with a lifetime history of suicidal behaviour vs. without. Although the compliance

rate of 59% (number of valid observations per participant = 35.3) was slightly lower than the

reported rate in schizophrenia (66%; Oorschot et al., 2009), it is important to note that the

sample in this present study were still at a difficult stage of recovery following the initial

episode of psychosis (Harrison & Fowler, 2004).

Given that this is the first study to have explored the occurrence, amplitude, and

fluctuation of hopelessness in everyday life, findings from laboratory-based studies that

investigated the link between hopelessness and suicidal behaviour will only be comparable to

a certain extent. Unlike the mood-primed data on hopelessness from previous laboratory-

based cross-sectional studies, the ESM data on momentary hopelessness were repeatedly

sampled from the individual’s natural environment for a prolonged period of time. For this

reason, only indirect comparisons were made in some parts of the discussions.

Consistent with the hypothesis, the suicidal history group exhibited greater amplitude

of momentary hopelessness on a day to day basis compared to the non-suicidal group. This

finding was consistent with other studies, which indicated a strong link between hopelessness

and suicidality in the FEP sample (Cohen et al., 1994; Klonksy et al., 2012; Nordentoft et al.,

2002; Robinson et al., 2009).

Also in keeping with the hypothesis, the suicidal history group also showed

significantly larger increase in momentary hopelessness linked to negative affectivity and

larger decrease in momentary hopelessness linked to positive affectivity. The pattern of

results also indicated that momentary hopelessness was more strongly linked with NA than

PA, which was in keeping with the DAH for suicidal relapse (Lau et al., 2004). Closer

inspection of the changes in momentary hopelessness linked to affectivity revealed that the

amount of increase in momentary hopelessness linked to NA in the suicidal history group (β

= 50) was 32% greater than the non-suicidal group (β=38). In contrast, the difference in the

79
amount of reduction in momentary hopelessness linked to PA in the suicidal history group (β

= 59) was 28% greater than the non-suicidal group (β = 46). This pattern of results was in

agreement with the findings on Hepburn et al.’s mood-priming study (2006) which indicated

that only the negative, but not the positive mood induction, prompted a change in the

individual’s positive future fluency (a behavioural feature of hopelessness).

Unexpectedly, the data on daily hassles (unpleasant events & challenging activities)

as a predictor of momentary hopelessness and affectivity (positive & negative) produced a

mixed outcome. Contrary to hypothesis, the suicidal history group only exhibited greater

increase in momentary hopelessness when confronted with unpleasant events but not with

challenging activities. One possible explanation is that a good number of the participants

were unemployed and had very limited range of social activities on a day to day basis. The

data from the ESM diary revealed that most commonly reported activities included “watching

telly or listening to music” and “sleeping or napping”, which accounts to 28% and 12% of the

total reported activity respectively. Given that the ESM items on the activity appraisal

section were devised to measure the subjective difficulty of the task, the nature of the

activities that most of the participants engaged themselves in seemed to be quite relaxing, less

varied, and less difficult as opposed to being challenging and complicated. In other words,

the reported activities were simply not stressful enough to trigger significantly different

amplitudes of hopelessness between the suicidal history group and non-suicidal group. The

most commonly reported events, on the other hand, included experiences or happenings that

were more personal to the participants such as face to face conversations, telephone calls, or

visits by family members or friends (31%). Given that the ESM item on event appraisal was

devised to measure the unpleasantness or pleasantness of the event, it is possible that

displeasing personal events were likely to be perceived as more unpleasant by the individual.

The significantly higher increase in momentary hopelessness in the suicidal history group

80
suggests that those with a history of suicidal behaviour were more prone to the activation of

attenuated hopeless cognitions when faced with unpleasant events in everyday life compared

to those without any history of suicidal behaviour.

As expected, further analyses revealed that unpleasant events not only impacts on

momentary hopelessness, but also on the positive and negative affectivity of the individual.

The results have shown that the suicidal history group had a significantly greater NA and less

PA than the non-suicidal group when confronted with unpleasant events. However, when

confronted with challenging activities, the suicidal history group only exhibited greater NA

than the non-suicidal group while the groups did not differ at all on their PA. The pattern of

results illustrating the significant impact of unpleasant events on the individual’s affectivity

and momentary hopelessness was in keeping with the assumption of the DAH for suicidal

relapse (Lau et al., 2004). Recalling the assumptions of the DAH for suicidal relapse,

affectivity/mood and hopelessness are strongly associated to each other such that the previous

determines the mechanism of the latter (Lau et al., 2004). It was therefore unsurprising that

the changes in affectivity (greater NA & less PA) and momentary hopelessness in the suicidal

history group were more distinct than the non-suicidal group. Interestingly, the affective

reactivity to unpleasant events in the suicidal history group was found to be stronger in NA

than PA. When faced with unpleasant events, the suicidal history group showed a bigger

increase in NA than the non-suicidal group. Specifically, the results have indicated that the

increase in NA in the suicidal history group was 70.83% greater than the non-suicidal group.

In contrast to this, the decrease in PA in the suicidal history group was only 32.20% more

than the non-suicidal group. In effect, this distinctly stronger link between unpleasant events

and NA in the suicidal history group supports the validity of the DAH framework in the

context of everyday life. It also Unlike the artificial setting of laboratory-based studies, the

real-life context of the ESM studies allows contextual factors such as the daily hassles (e.g.

81
unpleasant events) to influence the natural mechanism of affectivity. The results of the

present study illustrating the role of daily hassles (unpleasant events) as a predictor of NA

and momentary hopelessness extends the application of the DAH framework in the everyday

life of the FEP sample. In particular, the link between affective variability (e.g. increase in

NA or decrease in PA) and unpleasant events have important implications for the concept of

cognitive reactivity to hopeless and suicidal thoughts as proposed by the DAH of suicidal

relapse (DAH; Lau et al., 2004). First, the distinctly greater sensitivity to unpleasant events of

FEP individuals from the suicidal-history group (as illustrated by the increase in their NA)

than those from the non-suicidal group suggest that FEP individuals with a history of suicidal

behaviour are at a greater risk for future suicidal behaviour. On a day to day basis, it simply

means that unpleasant events are more likely to elicit negative affective responses amongst

individuals with histories of suicidal behaviour. These negative affective responses then

reactivate a network of maladaptive thinking patterns which, given the right intensity and

context, could potentially trigger reactivation of hopeless or suicidal thoughts. On the whole,

the pattern of results suggests that the occurrence of unpleasant events in the everyday lives

of FEP individuals with a history of suicidal behaviour can therefore act as a precursor to a

more negative mood/affect, which according to the DAH of suicidal relapse can potentially

trigger the recurrence of hopeless/suicidal cognitions.

Second, the evidence suggesting the mediating effect of psychosis as a traumatic life

experience on the affective responses to minor stresses in everyday life (e.g. unpleasant

events), supports the previous findings that FEP individuals were more vulnerable to suicidal

relapses. Dealing with the traumatic experience of psychosis and adjusting to changes

brought by the psychotic illness can be difficult for many individuals. Having to confront one

or both of these challenges at the same time is enough to render this particular group of

individuals more vulnerable to the effects of minor stresses in everyday life. As evidenced by

82
the pattern of results discussed previously, such vulnerability is even intensified when the

individual has previously felt hopeless/suicidal. Recalling the assumption of the DAH, the

individual’s vulnerability to suicidal relapse is determined by how strong the link is between

negative affect (e.g. depressed mood) and hopeless/suicidal thoughts. Given the enhanced

affective sensitivity to unpleasant events as moderated by the experience of psychosis as a

traumatic life event, and the susceptibility of FEP individuals with a history of suicidal

behaviour to hopeless/suicidal thoughts when affect is intensely negative (e.g. depressed), the

occurrence of severely unpleasant events to the lives of this group of individuals is almost

tantamount to the reactivation of hopeless/suicidal thoughts.

The pattern of results on challenging activities as a predictor of affectivity, on the

other hand, was more difficult to explain. The results from an earlier analysis indicated that

challenging activities did not significantly predict momentary hopelessness. However, when

challenging activities were tested as a predictor of affectivity, it predicted greater NA in the

suicidal history group than the non-suicidal group. It is possible that due to the lack of

complexity in the daily activities of the participants in the present study, the impact may have

simply been too subtle to reactivate hopeless thoughts, but enough to alter negative

affectivity. This further substantiates the concept of “differential activation” as the effects of

the daily hassles can vary greatly from one event/activity to another. Similarly, this may also

be the reason why challenging activities did not significantly predict greater reduction in PA

in the suicidal history group as hypothesised. It was noted earlier that there was a general lack

of complexity and variety in the day to day activities of the participants in the present study.

It is therefore possible that the activities that were particularly challenging were simply not

strong enough to predict differential reduction in PA between the two groups. It is plausible

that a similar pattern of relationship found between unpleasant events and affectivity also

83
applies to the relationship between challenging activities and affectivity, which suggests that

challenging activities might also have a stronger link with NA than PA.

As hypothesised, the suicidal history group scored higher in the LEIDS’ hopelessness

subscale than the non-suicidal group. This outcome is in agreement with the results by

Williams and colleagues (2008), who found that those who had suicidal thoughts when

feeling depressed in the past scored higher on the LEIDS hopelessness subscale. In keeping

with the assumption of the DAH of suicidal relapse (Lau et al., 2004), the cognitive reactivity

(CR) to hopelessness as measured by the LEIDS’ hopelessness subscale was predictive of the

individual’s susceptibility to momentary hopelessness when affectivity is negative. Further

analyses specifically showed that those who scored higher in the LEIDS’ hopelessness

subscale predicted greater increase in momentary hopelessness when affectivity is negative

compared to those who scored lower. This confirms the results found in previous mood-

priming studies (Hepburn et al., 2006; Williams et al., 2005, 2006, & 2007).

Finally, the data for the final hypothesis of this study revealed dissimilar results. It

was found that compared to those who have lower CR to hopelessness, those who have

higher levels of CR to hopelessness as measured by the LEIDS’ hopelessness subscale

exhibited greater increase in momentary hopelessness when confronted with unpleasant

events but not with challenging activities. These findings replicate the data on daily hassles as

a predictor of momentary hopelessness. Earlier it was found that when faced with unpleasant

events, the suicidal history group experienced greater increase in momentary hopelessness

than the non-suicidal group. Similarly, when faced with the same unpleasant events, those

who scored higher in LEIDS’ hopelessness subscale experienced a greater increase in

momentary hopelessness than those who scored lower. On the other hand, the same pattern of

results was observed with the data on challenging activities. Earlier it was found that when

faced with challenging activities, the changes in momentary hopelessness did not differ

84
between groups. Correspondingly, when faced with the same challenging activities, the

changes in momentary hopelessness also did not differ between the high- and low-scorers in

the LEIDS’ hopelessness subscale. Overall, these comparable results support the notion that

suicidality is strongly associated with higher CR to hopelessness, which is in keeping with the

assumptions of the DAH for suicidal relapse. It also supports the potential of the LEIDS’

hopelessness subscale as a measure of the individual’s CR to hopelessness. More importantly,

the results are also indicative of the potential of the ESM as a reliable measure of

vulnerability to hopelessness in everyday life. Unlike the LEIDS which is completed on the

basis of how the individual would react/behave when he/she is feeling sad, the ESM data are

collected from the individual’s real-time responses within his/her real-life environment. In

other words, the ESM data are based on naturally occurring behaviour in everyday life as

opposed to the imagined behaviour based on hypothetical mood condition. The ability of the

ESM to capture real-life contextual factors also makes the ESM a better measure than the

LEIDS.

All in all, the results of this study extend the relevance of the DAH of suicidal relapse

from being a model of suicidal vulnerability in a previously depressed sample to a potentially

feasible model of suicidal relapse in an FEP sample. It also brings to light the role of daily

hassles (e.g. minor unpleasant events & challenging activities) in the momentary changes in

affect, which determines the reactivation of low-level attenuated hopelessness. Finally, the

outcome of this study also adds an important contribution to the literature by illustrating the

DAH as a valid cognitive model of suicidal vulnerability in psychosis that can be tested via a

structured diary technique.

85
3.4.1. Strengths and limitations

To date, this is the first study to have used the ESM to investigate the validity of the

DAH for suicidal relapse in a sample of first episode psychosis individuals. More

importantly, this is also the first study to have investigated the underlying mechanism of

suicidal thinking process by looking into the ebb and flow of momentary hopelessness in

relation to the fluctuations of affectivity in everyday life. In addition, this is the very first

study which examined the use of the Leiden Index of Depression Scales’ hopelessness

subscale as a measure of CR to hopelessness in a sample of previous suicide attempters in

FEP. This is also the first move which examined the use of the ESM as a measure of

vulnerability to hopelessness (or momentary hopelessness) in everyday life.

A number of limitations have to be considered in understanding the findings of the

present study. First, the use of ESM had a number of methodological issues. Whereas most of

the participants did not find the ESM particularly difficult, a number of participants found the

task inconvenient and slightly irritating, which was mainly due to the overall duration of the

task (6 days) and the daily frequency of time sampling (total = 10). Even though a good

number of individuals thought that ESM was a good way of keeping them more mindful of

their mood/feelings and thoughts, only a few individuals remained keen to take part again if

given the opportunity. Due to the challenging nature of the ESM task, it simply dissuades the

participants from doing it again. Some of the participants in the present study found the ESM

quite interruptive to their activities, most especially outdoor activities (such as commuting on

a bus & shopping) as it meant that they had to fill in their ESM diaries in public places.

Whereas some took part for purely altruistic reasons, a lot of the participants were motivated

by the monetary incentive upon successful completion of the task. Although these factors had

no detrimental effect on the results of the present study, it may have contributed to the second

limitation of this study, which is the slightly lower compliance rate (59%) compared to the

86
previously reported rate of 66% (Oorschot et al., 2009). However, it is important to consider

that this is only speculative and there may be other reasons why the compliance rate was

slightly lower in this study. As noted earlier, it is also possible that undertaking the ESM

during a particularly difficult period (recovery following the FEP) might have been too

challenging for the participants in general. It is also possible that the lower compliance rate

might have been influenced by the time frame during which the sampling is conducted (7.30

until 22.30). The early start meant that the participants might have missed most of the early

samplings as majority of them start their day between the hours of 10am to 12 noon.

Finally, the data on challenging activities were not conclusive and should be treated

with some caution. As the participants in the present study were still at the recovery stage

following their first psychotic episode, their typical day were therefore limited to

unstructured and solitary activities such as “watching telly, listening to music, sleeping or

napping”. Altogether, these types of activities simply do not characterise challenging daily

hassles. It is also important to note that there was a lack of activity appraisal items as only

two out of five questionnaire items factored in the principal component analysis. The lack of

questionnaire items might have caused the appraisal of challenging activities to be less

effective.

The present study has a number of implications. First, the results of this research

demonstrated the link between momentary hopelessness and organic mood fluctuations in

everyday life, which confirmed the application of the DAH of suicidal relapse in psychosis.

With the link between momentary hopelessness and negative affectivity stronger in the

suicidal history group than the non-suicidal group, it therefore suggests that although

hopeless/suicidal thoughts are attenuated when the individuals are not currently suicidal, low

levels of hopeless/suicidal thoughts remain reactive to subtle changes in NA.

87
 Second, the feasibility of the ESM as an effective assessment tool for the individual’s

vulnerability to hopelessness in everyday life, particularly in the FEP sample, may provide

future researches an alternative means to further explore the mechanism of suicidal thinking

in a context that is more organic to the individuals.

Third, the confirmation of LEIDS’ hopelessness subscale as a valuable measure of

hopelessness or suicidal vulnerability may create a platform for both researchers and

clinicians to further pursue the potential of this scale and along with it, develop more

effective ways to manage and prevent suicidal behaviour.

In conclusion, the current study found that in there is a stronger link between NA and

momentary hopelessness in the suicidal history group than the non-suicidal group in the

context of everyday life, which is in keeping with the core idea of the Differential Activation

Hypothesis of suicidal relapse. It also identified the LEIDS’ hopelessness subscale as an

effective measure of CR to hopelessness in the FEP sample. The findings of this study may

represent a platform for both researchers and clinicians to further explore the mechanism of

suicidal thinking in everyday life and develop interventions for suicidal behaviour in

psychosis, which remains a serious challenge for clinical services.

88
 CHAPTER 4

Assessing the Link between Low Mood and Lack of Problem Solving Skills

as a Behavioural Feature of Hopelessness: A Mood Priming Study

4.0. Introduction

Hopelessness, according to MacLeod and his colleagues (2005) is a “multi-faceted

construct”. If defined literally, hopelessness is simply the absence of hope. As a symptom of

depression, hopelessness is a negative view of oneself and the future. In simpler words, it is a

belief that nothing is good enough, nothing will get better, and everything will only get

worse. The more severe hopelessness becomes, the worse the depression is, and the higher

the risks of a suicidal attempt. A number of studies have suggested that hopelessness is the

link between depression and suicidal behaviour (Dyer & Kreitman, 1984; Minkoff, Bergman,

et al., 1973; Nekanda-Trepka, Bishop, & Blackburn, 1983; Salter & Platt, 1990; Wetzel,

Margulies, Davies et al., 1980). Of the significant risk factors identified for suicidal

behaviour in both healthy and psychosis samples, hopelessness was found to be closely

linked to both completed and attempted suicide (Abramson, Alloy, Hogan et al., 1998; Beck,

Steer, Kovac et al., 1985; Beck, Brown, Berchick, Stewart, & Steer, 1990; Beck et al., 1993;

Cohen et al., 1994; Conner, Duberstein, Conwell et al., 2001; Hawton & van Heeringen,

2009; Klonksy et al., 2012; Nordentoft et al., 2002; Pinto & Whisman, 1996; Robinson et al.,

2009), along with greater insight (Barret et al., 2010; Crumlish et al., 2005; Flanagan &

Compton; 2012; Foley et al., 2008). Whereas a grave physical illness represents an obvious

threat to a person’s life, hopelessness characterises a more subtle yet often a very fatal killer.

Over the years, a huge amount of effort has been made to understand this complex construct

of hopelessness but there has been only a limited success in finding ways to effectively

89
manage it. There are two main reasons for this: One, the mechanism of hopelessness is so

complex that it is still not yet fully understood. Two, there is a limited amount of information

regarding the behavioural outcomes of hopelessness, which can be of practical use to both the

clinician and the sufferer. In other words, what makes it hard for a person who is feeling

hopeless to see alternative solutions to his/her problem besides pure pessimism? On a

practical level, what are the day to day things that most people do that a person who is feeling

hopeless struggles to do besides finding a reason to live? The present study aims to

demonstrate that hopelessness can be a measured precursor to suicidal thinking.

The relationship between hopelessness and problem solving in a psychiatric sample

was first explored by Schotte and Clum (1982). The results of their study prompted the

conception of the diathesis – stress model, which suggests that chronic experiences of stress

accompanied by lack of problem solving skills increases the individual’s vulnerability to

depression, hopelessness, and suicidal ideation (Schotte & Clum, 1982). To date, a number of

studies have shown that hopelessness is in fact, associated with a lack of problem solving

skills in a sample of suicidal individuals. Williams and his colleagues (2005) described the

relationship between hopelessness and problem solving impairment as that of a “vicious

circle”. The vicious circle starts with problem solving impairment triggering suicidal

ideation, the outcome of the combined effects between hopelessness, helplessness, and

entrapment prompted by the inability to think of alternative solutions to a problem, and

suicidal ideation further impairing the individual’s problem solving ability (Williams et al.,

2005). The most commonly used procedure to examine real life problem solving is Platt and

Spivack’s (1975) Means Ends Problem Solving task. The MEPS task employs a social

context for all of its problem scenarios, which makes the procedure relevant to the everyday

life context of a wide range of research samples. The MEPS was initially developed in 1972

(Platt & Spivack) to examine the problem solving abilities and adjustment of normal

90
adolescents. It is apparent that Schotte and Clum’s (1982) initial attempt to examine problem

solving in a psychiatric sample through the use of the MEPS procedure played a vital role in

establishing two findings in the literature of suicidal behaviour to date: (1) the link between

problem solving and suicidal behaviour, (2) the use of MEPS procedure as a valid and

reliable procedure to test problem solving impairment in a sample of suicidal individuals.

Studies that looked into the problem solving abilities of individuals with histories of suicidal

behaviour have consistently found a significantly impaired problem solving ability in this

particular sample. For example, the initial study conducted by Schotte and Clum (1982)

confirmed that suicidal individuals generated fewer numbers of relevant solutions in the

MEPS task compared to the non-suicidal individuals. Similarly, Sadowsky and Kelly (1993)

found when that previous suicide attempters exhibited greater problem solving impairment

than the psychiatric controls who had never attempted. They also found that whereas both

groups showed reduced problem solving abilities compared to healthy controls, problem

solving in previous attempters was far more impaired than the psychiatric controls.

Consistent with these findings, Pollock and Williams (2001) indicated that the severity of

problem solving impairment in suicidal psychiatric patients was significantly greater

compared to a sample of patients with a similar symptom level and after controlling for the

effects of depression in both groups. In 2004, Williams and Pollock obtained a similar pattern

of results as problem solving impairment was, once again, found to be greater in the suicidal

patients than the psychiatric and healthy controls. Whereas these studies provide useful

contributions to the literature, Williams and his colleagues (2005) pointed out the difficulty in

interpreting these results. To date, the majority of the studies that have investigated the role

of problem solving in suicidal behaviour had largely employed a retrospective approach

where problem solving impairment was measured following the incidence of a suicidal

behaviour. By using this approach, it is simply impossible to conclude whether problem

91
solving impairment was a stable trait that naturally characterises individuals with histories of

suicidal behaviour, or a mere state or crisis – dependent response that causes individuals to

behave in a certain way. Contrary to the popular belief that problem solving impairment is a

stable trait (Schotte & Clum, 1982), a growing number of evidence suggest otherwise

(Schotte et al., 1990; Ivanoff Smyth, Grochowski et al., 1992; Biggam & Power, 1999). For

instance, results from Ivanoff et al.’s (1992) study showed that the history of suicidal

behaviour had no effect on the problem solving performance of incarcerated offenders.

Despite their findings confirming that problem solving impairment is not a trait phenomenon,

Ivanoff and his colleagues (1992) suggested that “the role of problem solving deficits in

suicidal behaviour may be more complex and interactive than dichotomous – that is, neither

state nor trait”. With the trait phenomenon becoming increasingly contentious, Williams and

his colleagues (2005) pointed out a question that is of critical value for future clinical work -

“How can we determine which individuals remain vulnerable to future suicidal crises even

when they appear to have completely recovered?”

The concept of “cognitive reactivity” to hopelessness is at the heart of the Differential

Activation Hypothesis of suicidal relapse and is defined as the vulnerability to hopeless

thinking or thoughts. In brief, the DAH of suicidal relapse suggests that during the early

episodes of depression, a link is formed between low mood and a pattern of negative and

maladaptive thoughts, of which hopelessness becomes a part of as a result of an intensely

negative self-referential thinking during a severe episode of depression (Malone, Oquendo,

Haas et al., 2000; Lam, Schuck, Smith, Farmer, & Checkley, 2003). The link that is formed

between low mood and hopelessness is then reinforced every time the individual experiences

a depressive episode. The stronger the link between depressed mood and hopelessness is, the

more vulnerable the individual is to hopeless thoughts when feeling particularly low in mood.

According to authors of the DAH for suicidal relapse, “it is not the resting level of

92
hopeless/suicidal cognitions that is important in rendering someone vulnerable to future

suicidal crises... it is the ease with which these patterns of thinking can be activated that is

important” (Williams et al., 2006). Such ease refers to the individual’s level of cognitive

reactivity to hopelessness. Given that the precondition to testing cognitive reactivity requires

an appropriate trigger (depressed mood), mood priming techniques were previously used to

test the assumptions of the DAH.

In order to determine who remains vulnerable to suicidal relapse amongst the

recovered attempters, the present investigation attempts to replicate the mood priming study

conducted by Williams and colleagues (2005) in sample of previously depressed patients.

The key objective was to employ the DAH framework of suicidality to a sample whose

diagnosis is psychosis. Specifically, the sample consists of first episode psychosis individuals

who were within the first 3 years of psychosis onset. Studies have previously shown that the

risk of suicidal attempts and deliberate self-harm was usually highest during the first 5 years

following the onset of psychosis (Brown, 1997; Harris & Barraclough, 1997; Hawton et al.,

2005; Palmer et al., 2005). The first aim of the study is to examine suicidal vulnerability

amongst FEP patients by looking into their ability to generate solutions to real-life problems.

More importantly, it is the study’s particular interests to compare the effects of the

experimentally induced feelings of sadness on the problem solving ability of those at high

risk of suicidal relapse (with histories of lifetime suicidal attempt/s or DSH) and those at low

risk (without any history of suicidal attempt/s or DSH). The second aim of the study is to

assess the use of the Leiden Index of Depression Scale’s hopelessness subscale as a measure

of cognitive reactivity to hopelessness and test if scores on this subscale will be associated

with the pre- to post-induction change in problem solving performance. During the

conception of the DAH of suicidal relapse, the LEIDS’ hopelessness subscale was devised in

93
order to measure the individual’s susceptibility to hopeless/suicidal thoughts when in a sad

mood (see chapter 3’s “measures” section).

4.1. Hypotheses

Effects of the Mood Challenge on Problem Solving Ability

In keeping with the DAH, the impact of the sad mood induction procedure will be

more evident in the suicidal history group than the non-suicidal group. Specifically,

1. Compared to the non-suicidal group, the suicidal history group will exhibit a greater pre- to

post- induction decrease in the number of problem-solving solutions.

2. Compared to the non-suicidal group, the suicidal history group will show a greater pre- to

post-induction decrease in the effectiveness ratings of problem solving solutions.

Effects of the Mood Challenge on Happiness and Despondence Ratings

Also in line with the assumption of the DAH,

3. The suicidal history group will exhibit a greater pre- to post-induction decrease in

happiness ratings and a greater pre- to post-induction increase in despondence ratings

compared to the non-suicidal group.

The Validity of LEIDS as a Measure of Cognitive Reactivity to Hopelessness

Prior to the mood challenge, measurements of cognitive reactivity to hopelessness were taken

using the LEIDS’ hopelessness subscale. Measured CR to hopelessness will be tested using

the DAH framework. Consistent with the DAH,

4. The suicidal history group will also show significantly greater CR to hopelessness as

measured by the LEIDS compared to the non-suicidal group.

94
4.2. Method

4.2.1. Sampling

N.B. The same sampling procedure described in Chapter 3 was also employed in this

study.

4.2.2. Measures

N.B. The measures described in the Chapter 3 (CHSF, LEIDS-R, CDSS, BHS, and

ISST) were also employed in this study.

In order to avoid contamination of answers, the BHS was always administered first

followed by the Calgary Depression Scale for Schizophrenia and InterSept for Suicidal

Thinking. These measures for depression (CDSS) and suicidal thinking (ISST) may

potentially evoke feelings of hopelessness by bringing to mind certain thoughts and feelings

associated with the individual’s previous depressive experience.

Visual Analogue Scale - Mood Rating (McCormack, Horne, & Sheather, 1988; Appendix 13)

As the name suggests, the mood rating VAS measures the participant’s subjective

mood through the use of an analogue scale (a 10cm continuous line between end points “not

at all” and “extremely”). For the purpose of this study, only the two VAS items were used:

(1) happiness, and (2) hopelessness. Each item is preceded by a statement printed above the

10cm line “At this moment, I feel...” and a description of mood printed just under the line

“happy” or “hopeless”. Participants rate their agreement/disagreement to each of the VAS

mood rating items by marking a position (vertical line) along the 10cm continuous line that

best represents how they feel. In keeping with the methodology used in Williams et al.’s

study (2005), the VAS mood rating was administered on four different time points during the

testing session: once prior to starting the testing session, once before the sad mood induction

95
procedure, once immediately after the sad MIP, and once at the end of the mood priming

task.

Means-End Problem Solving Task (Platt & Spivack, 1975; Appendix 14)

The MEPS task consists of 10 short stories or social problem scenarios where each

scenario is presented with its own beginning and ending. The MEPS task aims to assess the

participant’s social problem skills by measuring his or her ability to generate step-by-step

means or solutions to the hypothetical social problem scenarios. Scoring is based on the

number of relevant and effective solutions generated for each of the problem scenarios. Due

to its good construct validity and internal consistency (from 0.80 to 0.84; Platt & Spivack

1972, 1975), the MEPS task remains as a widely used social problem solving skills test in

many depression studies. Having adopted the MEPS procedure used in Williams et al.’s

mood priming study (2005), this study only used six out of the original ten social problem

scenarios (numbers 2, 3, 4, 6, 8 &10). The version of the MEPS items used was determined

mainly by the gender of the participant. The female version was administered only to the

female participants while the male version was administered only to the male participants.

The MEPS items on both versions were identical with the exception of the names of the

protagonists.

4.2.3. Procedure

4.2.3. a. Case Identification

The participants in this study were recruited from the Early Intervention Service in

Birmingham from March 2009 to March 2011. The author of this study approached every

care coordinator within EIS to identify service users who conformed to the inclusion criteria.

96
As established in the earlier section, two groups of participants were identified: (1) suicidal

history group, and (2) non-suicidal group. In order to ensure that all of the participants fulfil

both the inclusion and exclusion criteria, care coordinators were provided with a leaflet that

briefly explained the study and its recruitment criteria. Following referral, participants were

approached over the phone or in person, depending on their preference. During the initial

meeting, the research was presented a three-part study [ESM, Problem-solving (MEPS), &

Future Thinking (FT)], with each study investigating the mechanism of hopeless thinking by

employing contrasting methodologies – the ecological and experimental approach. In order to

counterbalance the order to which the two sets of methodologies were conducted, the three

studies were split into two sets. Set A consisted of the ecological methodology (ESM) and set

B consisted of the experimental methodology (MEPS & FT studies). Those participants who

agreed to take part in all the three studies were randomly allocated to sets AB or BA.

Following consent, the participant was asked to complete the Columbia Suicidal History

Form in order to confirm any history of suicidal attempt or deliberate self-harm. In addition,

the author also conducted an audit on the participant’s clinical case notes at EIS in order to

check for any historical entries of DSH or suicidal attempt. The LEIDS questionnaire was

also conducted immediately following consent, which was on average at least a week prior to

the testing session, in order to avoid two possible sources of contamination: (1)

contamination from responses to other measures administered prior to the testing session (e.g.

BHS or CDSS), one of these measures might evoke an emotional response which could

potentially influence their responses on LEIDS or vice versa, and (2) contamination from any

residual effects of the sad mood induction procedure.

Prior to starting the testing session, participants were briefed about the details of the

study and given an opportunity to ask questions. Following this, a set of questionnaires

measuring hopelessness (BHS), depression (CDSS), and suicidal thinking (ISST) was

97
completed. The MEPS and Future Thinking tasks (chapter 5) were both carried out on two

occasions, once after the completion of questionnaires which is just prior to the sad mood

induction procedure (pre-induction) and once after the sad MIP (post-induction). In line with

Williams et al.’s study (2005), the tasks were completed in the same order for both pre-and

post-induction, with the MEPS task first followed by the FT task. A debriefing was carried

out at the end of the testing session to discuss the actual purpose of the MEPS task and more

importantly, to check if the participant’s mood had returned to its normal level. Participants

who remained upset at the end of the testing session were offered a Happy Mood Induction

Procedure to counteract the effects of the sad MIP. In keeping with what was agreed in the

consent form, participants were also informed that their care coordinators will be requested to

closely monitor on their mood for as long as they think it is necessary to do so. Out of the

three participants who reported feeling upset, only two agreed to complete the happy MIP.

All three participants consented to have their care coordinators informed in order to ensure

that their mood will be monitored closely until deemed necessary. As an appreciation of their

time and contribution, all of the participants received a payment of £20 at the end of the

testing session.

4.2.3. b. The Sad Mood Induction Procedure

The sad MIP used in this study was adopted from Williams et al.’s mood priming

study in 2005. Their version of the sad MIP employed the combined techniques of the Velten

procedure and musical mood induction procedure. Prior to starting the sad MIP, participants

were briefed about the purpose and details of this procedure. The researcher explained that

the sad MIP will induce them to feel sad by reading a set of 30 Velten negative statement

cards (Appendix 15) while listening to a sad music playing in the background. The music

used in this procedure was Prokofiev’s Russia Under the Mongolian Yoke, which was re-

98
mastered at half-speed using Wavepad Sound Editor version 5.13. Participants were

instructed to read each card carefully and internalize the thoughts and feelings evoked by the

negative statement written on each card (e.g. “I am discouraged and unhappy about

myself.”). While doing this, participants were asked to identify the cards which they felt were

more effective in making them feel sad and to put these cards on a separate pile. This set of

cards was then later used in the two booster versions of the sad MIP, one prior to post-mood

induction MEPS and another one prior to the post-mood induction Future Thinking task

(chapter 5). The booster sad MIP was simply a shorter version of the original sad MIP with

fewer negative Velten statement cards to go through.

4.2.3. c. Means-Ends Problem Solving Task

The MEPS task was presented to the participants as a ‘story-telling’ task, which aims

to explore their creativity. Six problem scenarios were split into two sets of three. Set 1

consisted of scenarios about ‘relationship difficulties with boyfriend/girlfriend’, ‘finding a

lost wristwatch’, and ‘making friends in a new neighbourhood’ (MEPS items 2, 3, & 4). Set 2

consisted of scenarios about ‘starting relationship ’, ‘difficulties with friends’, and

‘difficulties with supervisor at work’ (MEPS items 6, 8, & 10). Each participant was

randomly allocated to sets 1/2 or 2/1 in order to counterbalance the presentation of MEPS

items before (pre) and after (post) the sad mood induction procedure.

The participants were given one problem scenario to solve at a time. The researcher

read each problem scenarios to the participants who, at the same time, followed what was

being read on a separate card. Each scenario begins with a brief description of the protagonist

facing a problem and ends with the protagonist successfully solving it while leaving the

middle part of the scenario unknown. The participants were then given a time limit of 2

minutes to supply the middle part of the story by describing what they thought had happened,

99
which led to the successful ending of the story. All of the MEPS tasks were recorded using a

dictaphone in order to allow the raters to score the task at a later point. In keeping with

Williams et al.’s study (2005), each scenario was scored two ways: (1) for the number of

relevant means/solutions, and (2) for the effectiveness of the solutions. The scoring was

completed separately by two blind raters. The blind raters were psychology undergraduate

students who were properly trained by the author of the study prior to scoring the MEPS task.

A solution/mean was rated as “relevant” if the course of action led to the desired ending of

the story (Platt & Spivack, 1975). Only actions that were taken by the protagonist were rated

as valid. On the other hand, a 7-point Likert scale (1 = not at all & 7 = extremely effective)

was employed to rate the effectiveness of the solutions for each of the stories. The intra-class

correlation between the ratings of the two independents raters for the number of solutions was

r = .81, p <.001 whilst the intra-class correlation for the effectiveness ratings was r = .92, p

<.001. The two raters reviewed all of the recorded tasks again until 100% agreement was

reached on the number of solutions and effectiveness ratings. The average number of

solutions and effectiveness ratings for each task (pre- & post-induction MEPS) were

calculated by adding the scores of the three problem scenarios divided by three.

4.2.4. Analysis Strategy

To test the hypotheses, a mixed between/within repeated measures analysis of

variance was conducted using an IBM SPSS Statistics software version 21 for Windows. In

order to control for the possible effects of the key clinical symptoms (e.g. generalised

hopelessness, depression, and suicidal thinking), two sets of analysis of covariance using the

repeated measures design were conducted. The purpose of the initial ANCOVA was to test

for any clinical symptom that significantly interacts with the main outcome variable on the

whole. If a significant interaction is found, the ANCOVA was repeated between groups with

100
the specific clinical symptom entered as a covariate. An alpha level of .05 was used for all

statistical tests.

4.3. Results

4.3.1. Sample Characteristics

Of the 105 individuals who were approached, only 3 individuals responded with an

outright refusal. Out of the recruited 102 participants, a subsample of 3 (2.94%) changed their

mind about participating in the study (those who previously completed the LEIDS screening

measure but refused to do the mood-priming study) while the other 2 (1.96) opted out from

the MEPS task (but carried on completing the other task in the mood-priming study). In total,

the final sample consisted of 97 participants (37 females and 60 males) of which, 48

(48.98%) had a lifetime history of suicidal behaviour while 49 (50%) had no history of

suicidal behaviour in their lifetime. The participants’ age and key symptom scores are

summarised in Table 12.

Table 12. Means and Standard Deviations of Age and Key Symptom Scores for the Non-

Suicidal Group and Suicidal History Group

Non-suicidal group (N = 49) Suicidal history group (N = 48)

Variable M SD M SD
Demographic
Age 23.86 5.00 23.16 4.66

Symptom Score
BHS 5.61 4.43 9.92 5.90
CDSS 1.73 2.47 3.96 3.99
ISST .39 1.52 1.77 3.12
Note: BHS = Beck Hopelessness Scale, CDSS = Calgary Depression Scale, and ISST = InterSept Scale for
Suicidal Thinking

101
4.3.2. Descriptive Statistics and T-tests

Generalised Hopelessness

Current levels of generalised hopelessness were measured using the Beck

Hopelessness Scale. An independent t-test revealed a significant difference between groups,

with the suicidal history group (M = 9.92, SD = 5.90) showing a higher level of generalised

hopelessness compared to the non-suicidal group (M = 5.61, SD = 4.43), t(95) = 3.92, p <

.001, d = .82.

Depression

Symptoms of depression were assessed using the 10-item Calgary Depression Scale

for Schizophrenia. Scores between the two groups were compared and an independent t-test

revealed a significant difference between the suicidal history group (M = 3.96, SD = 3.99)

with the non-suicidal group (M = 1.73, SD = 2.47) with the suicidal history group showing

higher levels of depression than the non-suicidal group, t(95) = 3.23, p = .002, d =.67.

Suicidal Thinking

Levels of suicidal ideation a week prior to testing were measured using the InterSept

Scale for Suicidal Thinking. Scores from both groups were compared using an independent t-

test, which revealed a significant difference between the suicidal history group (M = 1.77, SD

= 3.12) and the non-suicidal group (M = .39, SD = 1.52) with the suicidal history group

showing higher levels of suicidal ideation than the non-suicidal group, t(95) = 2.91, p = .005,

d = .56.

102
4.3.3. Hypotheses Testing

Data transformation was carried out on all of the continuous variables (e.g. dependent

& covariates) prior to conducting the ANOVA in order to satisfy the assumption of normality

and equality of variances. The data were transformed using the square root conversion

following Tabachnick and Fidell’s (2007) and Howell’s (2007) suggested guideline in data

transformation. The guideline suggests that square root transformation was more appropriate

if the data distribution was moderately skewed (positive or negative skewness). Also, the use

of square root employs the minimum amount of transformation to improve normality

compared to the other transformation procedures (e.g. logarithmic & inverse). This was

evident when a set of data was converted using square root and logarithmic transformation

for the purpose of contrast. Overall, data transformed via square root had better improvement

in normality when contrasted against data transformed via logarithmic method.

In keeping with Tabachnick and Fidell (2007) and Howell (2007), all means and

standard deviations reported in the following analyses were original values from the

untransformed data.

An alpha level of .05 was used for all statistical tests. All analyses were carried out

with group (suicidal history group vs. non-suicidal group) as a between-subjects factor.

Effects of the Mood Challenge on Problem Solving Ability

In keeping with the DAH, the impact of the sad mood induction procedure will be

more evident in the suicidal history group than the non-suicidal group. Specifically,

1. Compared to the non-suicidal group, the suicidal history group will exhibit a greater pre- to

post-induction decrease in the number of problem-solving solutions.

103
 To test this hypothesis, a two-fold analysis was employed. First, independent t-tests

were conducted to compare the MEPS scores of each group (suicidal history group vs. non-

suicidal group) before and after the sad mood induction procedure. Second, a mixed repeated

measure ANOVA was conducted to examine the effect of the sad mood induction procedure

on the problem solving ability of the suicidal history group and the non-suicidal group. To

conduct the mixed-repeated measure ANOVA, the variable group (suicidal vs. non-suicidal)

was entered as a between-subjects factor and mood (mood 1 = pre-induction & mood 2 =

post-induction) as within factor. The main aim of employing an ANOVA was to look into the

interaction effect between mood (pre- & post-induction) and group (suicidal & non-suicidal)

on problem solving ability.

Results of the independent t-test on the pre-mood induction MEPS task showed that

the suicidal history group (M = 6.44, SD = 2.02) generated fewer relevant means than the

non-suicidal group (M = 7.37, SD = 1.79), t (95) = 2.40, p = 0.018, d = .31. Similar results

were found on the post-mood induction task as the suicidal history group (M = 4.17, SD =

1.80), once again, generated less number of relevant means than the non-suicidal group (M =

6.22, SD = 1.57), t (95) = 6.00, p < .001; d = .73. Summary of means and standard deviations

are displayed in Table 13.

In line with the hypothesis, a mood x group interaction effect was observed as the

suicidal history group had a greater pre- to post-induction decrease in the number of problem

solving means compared to the non-suicidal group [F(1, 95) = 13.19, p <.001, partial ƞ2 =

.12]. There was also a significant within-subjects main effect of mood brought by the

decrease in the overall MEPS scores following the mood-induction procedure [F (1, 95) =

80.78, p <.001, partial ƞ2 = .46], and a significant between-subjects main effect of group due

to the fewer problem solving solutions in the suicidal history group than the non-suicidal

group [F(1, 95) = 21.73, p <.001, partial ƞ2 = .19]. The results remained significant after

104
controlling for generalised hopelessness (BHS), depression (CDSS), and suicidal thinking

(ISST).

Table 13. Means and Standard Deviations of the Pre- and Post- Mood Induction Number of

Problem Solving Solutions

Non-suicidal group (N = 49) Suicidal history group (N = 48)

Variable M SD M SD

Pre-induction 7.37 1.79 6.44 2.02

Post-induction 6.22 1.57 4.17 1.80

In summary, the results confirmed the hypothesis as the suicidal history group

exhibited a greater pre- to post-induction decrease in the number of problem solving solutions

compared to the non-suicidal group. Figure 2 illustrates the average number of solutions

generated by the suicidal history group and the non-suicidal group before and after the sad

mood induction procedure.

105
Figure 2. Average Pre- and Post-Induction Number of Problem Solving Solutions for the
Suicidal History Group and Non-Suicidal Group

2. Compared to the non-suicidal group, the suicidal history group will show a greater pre- to

post-induction decrease in the effectiveness ratings of problem solving solutions.

Following the analysis in hypothesis 1, independent t-tests were conducted in order to

compare the effectiveness of the problem solving means generated by each group before and

after the mood induction. A mixed repeated measure ANOVA was also conducted to

determine if the mood challenge had a differential effect on the effectiveness ratings of the

problem solving solutions generated by the suicidal history group and the non-suicidal group.

Similar to the repeated measures ANOVA conducted in hypothesis 1, group was entered as a

between-subjects factor and mood (mood 1 = pre-induction & mood 2 = post-induction) as a

within-subjects factor. Again, the interaction effect between mood (pre- & post-mood

induction) and group (suicidal & non-suicidal) was of key interest in this analysis.

Results of independent t-test on the pre-mood induction effectiveness ratings showed

no significant difference between the suicidal group (M = 5.16, SD = 1.45) and the non-

106
suicidal group (M = 5.99, SD = 1.90), t (95) = 1.30, p = .196. In contrast, t-test results on the

post-mood induction effectiveness ratings showed that the suicidal group (M = 3.91, SD =

2.01) scored significantly lower than the non-suicidal group (M = 5.06, SD = 1.96), t (95) =

2.43, p = .017, d = .50. Table 14 displays summary of means and standard deviations.

Table 14. Means and Standard Deviations of the Pre- and Post- Mood Induction Effectiveness

Ratings of Problem Solving Solutions

Non-suicidal group (N = 49) Suicidal history group (N = 48)

Variable M SD M SD

Pre-induction 5.99 1.90 5.16 1.45

Post-induction 5.06 1.96 3.91 2.01

Contrary to the hypothesis, no significant group x mood interaction effect was found

[F(1, 95) = 1.42, p = .237]. However, a significant within-subjects main effect on mood was

found as caused by the decrease in the effectiveness ratings following the mood induction

[F(1, 95) = 16.25, p <.001, partial ƞ2 = .15]. Between-subjects main effect of group was also

found as caused by the considerably lower effectiveness ratings in the suicidal history group

than the non-suicidal group [F(1, 95) = 5.70, p =.019, partial ƞ2 = .06].

In summary, results of the repeated measures ANOVA did not support the hypothesis.

The suicidal history group did not show a greater pre- to post-induction decrease in the

effectiveness ratings of problem solving solutions as predicted. Figure 3 displays the average

effectiveness ratings for each group before and after the sad mood induction procedure.

107
Figure 3. Average Pre- and Post-Induction Effectiveness Ratings of Problem Solving

Solutions for the Suicidal History Group and Non-Suicidal Group

Effects of the Mood Challenge on Happiness and Despondence Ratings

Also in line with the assumption of the DAH,

3. The suicidal history group will exhibit a greater pre- to post-induction decrease in

happiness ratings and a greater pre- to post-induction increase in despondence ratings

compared to the non-suicidal group.

Replicating the two-step analysis in hypothesis 1 and 2, independent t-tests were

conducted in order to compare the levels of momentary happiness and hopelessness in each

group before and after the mood induction. A mixed repeated measure ANOVA was also

conducted to determine if the effect of the sad mood induction procedure on the levels of

happiness and despondence and whether the effect will vary between the suicidal history

group and the non-suicidal group. Following the ANOVA analyses in hypotheses 1 and 2, the

108
group variable was entered as the between-subject factor while mood (mood 1 = pre-

induction & mood 2 = post-induction) was entered as the within-subject factor. Again, the

interaction effect between mood (pre- & post-mood induction) and group (suicidal & non-

suicidal) was of key significance in this hypothesis testing.

Happiness Ratings

Independent t-test on pre-mood induction happiness ratings revealed no significant

difference between the suicidal history group and the non-suicidal group [mean (SD) = 5.33

(2.09) & 5.73 (2.00), respectively; t (95) = .96, p = .337]. In contrast, independent t-test on

post-mood induction happiness ratings showed a significant difference with the suicidal

history group scoring lower than the non-suicidal group [mean (SD) = 4.06 (1.72) & 5.04

(1.98), respectively; t (95) = 2.48, p = .015; d = .50]. Summary of means and standard

deviations are displayed in Table 15.

As predicted, results indicated a significant time x group interaction effect due to the

greater pre- to post-induction decrease in happiness ratings within the suicidal history group

in comparison to the non-suicidal group following the mood induction [F(1, 95) = 4.723, p

=.082, partial ƞ2 = .032]. Results also showed a significant within-subjects main effect of

mood as caused by the decrease in the happiness ratings following the mood induction [F(1,

95) = 42.68, p <.001, partial ƞ2 = .31]. There was, however, no significant between-subjects

main effect of group [F(1, 95) = 3.091, p =.082]. The results remained significant after

controlling for generalised hopelessness (BHS), depression (CDSS), and suicidal thinking

(ISST).

109
Despondence Ratings

Independent t-test on pre-mood induction momentary despondence ratings revealed

no significant difference between the suicidal history group and the non-suicidal group [mean

(SD) = 3.28 (2.34) & 2.80 (2.51), respectively; t (95) = 1.10, p = .276]. In contrast,

independent t-test on post-mood induction hopelessness ratings showed a significant

difference with the suicidal history group scoring lower than the non-suicidal group [mean

(SD) = 4.87 (2.44) & 3.36 (2.48), respectively; t(95) = 2.78, p = .007; d = .56]. Table 15

displays summary of means and standard deviations.

Table 15. Means and Standard Deviations of the Pre- and Post- Mood Induction Happiness

and Despondence Ratings

Non-suicidal group (N = 49) Suicidal history group (N = 48)

Variable M SD M SD

Happiness Ratings

Pre-induction 5.73 2.00 5.33 2.09

Post-induction 5.04 1.98 4.06 1.72

Despondence Ratings

Pre-induction 2.80 2.51 3.28 2.34

Post-induction 3.36 2.48 4.87 2.44

In agreement with the hypothesis, a significant group x time interaction effect [F (1,

95) = 4.48, p =.037, partial ƞ2 = .04] was found due to the greater pre- to post- induction

increase in despondence ratings within the suicidal history group in comparison to the non-

suicidal group. A within-subjects main effect of mood was also found due to the decrease in

post-mood induction despondence ratings [F(1, 95) = 32.71, p <.001, partial ƞ2 = .26].

110
Additionally, a significant between-subjects main effect of group was found, with the suicidal

history group showing higher despondence ratings compared to the non-suicidal group [F(1,

95) = 4.18, p =.044, partial ƞ2 = .04]. The results remained significant after controlling for

generalised hopelessness (BHS), depression (CDSS), and suicidal thinking (ISST).

In summary, the results were consistent with the hypothesis as the suicidal history

group exhibited a greater pre- to post-induction decrease in happiness ratings and a greater

pre- to post-induction increase in hopelessness ratings compared to the non-suicidal group.

Figure 4 illustrates the fluctuation of happiness ratings and despondence throughout the

testing session.

The Validity of LEIDS as a Measure of Cognitive Reactivity to Hopelessness

Prior to the mood challenge, measurements of cognitive reactivity to hopelessness were taken

using the LEIDS’ hopelessness subscale. Measured CR to hopelessness will be tested using

the DAH framework. Consistent with the DAH,

4. The suicidal history group will also show significantly greater CR to hopelessness as

measured by the LEIDS compared to the non-suicidal group.

An independent t-test was conducted in order to compare the means of the two groups

on LEIDS’ hopelessness subscale. As hypothesised, the suicidal history group (M = 12.63,

SD = 5.25) showed significantly higher mean score than the non-suicidal group (M = 6.20,

SD = 4.13), t(95) = 6.34, p < .001, d = 1.36).

111
Figure 4. Average VAS Despondence (A) and Happiness (B) Ratings on the Pre- and Post-Induction Tasks for the Suicidal History Group

and the Non-Suicidal Group

A B

112
The DAH of suicidal relapse suggests that the occurrence of low mood will trigger

hopelessness. As a behavioural feature of hopelessness,

5. The deterioration in problem solving ability following the mood challenge will be

correlated with greater levels of CR to hopelessness as measured by the LEIDS.

To test this hypothesis, a difference score was first calculated by subtracting the pre-

mood induction number of problem solving means (PSM) from the post-mood induction

number of PSM. Next, a bivariate correlation was carried out on the difference score (or the

pre- to post-induction change in the number of problem solving means) and the LEIDS’

hopelessness subscale scores. Contrary to the hypothesis, there was no significant correlation

between scores on LEIDS’ hopelessness subscale and the pre- to post-induction decrease in

the number of problem solving solutions (r = -.11, N = 97, p = .280). Similar results were

found when correlation was conducted on the original data.

4.4. Discussion

The key objective of this study was to explore the applicability of the DAH of suicidal

relapse in understanding the suicidal thinking mechanism in FEP individuals, who are at high

risks of suicidality as a result of their history of suicidal attempt or DSH. Encouraged by the

studies conducted by a group of researchers who developed the idea of the DAH (e.g.

Teasdale, Williams, Lau, Segal, & Barnhofer) along with the aspiration to make a valuable

contribution to the literature of suicidal behaviour, the current study was conducted with the

aim of shedding light on the phenomenon of suicide in a clinical group which is at a higher

risk of hurting or killing themselves. Previous studies have shown that suicide in

schizophrenia was highest during the early phase of the illness, typically during the first five

years after the initial psychotic episode (Brown, 1997; Harris & Barraclough, 1997; Palmer et

113
al., 2005). But in keeping with suicide in other high risk groups, understanding who and

when individuals may choose to harm themselves, remains a major challenge for clinicians.

In line with the literature on social problem solving as a behavioural marker

associated with suicidal behaviour, this study employed the Means Ends Problem Solving

task (Platt & Spivack, 1975) with the aim of comparing the performances of a high risk

suicidal history group and low risk non-suicidal group, in a baseline mood (pre-induction) vs.

sad induced mood condition (post-induction). The use of the sad mood induction procedure

played a crucial part in assessing if the assumption of the DAH on cognitive vulnerability to

suicidal relapse was applicable to the FEP sample.

As predicted, the results indicated that the suicidal history group had a significantly

greater pre- to post-induction decrease in the problem-solving means compared to the non-

suicidal group. The same findings were found when current levels of hopelessness (BHS),

depression (CDSS), and suicidal thinking (ISST) were used as covariates. This suggests that

the decrease in the post-induction number of relevant solutions in the suicidal history group

was mainly due to the downward shift in mood and the group’s pre-existing CR to

hopelessness as suggested by the DAH framework. These findings were consistent with

previous studies (Sadowsky & Kelly 1993, Pollock & Williams, 2001; Pollock & Williams,

2004) but were at variance with the data from Williams et al.’s mood priming study (2005),

which reported no significant difference in the number of post-induction problem-solving

solutions between those with a history of mood depressive disorder and suicidal ideation,

those with MDD but without the history of suicidal ideation, and those with neither MDD or

suicidal ideation. These conflicting results might be partially due to the dissimilar sample

characteristics. Firstly, whereas their study only recruited those who were symptom-free from

depression for at least 8 weeks, this study only excluded those who were severely depressed

and suicidal at the time of assessment; this was because low mood is a prevalent feature of

114
psychosis at all phases of the disorder (Birchwood et al., 2000). However, it is important to

note that due to the active involvement of the care coordinators in the recruitment process of

this study, only those who were relatively stable in mood were actually referred for

recruitment. Secondly, whereas Williams et al.’s (2005) previous study included the

experience of suicidal ideation as part of the criteria for suicidal behaviour, the current

study’s criteria for suicidality were strictly limited to actual suicidal attempts or incidents of

deliberate self-harm. Thirdly, whereas the main clinical diagnosis of the sample in their

study was MDD, the clinical diagnosis and focus of this study was early psychosis and

schizophrenia. It is vital to note, however, that the particular characteristics of the sample

used in this study was key to extending the use of the DAH from being a general theory of

suicidal relapse in previously depressed individuals to a framework that can potentially

elucidate suicidal vulnerability in early psychosis, a particularly crucial period for young

individuals who are still trying to come to terms with the trauma of the initial episode

(Harrison & Fowler, 2004; Jackson & Iqbal, 2000; Jackson, et al., 2004; Riedesser, 2004;

Tarrier, Khan, Cater, & Picken, 2007) and the subjective distress associated with this

experience (Brunet, Birchwood, Upthegrove, Michail, & Ross, 2012; McGorry, Patrick,

Chanen et al., 1991). Whereas depression in its severe form can develop into psychosis,

depression can also develop out of the traumatic experience that an episode of psychosis can

bring (Birchwood, Iqbal, Trower et al., 2000). Hafner and his colleagues (1998) have

suggested that the adolescence of young people with schizophrenia is characterised by

periods of low mood and crises of self-esteem. By using the first-episode of psychosis

sample, the present study was able to explore if the DAH also applies to psychosis. The DAH

suggests that suicidal relapse occurs when depressed mood and hopelessness are strongly

linked to each other such that the experience of low mood will trigger hopelessness, which is

a known risk factor for suicidal behaviour. If this link between low mood and hopelessness

115
were assumed to develop through repeated depressive episodes, how many episodes are

sufficient enough to create such a robust link between the two? Do traumatic life events (i.e.

psychotic experiences & hospitalisation) provide a context that hastens the formation of a

strong link between the depressed mood and feelings of hopelessness? Whereas the answers

to this question remain unknown, the fact that the present study found significantly fewer

numbers of relevant means compared to the non-suicidal group following the mood challenge

and Williams et al.’s (2005) study did not find any differences in a sample of MDD patients,

suggests that there seems to be a greater vulnerability to hopeless thinking within the

psychosis sample when mood is relatively low.

Contrary to the hypothesis, there was no significant difference in the degree of pre- to

post-induction change in the effectiveness ratings of the problem-solving means generated in

both groups. Although the effectiveness ratings were considerably reduced following the sad

MIP and the effectiveness ratings of the suicidal history group in general was significantly

lower than the non-suicidal group, the effect of the sad MIP on the effectiveness ratings

simply did not differ between groups. These results suggest that although the difference in the

post-induction effectiveness ratings between the suicidal history group and the suicidal

history group was not substantial enough to be detected, the overall effectiveness ratings of

the suicidal history group was significantly lower than the non-suicidal group. Interestingly,

this significant between-group distinction on the effectiveness of their problem-solving

solutions was not simply caused by their differences in current levels of generalised

hopelessness (BHS), depression (CDSS), or suicidal thinking (ISST), as the results were re-

tested with these key clinical symptoms as covariates. Intriguingly, it remains a mystery as to

why the groups did not differ in their effectiveness ratings following the sad mood induction

procedure. The only possible reason for this is that the pre-existing group differences on the

pre-induction effectiveness scores caused the decrease in post-induction effectiveness scores

116
to diminish. Such a possibility was demonstrated when the groups significantly differed in the

post-mood induction effectiveness ratings after controlling for the effects of the pre-mood

induction effectiveness ratings. However, should pre-existing differences need controlling

when the groups were naturally different to start with? Whereas pre-existing differences are

customarily controlled for in standard pre-post experimental designs, the current study

embraced the notion that the suicidal history group and non-suicidal group have naturally

occurring, if not acquired, intrinsic differences. The fact that one group of individuals have

attempted to kill or hurt themselves at some point in their lives when the other group of

individuals have not, underlines that the two groups were distinct in significant ways. For

instance, a number of previous studies have already illustrated the differential problem-

solving abilities between those with and without histories of suicidal behaviour in a

psychiatric sample (Curry et al., 1992; Reinecke et al., 2001; Watkins & Baracaia, 2002;

Williams et al., 2005; Arie, Apter, Orbach et al., 2008).

Also in agreement with the predictions of the present study, the suicidal history group

exhibited a greater pre- to post-induction decrease in happiness ratings and a greater pre- to

post- induction increase in despondence ratings as measured by the Visual Analogue Scale

compared to the non-suicidal group. Importantly, these distinctly greater degree of changes in

the pre- to post-induction mood ratings (e.g. greater decrease in happiness & greater increase

in despondence) exhibited by the suicidal history group were not just consequences of higher

levels of generalised hopelessness (BHS), depression (CDSS), and suicidal thinking (ISST)

as appropriate tests were made to check if the results remain significant after controlling for

the key symptoms. Overall, this pattern of results was consistent with the findings in

Williams et al.’s (2005) study where participants with mood depressive disorder and histories

of suicidal ideation exhibited less happiness and more despondence following the sad mood

induction procedure. The present study confirms the results of the previous investigation but

117
in a sample of first-episode of psychosis whose vulnerability to suicide is not only at high risk

(Brown, 1997; Harris & Barraclough, 1997; Palmer et al., 2005), but also largely

unpredictable at the individual level (Power, 2010). Current findings indicate that despite the

particularly difficult and emotionally stressful post-psychotic period currently experienced by

both groups, only those with histories of suicidal behaviour exhibited greater sensitivity to the

sad MIP as evidenced by their considerably greater degree of changes in their pre- to post-

induction happiness and despondence ratings. However, it is also important to consider that

the differential effects of the sad MIP on both groups may only be due to the natural, pre-

existing group differences in mood prior to the mood induction. It can be argued that the

suicidal-history group might already had significantly greater despondence ratings prior to

the sad MIP than the non-suicidal group and was therefore more likely to respond with

greater despondence than the latter group to the sad MIP. Results of independent t-tests,

however, revealed that the groups did not differ in their overall despondence and happiness

ratings. In other words, there were no pre-existing group differences in the pre-induction

despondence ratings that could have biased the data in favour of the suicidal-history group.

As the DAH focuses on the individual’s cognitive vulnerability to hopelessness, it was crucial

that the results of independent t-tests have established that the degree of change in pre- to

post-induction mood ratings was not simply due to the pre-existing vulnerability to hopeless

thinking during the pre-induction stage. Due to this, it was easier to determine that the degree

of change in the pre- to post-induction problem-solving abilities was mainly due to the

individual’s cognitive vulnerability to hopelessness when in a sad mood and not simply due

to the worsening of a pre-existing vulnerability or mood state.

In line with the assumption of the DAH, the suicidal history group exhibited greater

cognitive reactivity to hopelessness by scoring higher in the hopelessness subscale of the

Leiden Index of Depression Scale – Revised version. However, contrary to the hypothesis,

118
the CR to hopelessness as measured by the LEIDS’s hopelessness subscale did not correlate

well with the number of problem-solving solutions following the mood induction. One

possible reason is that a number of participants (conservative estimate of less than 10)

reported filling in the LEIDS questionnaire to be a difficult challenge. This particular group

of participants expressed some difficulty imagining a hypothetical sad mood state, which was

necessary if the LEIDS was to measure CR to hopelessness properly. Nevertheless, this group

of participants was relatively small in number and cannot account for the lack of association

between CR to hopelessness and number of relevant problem solving solutions. It is also

possible that the lack of association between the two was due to the fact that experimentally

induced change in number of problem solving solutions does not accurately represent suicidal

vulnerability in real life thus, was unable to demonstrate a detectable link with CR to

hopelessness. Finally, there is also a possibility that the LEIDS’ hopelessness subscale was

simply not able to effectively capture the key elements that embody CR to hopelessness in

this particular sample.

In conclusion, the results of this study were consistent with previous findings that

individuals with histories of suicidal behaviour were more impaired at solving problems

particularly when mood is low, which were in keeping with the assumptions of the

Differential Activation Hypothesis of suicidal relapse. This significantly noticeable problem

solving impairment found in participants with histories of suicidal behaviour, especially

following the sad mood induction, suggest that a subtle shift in mood (from neutral to sad

mood induced) can impair the problem solving ability of this sample and reactivate some low

level feelings of hopelessness. All in all, the results of this study support the assumptions of

the Differential Activation Hypothesis in a number of ways. First, it illustrated the that DAH,

as a hypothesis of suicidal relapse in a previously depressed sample, is also a valid model of

suicidal relapse in a sample whose primary diagnosis is psychosis instead of depression.

119
Second, this study contributes more evidence to the literature that supports the DAH’s as a

valuable model of suicidal vulnerability by confirming that the DAH is not just a mere

cognitive paradigm, but also a model that can be tested concretely via quantifiable

behavioural markers (e.g. problem solving ability).

4.4.1. Strengths and limitations

This study has some limitations which need to be borne in mind. First, there was only

one manipulated treatment variable (sad MIP) employed in this study, which meant that the

comparison of problem-solving abilities between the suicidal and non-suicidal groups were

only limited to the effect of this particular manipulation. Whereas the DAH only accounts for

the individual’s CR to hopelessness when in a depressed mood, other manipulations (i.e.

happy or neutral MIP) could have illustrated the mechanism of hopelessness when the

individual’s mood is happy or neutral. Although the absence of a neutral or a happy mood

induction did not have an unfavourable effect in the results of the present study, the neutral

mood, in particular, could help establish if the changes in the number of relevant solutions

following the mood induction were indeed due to shift in mood and not from other

undesirable factors (e.g. boredom or loss of interest in the study, tiredness, & possible

participant bias). Second, the results for the effectiveness ratings of the relevant means

showing no difference between the suicidal history group and non-suicidal group following

the mood induction contradicted previous research (Williams et al., 2005). It is vital to

consider, however, that the suicidal history group showed significantly lower effectiveness

ratings than the non-suicidal group after controlling for the pre-induction effectiveness

ratings. Finally, although the key predictions in the study were confirmed and in keeping with

the assumptions of the DAH as a model for suicidal relapse, this study was only able to

illustrate the effects of minor shifts in mood on the problem solving ability of an individual.

120
Although the ESM study (chapter 3) suggests that the DAH is ecologically valid and the

results of the present study are consistent with the ESM data, a follow study is still needed in

order to verify if the observed suicidal vulnerability as measured in the problem-solving task

following the mood induction will predict suicidal relapse in real life. To date, no studies

were able to demonstrate this and should therefore also be seen as a useful avenue for future

research.

This study has a number of strengths. First, to the best of the author’s knowledge, this

study is the first to explore the mechanism of suicidal thinking in psychosis using the mood

priming technique. Second, the sample is of a reasonably size compared to the sample size of

a similar study conducted by Williams et al.’s study (N = 34); having recruited nearly three-

fold of the sample size of a clinical group that is often not easy to engage, this is a positive

achievement. Third, Birmingham as a culturally diverse city offered this study an excellent

opportunity to investigate a sample that was of a good mix in terms of their ethnicity and

social backgrounds (i.e. religion & family structures).

Overall, the findings of this study have important clinical implications. The

prevention and management of suicidal behaviour in psychosis have not been greatly

successful so far. To date, this is the first study to have explored the suicidal thinking

mechanism in early psychosis and the significant results from this study present two valuable

implications: (1) that the mood priming technique is a safe and effective method for studying

the suicidal thinking processes, and (2) that the use of behavioural measures (e.g. problem

solving tasks) following a mood challenge is a useful way to compare suicidal/hopeless

thoughts relative to mood. Further, the findings of this study could serve as a platform for

other researchers to further explore problem solving ability as one of the key behavioural

markers for suicidal vulnerability in psychosis. Most importantly, the results supporting the

assumption of the DAH for suicidal relapse could also serve as a platform for other

121
researchers to further investigate the application of the DAH as a framework of suicidality in

psychosis. Specifically, extending the DAH as a guide to suicide risk assessments in first

episode psychosis could be of great value to clinicians. Previous studies have shown that the

stage following the initial psychotic episode is particularly crucial as the risks for both

attempted and completed suicide are not only high but also largely unpredictable (Brown,

1997; Harris & Barraclough, 1997; Palmer et al., 2005; Power, 2010).

122
 CHAPTER 5
Assessing the Link between Low Mood and Lack of Positive Future

Fluency as a Behavioural Feature of Hopelessness: A Mood Priming Study

5.0. Introduction
Individuals who are at the early stages of psychosis, especially those who are still

recovering from an initial psychotic episode, have been found to have a high level of suicidal

ideations (Birchwood, Mason, MacMillan, & Healy, 1993; Rooke & Birchwood, 1998; Iqbal,

Birchwood, Chadwick, & Trower, 2000) and suicidal attempts (Brown, 1997; Harris &

Barraclough, 1997; Heila Isometsa, Henriksson et al., 1997, 1999; King, Baldwin, Sinclair et

al., 2001; Nordentoft et al., 2002; Palmer et al., 2005; Power, 2010; Walsh et al., 2001). The

role of hopelessness as a risk factor for suicidal behaviour (both ideations & attempts) in

early psychosis has already been illustrated in previous studies (Nordentoft et al., 2002;

Hawton et al., 2005; Pompili, Lester, Grispini et al., 2009). Theoretically, the function of

hopelessness as a part of the suicidal thinking mechanism has also been demonstrated in

different ways (Schotte & Clum, 1987; Williams et al.’s, 2005; Johnson, Gooding, &Tarrier,

2008). Empirically, hopelessness has been found to be associated with certain cognitive and

behavioural characteristics, such as deficits in autobiographical memory (Williams, 1996;

Goddard, Dritschel, & Burton, 1996; Pollock & Williams, 2001; Arie et al., 2008), impaired

interpersonal problem solving (O’Connor, R., O’Connor, D. et al., 2004; Pollock & Williams,

1998; Pollock & Williams, 2001; Williams, 1996; Goddard et al., 1996; Williams et al.,

2005), and lack of fluency for positive events (Hepburn et al., 2006; MacLeod et al., 1993;

1997; 2005; MacLeod & Cropley, 1995; MacLeod & Byrne, 1996; O’Connor, Connery, &

Cheyne, 2000).

123
 Over the last two decades, the conceptualisation of hopelessness has changed

somewhat. A growing number of evidence suggests that hopelessness is more than just an

expectation of more negative events instead; it is an expectation of fewer positive events

happening in someone’s future (MacLeod & Byrne, 1996). Until the beginning of the 1990’s,

the characterisation of hopelessness as a risk factor for suicidal behaviour was somewhat

vague. Despite numerous attempts to examine this multifaceted construct, hopelessness

remained as something whose pernicious effects have been seen and heard of, but was never

quite fully understood amongst clinicians and researchers alike. In 1993, MacLeod and his

colleagues devised a task-based measure called the “Future Thinking Task” in order to

explore the individual’s ability of to generate examples of positive and negative, personal

future events. The FT task’s procedure was originally based from the verbal future fluency

task (Lezak, 2004) except in the FT task; fluency was based on the generation of future

expectations or example of future events rather than words. The initial version of the FT task

involved asking participants to think of as many examples of future events as they can, across

various time periods (next week, next year, & next 5 – 10 years). The task was performed

under two conditions: (a) negative and (b) positive. In the positive condition, participants

were asked to think of examples of pleasurable future events (e.g. “things that they are

looking forward to”) while in the positive condition, they were asked to think of examples of

unpleasant future events (e.g. “things that they are not looking forward to”; MacLeod et al.,

1993). Findings from the initial use of the FT task indicated that the previously suicidal group

generated more examples of positive events than the control group. However, the groups did

not differ in their number of negative future events (MacLeod et al., 1993). Similar results

were found in MacLeod & Byrne’s study in 1996 on a sample of depressed individuals,

which indicated a markedly reduced fluency for positive events in the depressed group

compared to the control group. However, whereas the groups did not differ in the number of

124
negative events in MacLeod et al.’s study (1993), the depressed group showed more fluency

for negative events than the control group (MacLeod & Byrne, 1996). A year later, MacLeod

and colleagues (1997) found the same pattern of results that was illustrated in MacLeod &

Byrne’s study (1996). Intriguingly, an opposite pattern of results was revealed when the FT

task was conducted in a sample of non-clinical adolescents with symptoms of depression and

anxiety. Specifically, Miles, MacLeod and Pote’s study (2004) indicated that the participants

with greater depressive and anxiety symptoms exhibited more fluency for negative events

than the control group, but both groups did not differ in their fluency for positive events.

Despite the inconsistent findings, the link between depression and reduced fluency for

positive events as a proxy for hopelessness remained evident (MacLeod et al., 1996, 1998,

2005; Sidley, Calam, Wells, Hughes, & Whitaker, 1999). Then again, it is important to note

that a large number of these previous studies on future thinking and hopelessness have

focused mainly on healthy, or clinically depressed, sometimes in-patient, previously suicidal

individuals. Most of these studies also measured future fluency following identification of

suicidal behaviour (ideations & attempts), which suggest that interpretations are leaning

towards the idea that the lack of positive future fluency is a stable trait of previously suicidal

individuals. However, it also a fact that the life circumstances of the clinically depressed and

suicidal individuals are characterised by a number of emotional, social, and economic

difficulties, and traumatic events (Hawton et al., 2005; Isometsa, Heikkinen, Henriksson,

Aro, & Lonqvist, 1995; Leverich, Altshuler Frye, Suppes et al., 2003; O’Connor, 2011;

Rihmer, 2005, 2007). Altogether, these difficult circumstances represent a context that

renders positive future fluency impairment as a state-phenomenon in this particular sample.

Having combined the trait and state features of suicidality, the Differential Activation

Hypothesis of suicidal relapse (Lau et al., 2004) suggests a model that puts emphasis on the

role “cognitive reactivity” to hopelessness (trait) when in a sad mood (state). As already

125
discussed in chapter 1, the DAH framework proposes that the occurrence of a sad mood

determines the mechanism of suicidal/hopeless thoughts (Lau et al., 2004).

As a behavioural outcome that is closely linked with hopelessness, it is important for

clinicians to understand how positive future fluency responds to negative shifts in mood. It is

equally crucial to find out if positive future fluency remains reactive to mood changes

following a full recovery from the depression that facilitated the occurrence of hopeless

thoughts. By employing a mood priming technique, the present study will be able to test if the

assumption of the Differential Activation Hypothesis on hopelessness as a sad mood-

dependent cognition holds true for the first episode psychosis sample with a history of

suicidal attempt or deliberate self-harm. The present study will therefore explore if the

experience of psychosis will influence the relationship between mood and hopeless thoughts

as suggested by the DAH.

The first aim of this study is to examine the link between hopelessness and future

fluency in psychosis using the assumptions of the DAH of suicidal relapse (Lau et al., 2004).

Specifically, the present study seeks to examine the effect of mood on the positive and

negative future fluency of FEP individuals, with and without a history of suicidal attempt or

deliberate self-harm, by employing the mood challenge in order to evoke a sad mood in the

participants. The mood challenge is crucial in testing the assumptions of the DAH as the post-

induction future fluency will provide a valuable contrast to the future fluency prior to the

effects of the “induced” sadness. Whereas Williams and his colleagues have already tested

the DAH of suicidal relapse in both healthy and previously depressed samples with histories

of suicidal ideations, to date, this is the very first study to test the application of the DAH as a

model of suicidal vulnerability in a sample of FEP patients.

126
 The second aim of the study is to look into the effects of the mood challenge on the

perceived valence and likelihood of future events. The study also intends to look into the

effects of the mood challenge on the levels of momentary happiness and hopelessness.

The final aim of this study is to assess the use of the Leiden Index of Depression

Scale’s (Van der Does & Williams, 2003) hopelessness subscale as a measure of cognitive

reactivity to hopelessness and test if scores on this subscale will be associated with the pre- to

post-induction change in positive future fluency. As mentioned in chapter 4, the LEIDS’

hopelessness subscale was devised during the conception of the DAH for suicidal relapse in

order to measure the individual’s susceptibility to hopeless/suicidal thoughts when in a sad

mood (chapter 3).

5.1. Hypotheses
Effects of the Mood Challenge on Future Fluency

In keeping with the DAH, the impact of the sad mood induction procedure will be

more evident in the suicidal history group than the non-suicidal group. Specifically,

1. The suicidal history group will exhibit a greater pre- to post-induction decrease in the

number of positive events than the non-suicidal group. However, the degree of change in the

pre- to post-induction number of negative events will not differ between the two groups.

Effects of the Mood Challenge on the Perceived Valence and Likelihood of Future Events

Compared to the non-suicidal group, the suicidal history group will –

2. Demonstrate a greater pre- to post-induction decrease in the positive valence ratings on

good events (positive events) and a greater pre- to post-induction increase in the negative

valence ratings on bad events (negative events).

127
3. Exhibit a greater pre- to post-induction decrease in the likelihood ratings of good events

and a greater pre- to post-induction increase in the likelihood ratings of bad events.

Effects of the Mood Challenge on Happiness and Despondence ratings

In keeping with the assumption of the DAH,

4. The suicidal history group will reveal a greater pre- to post-induction decrease in happiness

ratings and a greater pre- to post-induction increase in despondence ratings.

The Validity of LEIDS as a Measure of Cognitive Reactivity to Hopelessness

Prior to the mood challenge, measurements of cognitive reactivity to hopelessness using the

LEIDS’ hopelessness subscale were taken. Measured CR to hopelessness will be tested using

the DAH framework. In line with the DAH,

5. The suicidal history group will exhibit greater CR to hopelessness, as measured by the

LEIDS, compared to the non-suicidal group.

The DAH suggests that the occurrence of low mood will trigger hopelessness. As a

behavioural feature of hopelessness,

6. The decline in fluency for positive events following the mood challenge will be associated

with greater levels of cognitive reactivity to hopelessness as measured by the LEIDS.

5.2. Method

5.2.1. Sampling

N.B. The sampling procedure of the present study was identical to the one employed

in the Experience Sampling Method study in Chapter 3.

128
5.2.2. Measures

N.B. Measures described in Chapter 3 (CHSF, LEIDS-R, CDSS, BHS, and ISST)

were also employed in this study.

Visual Analogue Scale - Mood Rating

N.B. The mood rating VAS described in the Chapter 4 was also employed in this

study.

Future Thinking Task (MacLeod et al., 1993)

The Future Thinking Task (MacLeod et al., 1993, 1998) is a verbal task where

participants are instructed to generate examples of personal experiences or events that they

think are likely to happen in their future. In the original FT task developed by MacLeod et al.

(1993), participants were asked to generate examples of future events in two different

conditions (positive and negative) over three different time periods (next week, next year, and

next five to ten years). In the positive condition, participants were asked to generate examples

of pleasurable future events, or “experiences that they were looking forward to”. In the

negative condition, participants were asked to generate examples of unpleasant future events,

or “experiences that they were not looking forward to”. In keeping with the version of the FT

task employed in Hepburn et al.’s (2006) mood-priming study, the FT task in this study was

conducted over four different time periods (next week including today, next month, next year,

and next 5 to 10 years) as opposed to the standard 3 time periods (next week, next year, &

next 5 to 10 years). The reason for this was to achieve an equal number of time periods for

the pre- and post-induction tasks. For example, the pre-induction FT task covers the next

week and the next month time periods while the post-induction FT task covers the next year

and the next 5 to 10-year time periods.

129
5.2.3. Procedure

5.2.3. a. Case Identification

The current study and the Means Ends Problem Solving task (Chapter 4) were both

conducted in a single testing session, using exactly the same sample. However, the sample

sizes of these two studies were slightly different as two participants opted out from the MEPS

task, but both agreed to complete the current study (MEPS study N = 97, FT study N = 99).

The two participants who opted out of the MEPS agreed to complete the FT task simply

because they felt that the FT task is less challenging than the MEPS task. As previously

discussed in the MEPS study (see chapter 5), the participants in this study were recruited

from the Early Intervention Service in Birmingham from March 2009 to March 2011. The

participants were split into two groups: (1) suicidal history group, and (2) non-suicidal group.

Care coordinators were informed about the inclusion and exclusion criteria of the study in

order to ensure that only eligible service users were approached. Following referral from the

care coordinators, participants were approached over the phone or in person, depending on

their preference. Once consent was obtained, the Columbia Suicidal History Form was

conducted to assess for the individual’s lifetime histories of suicide attempt or DSH. In

addition, the author of this study also carried out an audit on the participant’s clinical case

notes at EIS in order to check for any historical records of suicidality.

As explained in the testing procedures of the MEPS study in the previous chapter, the

testing session began with a briefing about the details of the study. Participants were given an

opportunity to ask questions and/or clarify any issues or concerns about their participation

and/or the nature of the study. Following this, a set of questionnaires measuring hopelessness,

depression, and suicidal thinking (BHS, CDSS, & ISST respectively) was completed. The

Future Thinking and the MEPS tasks (chapter 5), as mentioned earlier, were both carried out

on two occasions, once after the completion of questionnaires which was prior to the sad

130
mood induction procedure (pre-induction) and once immediately after the sad MIP (post-

induction). Having based the present investigation on Hepburn and colleagues’ study (2005),

the tasks were completed in exactly the same order for both pre and post-sad MIP. The MEPS

task was always presented first followed by the FT task. As mentioned in the previous

chapter, a debriefing was carried out at the end of the testing session for two main purposes:

(1) to discuss the actual purpose of the FT task (and the MEPS), and (2) to check if the

participant’s mood had returned to its normal level. Participants who remained upset at the

end of the testing session were offered a Happy Mood Induction Procedure to counteract the

effects of the sad MIP. Of the three participants who reported feeling upset, only two

completed the happy MIP. Participants were also informed that their care coordinators will be

requested to closely monitor on their mood for as long as they think it is necessary to do so.

This issue on risk overruling confidentiality was carefully discussed with the participants

prior to them signing the consent form. All three participants agreed to have their care

coordinators informed and all of them recovered well without any further deterioration of

their mental health.

5.2.3. b. Sad Mood Induction Procedure

As the FT task and MEPS tasks were both conducted in one testing session, the

participants undertook the same sad MIP procedure described in chapter 5.

5.2.3. c. Future Thinking Task

As the current study was a replication of Hepburn and her colleagues’ (2006) mood

priming study on future fluency, the FT task was also conducted in two blocks of trials, one

prior to and another one following the sad mood induction procedure. Each block of trials

contained equal numbers of conditions over four different time periods (Block A = positive

131
week, negative month, positive year, negative 5–10 years & Block B = negative week,

positive month, negative year, positive 5–10 years). Each participant was randomly allocated

to blocks A/B or B/A in order to counterbalance the presentation of conditions before (pre)

and after (post) the sad mood induction procedure. The time periods were presented one at a

time in chronological order. The participants were given a time limit of 1 minute to generate

as many future events as they can think of within the time period and condition specified by

the researcher. Examples of future events generated were then recorded by the researcher on

an FT task response sheet while making sure that the participant remained focused in

finishing the task. Upon completion of all time periods, the researcher read each example of

future events and asked the participant to rate it in two ways: (1) perceived valence, and (2)

likelihood. Valence ratings were obtained by asking the participants to rate how

positive/negative they would feel if the events were to actually happen using a 7-point Likert

scale (1 = not at all positive/negative & 7 = extremely positive/negative). Alternatively,

likelihood ratings were obtained by asking participants to rate the probability that their future

expectations were to occur using a similar 7-point Likert scale (1 = not at all likely & 7 =

extremely likely). In keeping with Hepburn et al.’s (2005) rating procedures, future

expectations were all rated first for positivity and then negativity. According to Hepburn’s

and her colleagues (2005), ratings for positivity (positive valence) and negativity (negative

valence) should not be performed concurrently as participants might rate negativity as an

inverse of positivity or vice versa. By rating them separately, we were able to measure

negativity and positivity as two separate dimensions of affect. A number of studies have

already demonstrated that negativity is not merely the opposite equivalent of positivity

(MacLeod & Byrne, 1996). However, it is important to note that the hypotheses of the present

study only examined the positive valence of good events and negative valence of bad events.

132
In keeping with hypotheses and for the sake of brevity, results and discussion were therefore

limited only to the valences specified previously.

Overall, a set of three scores were calculated upon completion of the future thinking

task (FT task). Calculations were based on the formulae used in Hepburn et al.’s study. First,

the total future fluency scores were calculated by summing the total number of future events

generated in each of the four time periods within the specified condition (positive or negative

FT task) and mood state (pre- & post-induction). Examples of future events across all time

periods must be unique and any repeated examples were only counted the first time they were

cited. Second, the valence scores were calculated by summing the total valence ratings

divided by the total number of future events generated within the specified condition and

mood. Third, following the calculation of the valence scores, the average likelihood scores

were calculated by adding the total likelihood ratings of each future event divided by the total

number of future expectations within the specified condition and mood state. The analyses in

the hypotheses testing focused mainly on the post-mood induction scores for future fluency

(positive & negative) and the associated features of future thinking (valence & likelihood).

The key objective was to compare the effects of the sad mood induction procedure on the

overall performance of the suicidal history group and the non-suicidal group in the FT task.

Although examples of future events were not identical on the pre- and post-mood induction

tasks (due to alternating positive & negative conditions within the 4 time periods),

comparisons on pre- and post-induction valence and likelihood scores were therefore

interpreted with caution. Instead of examining how each group of participants re-rated the

same events on two varying mood states, analyses in the current study were mainly focused

on comparing the general level of perceived valence (positive & negative) and likelihood

ratings between groups prior to and after the mood induction procedure.

133
5.2.4. Analysis Strategy

To test the hypotheses, a mixed between/within repeated measures analysis of

variance was conducted using an IBM SPSS Statistics software version 21 for Windows. In

order to control for the possible effects of the key clinical symptoms (e.g. generalised

hopelessness, depression, and suicidal thinking), an analysis of covariance using the repeated

measures design were conducted.

5.3. Results

5.3.1. Sample Characteristics

Of the 105 individuals who were approached, only 3 individuals responded with an

outright refusal. Out of the recruited 102 participants, only a subsample of 2 (1.96%) changed

their mind about participating in the study (those who previously completed the LEIDS

screening measure, but refused to do the mood-priming study). In total, the final sample

consisted of 99 participants of which, 49 (49.49%) had a lifetime history of suicidal

behaviour while 50 (50.51%) had no history at all of suicidal behaviour in their lifetime. Of

the 49 participants with a lifetime history of suicidal behaviour, 27 (55.10%) were males and

22 (44.90%) were females. Alternatively, of the 50 participants without a lifetime history of

suicidal behaviour, 35 (70%) were males and only 15 (30%) were females. Due to the

relatively small discrepancy in the sample size between the study discussed in the Chapter 4

(N = 97) and the present study (N = 99), the mean age and key symptom scores were almost

identical. However, for the sake of accuracy, the participants’ age and key symptom scores in

the present study are summarised in Table 16.

134
Table 16. Means and Standard Deviations of Age and Key Clinical Symptoms for the

Non-Suicidal Group and Suicidal History Group

Non-suicidal group (N = 50) Suicidal history group (N = 49)

Variable M SD M SD
Demographic
Age 23.86 4.95 23.08 4.65
Symptom Score
BHS 5.58 4.39 9.96 5.84
CDSS 1.70 2.46 3.88 3.99
ISST 0.38 1.51 1.73 3.10
Note: BHS = Beck Hopelessness Scale, CDSS = Calgary Depression Scale, and ISST = InterSept Scale for
Suicidal Thinking

5.3.1. Descriptive Statistics and T-tests

Generalised hopelessness

Current levels of generalised hopelessness were measured using the Beck

Hopelessness Scale. An independent t-test revealed a significant difference between groups,

with the suicidal (M = 9.96, SD = 5.84) group showing a higher level of generalised

hopelessness compared to the non-suicidal group (M = 5.58, SD = 4.39), t(97) = 4.07, p <

.001, d = .85. The BHS scores for each group are shown in Table 16.

Depression

Symptoms of depression were assessed using the 10-item Calgary Depression Scale

for Schizophrenia. Scores between the two groups were compared and an independent t-test

revealed a significant difference between the suicidal history group (M = 3.88, SD = 3.99)

and the non-suicidal group (M = 1.70, SD = 2.46), with the previous group showing higher

135
levels of depression than the latter group, t(97) = 3.21, p = .002, d =.66. The CDSS scores for

each group are shown in Table 16.

Suicidal Thinking

Levels of suicidal ideation during the past 7 days prior to testing were measured using

the InterSept Scale for Suicidal Thinking. Scores from both groups were compared using an

independent t-test, which revealed a significant difference between the suicidal history group

(M = 1.73, SD = 3.10) and the non-suicidal group (M = .38, SD = 1.51), with the previous

group showing higher levels of suicidal ideation than the latter group, t(97) = 2.92, p = .004,

d = .61. The ISST scores for each group are also shown in Table 16.

5.3.2. Hypotheses Testing

Following the data conversion in the previous chapter, square root data transformation

was also employed on all of the continuous dependent variables and covariates prior to

conducting the analysis of variance in order to satisfy the assumption of normality and

equality of variances. Again, in keeping with Tabachnick and Fidell (2007) and Howell

(2007), all means and standard deviations reported in this chapter were original values from

the untransformed data.

An alpha level of .05 was used for all statistical tests. All analyses were carried out

with group (suicidal history group vs. non-suicidal group) as a between-subjects factor.

Effects of the Mood Challenge on Future Fluency

In keeping with the DAH, the impact of the sad mood induction procedure will be more

evident in the suicidal history group than the non-suicidal group. Specifically,

136
1. The suicidal history group will exhibit a greater pre- to post-induction decrease in the

number of positive events than the non-suicidal group. However, the degree of change in the

pre- to post-induction number of negative events will not differ between the two groups.

To test this hypothesis, a two-step analysis was conducted. To test this hypothesis, a

three-step analysis was conducted. First, independent t-tests were conducted on future

fluency scores to test the difference between the means of each group in both conditions

(positive & negative events) conducted on two separate mood states (pre- & post-mood

induction). Future event scores were obtained by summing the total number of expectations

generated in each condition (positive & negative) during the two testing time points (pre- &

post-mood induction). Second, a mixed repeated measures ANOVA was carried out to

examine the effect of the sad mood induction procedure (sad MIP) on positive and negative

future fluency, and most importantly, to determine if the effect of the sad MIP differed

between the suicidal history group and the non-suicidal group. The hypothesis will be

validated on the basis of a significant interaction effect. To conduct the ANOVA, the variable

mood (pre- vs. post-induction future fluency scores) was entered as the within-subject factor

and group (suicidal history group vs. non-suicidal group) as the between-subject factor.

Third, an Analysis of Covariance using the repeated measure design was conducted in order

to test if the key clinical symptoms (e.g. generalised hopelessness, depression, and suicidal

thinking) have an effect of the mood x group interaction. To perform the ANCOVA, mood

(pre- vs. post-mood induction) was entered as the dependent variable, group (suicidal history

group vs. non-suicidal group) as the fixed factor, while generalised hopelessness (as

measured by the BHS), depression (as measured by the CDSS), and suicidal thinking (as

measured by the ISST) were entered as covariates. Separate ANCOVA’s were conducted for

each of the covariates to ensure better accuracy.

137
Positive events

Results of independent t-test on pre-mood induction positive Future Thinking Task

(FTT) scores showed that the suicidal history group (M = 8.86, SD = 2.43) had significantly

fewer number of positive events than the non-suicidal group (M = 10.02, SD = 2.68), t(97) =

2.10, p = .031, d = .44). Similar results were found from the independent t-test on post-mood

induction positive event scores as the suicidal history group (M = 6.78, SD = 2.18) had

significantly less number of positive events than the non-suicidal group (M = 9.26, SD =

3.73), t(84.70) = 3.55, p = .001; d =.72). The Levene’s test for the post-induction t-test

indicated unequal variances (F = 7.48, p = .007) so degrees of freedom were adjusted from 97

to 84.70. Summary of means and standard deviations are displayed in Table 17.

Consistent with the hypothesis, there was a significant group x mood interaction

effect with the suicidal history group showing a greater pre- to post-induction decrease in the

number of positive events compared to the non-suicidal group [F(1, 97) = 4.91, p =.029,

partial ƞ2 = .05]. This finding held true after controlling for generalised hopelessness (BHS),

depression (CDSS), and suicidal thinking (ISST).

The ANOVA also confirmed a significant within-subjects effect of mood as caused by

the decrease in post-induction number of positive events [F(1, 97) = 35.62, p <.001, partial ƞ2

= .27]. There was also a between-subjects main effect of group due to the suicidal history

group showing significantly fewer examples of positive events compared to the non-suicidal

group [F(1, 97) = 11.12, p =.001, partial ƞ2 = .10]. The pattern of results remained unaffected

after controlling for the previously identified key clinical symptoms.

Negative events

The independent t-test on the negative Future Thinking Task (FTT) scores showed no

significant difference between the suicidal history group and non-suicidal group on both the

138
pre-mood induction [mean (SD) = 7.49 (3.21) and 6.54 (3.22), respectively; t(97) = 1.59, p =

.114] and post-mood induction number of negative events [mean (SD) = 8.04 (3.56) and 6.92

(2.70), respectively; t(97) = 1.66, p = .101]. Table 17 displays summary of means and

standard deviations.

Also in agreement with the hypothesis, there was no significant mood x group

interaction effect [F(1, 97) = 3.30, p =.072]. There was also no between-subjects main effect

of group as the number of negative events did not differ between the suicidal history group

and non-suicidal group [F(1, 97) = .09, p =.080]. In contrast, there was a significant within-

subject effect of mood due to the decrease in the number of negative events following the

mood challenge [F(1, 97) = 4.30, p =.041, partial ƞ2 = .04]. The pattern of results was

unaffected following an ANCOVA to control for generalised hopelessness (BHS),

depression (CDSS), and suicidal thinking (ISST).

Table 17. Means and Standard Deviations of the Pre- and Post- Mood Induction Number of

Positive and Negative Events

Non-suicidal group (N = 50) Suicidal history group (N = 49)

Variable M SD M SD

Positive Events

Pre-induction 10.02 2.68 8.86 2.43

Post-induction 9.26 3.73 6.78 2.18

Negative Events

Pre-induction 6.54 3.22 7.49 3.21

Post-induction 6.92 2.70 8.04 3.56

139
 In summary, the results confirmed the hypothesis as the suicidal history group

exhibited a significantly greater pre- to post-induction decrease in the number of positive

events than the non-suicidal group. Also as predicted, both groups did not differ in their pre-

to post-induction changes in their number of negative events. Figure 5 illustrates the number

of positive and negative events for each group.

Effects of the Mood Challenge on the Perceived Valence and Likelihood of Future Events

Compared to the non-suicidal group, the suicidal history group will –

2. Demonstrate a greater pre- to post-induction decrease in the positive valence ratings on

good events (positive events) and a greater pre- to post-induction increase in the negative

valence ratings on bad events (negative events).

Following the three-step analysis in the hypothesis 1, independent t-tests, ANOVA,

and ANCOVA were conducted to test this assumption. The ANOVA and ANCOVA

variables were all identical to the ones used in hypothesis 1 with exception of the within-

subjects factor levels for mood as the current analyses requires pre- and post-induction

valence scores. The valence scores were obtained by asking the participants to rate how

positive or negative they would feel if their future expectations (positive & negative future

events) were to occur. The positive and negative valence ratings were calculated by summing

the individual valence ratings of each future event divided by the total number of future

events in each of the condition (positive & negative).

Positive Valence Ratings on Good Events

The independent t-tests showed that the suicidal history group had significantly lower

positive valence ratings on good events compared to the non-suicidal group on both the pre-

140
Figure 5. Average Pre- and Post-Induction Number of Positive Events (A) and Negative Events (B) for the Suicidal History Group and

Non-Suicidal Group

A B

141
mood induction [mean (SD) = 5.52 (.94) and 6.24 (.64), respectively; t(97) = 4.31, p <.001; d

= .83] and post-mood induction tasks [mean (SD) = 4.47 (1.03) & 5.72 (1.02), respectively;

t(97) = 5.90, p <.001; d = 1.19]. Summary of means and standard deviations are displayed in

Table 18 below.

As predicted, there was a mood x group interaction effect as caused by the

significantly greater pre- to post-induction decrease in the positive valence ratings on good

events in the suicidal history group compared to the non-suicidal group [F (1, 97) = 7.56, p

=.007, partial ƞ2 = .07]. The significant mood x group interaction held true after controlling

for depression (CDSS) and suicidal thinking (ISST). However, controlling for generalised

hopelessness (BHS) reduced the mood x group interaction effect to non-significance [F (1,

95) = 2.14, p =.147].

Similarly, the initially significant main effect of mood [F (1, 97) = 60.65, p <.001,

partial ƞ2 = .38] was also reduced to non-significance after controlling for generalised

hopelessness as measured by the BHS [F (1, 95) = 1.31, p =.255]. However, the significant

finding was unaffected after controlling for depression (CDSS) and suicidal thinking (ISST).

On the other hand, there was a significant between-subjects main effect of group as caused by

the considerably lower positive valence ratings on good events in the suicidal history group

compared to the non-suicidal group, and this held true after controlling for the previously

named key clinical symptoms [F(1, 97) = 37.70, p <.001, partial ƞ2 = .28].

Negative Valence Ratings on Bad Events

Results of independent t-tests indicated significantly higher negative valence ratings

on good events in the suicidal history group than the non-suicidal group on both the pre–

mood induction [mean (SD) = 1.65 (.80) & 1.30 (.82), respectively; t(97) = 2.54, p = .013; d

142
= .52] and post-mood induction tasks [mean (SD) = 1.85 (.94) & 1.47 (.81), respectively;

t(97) = 2.30, p = .024; d = .45]. Table 18 displays summary of means and standard deviations.

Again as predicted, there was a significant mood x group interaction effect due to the

suicidal history group exhibiting greater pre- to post-induction increase in the negative

valence ratings on bad events compared to the non-suicidal group [F(1, 97) = 7.20, p =.009,

partial ƞ2 = .07]. This finding remained significant after controlling for generalised

hopelessness (BHS), depression (CDSS), and suicidal thinking (ISST).

The results also revealed a significant within-subjects main effect mood as caused by

the significantly higher negative valence ratings following the post-mood induction [F(1, 97)

= 21.67, p <.001, partial ƞ2 = .18]. There was also a significant between-subjects main effect

of group as the suicidal history group had considerably higher negative valence ratings

compared to the non-suicidal group [F(1, 97) = 38.55, p <.001, partial ƞ2 = .28]. Again, the

results remained significant after controlling for the identified key clinical symptoms.

Table 18. Means and Standard Deviations of the Pre- and Post- Mood Induction Valence

Ratings on Positive and Negative Events

Non-suicidal group (N = 50) Suicidal history group (N = 49)

Variable M SD M SD

Positive Valence on Good Events

Pre-induction 6.24 .64 5.92 .94

Post-induction 5.72 1.02 4.47 1.03

Negative Valence on Bad Events

Pre-induction 1.30 .82 1.65 .80

Post-induction 1.47 .81 1.85 .94

143
 In summary, although the suicidal history group initially exhibited significantly

greater pre- to post-induction decrease in the positive valence ratings on good events than the

non-suicidal group as predicted, the groups failed to differ after controlling for generalised

hopelessness. In contrast, results of on negative valence ratings on bad events were consistent

with the hypothesis as the suicidal history group showed a greater pre- to post-induction

increase in the negative valence ratings on bad events compared to the non-suicidal group.

Figure 6 illustrates the pre- and post-induction positive valence ratings on good events and

negative valence ratings on bad events for the suicidal history group and non-suicidal group.

3. Exhibit a greater pre- to post-induction decrease in the likelihood ratings of good events

and a greater pre- to post-induction increase in the likelihood ratings of bad events.

To test the hypothesis, the three-step analysis employed in hypotheses 1 and 2 was

carried out. The ANOVA and ANCOVA variables match the ones used in hypothesis 1 and 2

with the exception of the within-subjects factor levels for mood as the current analyses

requires pre- and post-induction likelihood ratings. The likelihood ratings were obtained by

asking the participants how likely do they think their future expectations were to actually

happen. The overall likelihood score of good/positive events were calculated by summing the

individual likelihood ratings of all the good events divided by the number of good events

generated. The overall likelihood score of bad/negative events, on the other hand, were

calculated by summing the individual valence ratings of all the bad events divided by the

total number of bad events generated.

Likelihood Ratings on Good Events

Results of independent t-tests confirmed that the suicidal history group in general had

lower likelihood ratings on good events than the non-suicidal group on both pre-mood

144
Figure 6. Average Pre- and Post-Induction Positive Valence Ratings of Good Events (A) and Negative Valence Ratings of Bad Events (B) for

the Suicidal History Group and Non-Suicidal Group

A B

145
induction [mean (SD) = 5.01 (.94) & 6 (.67), respectively; t(81.01) = 5.91, p <.001; d = 1.19]

and post-mood induction tasks [mean (SD) = 3.98 (.81) and 5.27 (1.16), respectively; t(97) =

6.60, p <.001; d = 1.35]. Levene’s test indicated unequal variances on pre-mood induction t-

test (F = 6.89, p = .010) so degrees of freedom were adjusted from 97 to 81.01. Summary of

means and standard deviations are presented in Table 19.

As expected, mood x group interaction effect was found as the suicidal history group

showed a significantly greater pre- to post-induction decrease in the likelihood ratings of

good events [F(1, 97) = 5.42, p =.022, partial ƞ2 = .05]. The interaction effect remained

significant after controlling for generalised hopelessness (BHS), depression (CDSS), and

suicidal thinking (ISST). A significant main effect of mood was also found due to the

particularly lower likelihood ratings of good events following the mood induction [F(1, 97) =

98.06, p <.001, partial ƞ2 = .50]. Similarly, a significant between-subjects main effect of

group was found due to the considerably lower likelihood ratings in the suicidal history group

compared to the non-suicidal group [F(1, 97) = 3.50, p <.001, partial ƞ2 = .34]. The results

remained unaffected after the key clinical symptoms were controlled for.

Likelihood Ratings on Bad Events

Independent t-tests confirmed that the suicidal history group had higher likelihood

ratings on bad events or negative future expectations than the non-suicidal group on both pre-

mood induction [mean (SD) = 4.56 (1.20) & 3.36 (1.44), respectively; t(83.70) = 4.23, p

<.001; d = .84] and post-mood induction tasks [mean (SD) = 5.48 (1.13) and 4.53 (1.40),

respectively; t(97) = 3.65, p <.001; d = .75]. Levene’s test indicated unequal variances on

pre-mood induction t-test (F = 6.55, p = .012) so degrees of freedom were adjusted from 97

to 83.70. Table 19 displays summary of means and standard deviations.

146
 The predicted mood x group interaction effect was found as the suicidal history group

exhibited a significantly greater pre- to post-induction increase in the likelihood ratings of

bad events compared to the non-suicidal group [F(1, 97) = 4.00, p =.048, partial ƞ2 = .04].

This finding held true after controlling for generalised hopelessness (BHS), depression

(CDSS), and suicidal thinking (ISST).

A significant within-subjects main effect of mood was also found due to the

substantially lower likelihood ratings of bad events following the sad mood induction

procedure [F(1, 97) = 94.78, p <.001, partial ƞ2 = .49]. Likewise, a significant between-

subjects main effect of group was found as caused by the higher likelihood ratings of bad

events in the suicidal history group compared to the non-suicidal group [F(1, 97) = 18.16, p

<.001, partial ƞ2 = .16]. This pattern of results held true after controlling for the key clinical

symptoms.

Table 19. Means and Standard Deviations of the Pre- and Post- Mood Induction Likelihood

Ratings on Positive and Negative Events

Non-suicidal group (N = 50) Suicidal history group (N = 49)

Variable M SD M SD

Likelihood Ratings on Good Events

Pre-induction 6.00 .67 5.01 .94

Post-induction 5.27 1.16 3.98 .81

Likelihood Ratings on Bad Events

Pre-induction 3.36 1.44 4.56 1.20

Post-induction 4.53 1.40 5.48 1.13

147
 In summary, the pattern of results were consistent with the hypothesis as the suicidal

history group had a significantly greater pre- to post-induction decrease in the likelihood

ratings of good events and a significantly greater pre- to post-induction increase in the

likelihood ratings of bad events in comparison to the non-suicidal group. Figure 7

demonstrates the pre- and post-induction likelihood ratings for both good and bad events for

the suicidal history group and the non-suicidal group.

Effects of the Mood Challenge on Happiness and Despondence Ratings

In keeping with the assumption of the Differential Activation Hypothesis (DAH),

4. The suicidal history group will reveal a greater pre- to post-induction decrease in happiness

ratings and a greater pre- to post-induction increase in despondence ratings.

In keeping with the three-step analysis conducted in the previous hypotheses, first,

independent t-tests were conducted on the happiness and despondence ratings (as measured

by the Visual Analogue Scale) to test if the means of each group differed on both testing time

points (pre- & post-mood induction). Second, a mixed repeated measure ANOVA was carried

out to examine the effect the sad MIP on the happiness and despondence ratings and to check

if the effects of the sad MIP differed between groups. Third, an ANCOVA was conducted to

check for any statistically relevant covariates that could possibly account for the significant

interaction effect between the mood (dependent variable) and group (independent variable).

Again, the variables for the ANOVA and ANCOVA analyses were kept identical to the

variables in the previous analyses with the exception of within-subjects levels of mood as the

current hypothesis examines the pre- and post-induction mood ratings (happiness &

despondence).

148
Figure 7. Average Pre- and Post-Induction Likelihood Ratings of Good Events (A) and Likelihood Ratings of Bad Events (B) for the Suicidal

History Group and Non-Suicidal Group

A B

149
Happiness Ratings

Independent t-test on pre-mood induction happiness ratings confirmed no significant

difference between the means of the suicidal history group (M = 5.34, SD = 2.09) and the

non-suicidal group (M = 5.73, SD = 2.00), t(97) = .96, p = .377. In contrast, independent t-test

on post-mood induction happiness ratings showed a significant difference as the suicidal

history group (M = 4.09, SD = 1.71) in general scored lower than the non-suicidal group (M =

5.04, SD = 1.98), t(97) = 2.42, p = .017; d = .48). Summary of means and standard deviations

are shown in Table 20.

In agreement with the hypothesis, there was a significant mood x group interaction

effect due to the significantly greater pre- to post-induction decrease in happiness ratings in

the suicidal history group compared to the non-suicidal group [F(1, 97) = 4.38, p =.039,

partial ƞ2 = .04]. The same held true after controlling for generalised hopelessness (BHS),

depression (CDSS), and suicidal thinking (ISST).

There was also a significant within-subjects main effect of mood due to the decrease

in happiness ratings following the mood induction [F(1, 97) = 44.25, p <.001, partial ƞ2 =

.31]. The main effect of mood remained significant after controlling for the key clinical

symptoms. Interestingly, there was no significant between-subjects main effect on group

[F(1, 97) = 3.35, p =.070] therefore, no further analyses were conducted although the p-value

was only marginally over the conventional .05 significance level.

Despondence Ratings

Independent t-test on pre-mood induction momentary hopelessness ratings revealed

no significant difference between the suicidal group (M = 3.29, SD = 2.32) and the non-

suicidal group (M = 2.78, SD = 2.49), t(97) = 1.19, p = .239. In contrast, independent t-test on

post-mood induction hopelessness ratings showed the suicidal group (M = 4.95, SD = 2.38)

150
scoring significantly higher than the non-suicidal group (M = 3.38, SD = 2.46), t(97) = 3.01, p

= .003, d = .61). Table 20 displays summary of means and standard deviations for both

happiness and despondence ratings.

Also consistent with the hypothesis, there was a significant mood x group interaction

effect due to the significantly greater pre- to post-induction increase in despondence ratings

in the suicidal history group compared to the non-suicidal group [F(1, 97) = 4.90, p =.029,

partial ƞ2 = .05]. The interaction effect remained significant after controlling for generalised

hopelessness (BHS), depression (CDSS), and suicidal thinking (ISST).

Table 20. Means and Standard Deviations of the Pre- and Post- Mood Induction Happiness

and Despondence Ratings

Non-suicidal group (N = 50) Suicidal history group (N = 49)

Variable M SD M SD

Happiness Ratings

Pre-induction 5.73 2.00 5.34 2.09

Post-induction 5.04 1.98 4.09 1.71

Despondence Ratings

Pre-induction 2.78 2.49 3.29 2.32

Post-induction 3.38 2.46 4.95 2.38

The ANOVA showed a significant within-subjects effect of mood due to the increase

in despondence ratings following the mood induction [F(1, 97) = 37.08, p <.001, partial ƞ2 =

.28]. There was also a between-subjects main effect of group as caused by the higher

despondence ratings in the suicidal history group in comparison to the non-suicidal group

151
[F(1, 97) = 4.90, p =.029, partial ƞ2 = .05]. The main effects of mood and group remained

unaffected after controlling for the key clinical symptoms.

In summary, overall results were in agreement with the hypothesis as the suicidal

history group exhibited a significantly greater pre- to post-induction decrease in happiness

ratings and a significantly greater pre- to post-induction increase in despondence ratings in

comparison to the non-suicidal group. Figure 8 illustrates the fluctuation of momentary

happiness and despondence for the suicidal history group and the non-suicidal group.

The Validity of LEIDS as a Measure of Cognitive Reactivity to Hopelessness

Prior to the mood challenge, measurements of cognitive reactivity to hopelessness using the

LEIDS’ hopelessness subscale were taken. Measured CR to hopelessness will be tested using

the DAH framework. In line with the DAH,

5. The suicidal history group will exhibit greater CR to hopelessness, as measured by the

LEIDS, compared to the non-suicidal group.

As predicted, the suicidal history group (M = 12.67, SD = 5.20) showed significantly

higher cognitive reactivity to hopelessness as measured by the hopelessness subscale of the

Leiden Index of Depression Scale – revised version (LEIDS) than the non-suicidal group (M

= 6.40, SD = 4.32), t(97) = 6.21, p <.001, d = 1.31).

The DAH suggests that the occurrence of low mood will trigger hopelessness. As a

behavioural feature of hopelessness,

6. The decline in fluency for positive events following the mood challenge will be associated

with greater levels of cognitive reactivity to hopelessness as measured by the LEIDS.

To test this hypothesis, first, an overall positive future fluency difference score was

calculated by subtracting the pre-induction number of positive events from the post-induction

152
Figure 8. Average VAS Despondence (A) and Happiness (B) Ratings on Pre-Task, Pre-Induction, Post-Induction, and Post-Task Mood States

in the Suicidal History Group and Non-Suicidal Group

A B

153
number of positive events. Second, a bivariate correlation was carried out on the overall

difference score of positive future fluency and scores from the LEIDS’ hopelessness subscale.

Contrary to the hypothesis, there was no significant correlation between the difference

score of positive future fluency and scores on LEIDS’ hopelessness subscale (r = .01, N = 99,

p = .936). In other words, the pre- to post-induction change in fluency for positive events was

not associated with the level of cognitive reactivity as measured by the LEIDS’ hopelessness

subscale.

5.4. Discussion

The main aim of this study was to investigate the application of the DAH framework

in understanding the suicidal thinking process of FEP individuals with or without a lifetime

history of suicidal attempts or DSH. Specifically, the intent was to examine if the future

fluency of those with a history of suicidal behaviour was influenced by the subtle changes in

mood as caused by the sad MIP. A number of previous studies have indicated that the lack of

positive future fluency is strongly associated with hopelessness, which is a key risk factor for

suicidal behaviour (MacLeod & Byrne, 1996; MacLeod, et al., 1997, Hunter & O’Connor,

2003; Hepburn et al., 2006). The design of this study was based from the previous mood

priming study on future fluency conducted by Hepburn and her colleagues in 2006.

In agreement with the key hypothesis of this study, the shift in mood as caused by the

sad MIP significantly reduced the positive future fluency or the individual’s ability to

generate examples of positive/good future events, with a particularly more pronounced effect

in the suicidal history group than the non-suicidal group. In contrast, although the shift in

mood increased the negative future fluency or the ability to generate examples of negative/bad

future events of the entire sample following the sad MIP, the degree of pre- to post-induction

154
change did not significantly differ between groups as expected. This pattern of results was

consistent with the data from Hepburn et al.’s (2006) study, which indicated a reduced

positive future fluency and an unchanged negative future fluency following the sad MIP in a

sample of non-depressed participants. Such a pattern of results was also found in a number of

studies that examined future fluency as a behavioural feature of hopelessness in individuals

who were suicidal and depressed (MacLeod et al., 1993; MacLeod, Tata, Kentish, &

Jacobsen, 1997; MacLeod, Pankhania, Lee, & Mitchell, 1997; MacLeod & Salaminiou, 2001;

Conaghan & Davidson, 2002; Hunter & O’Connor, 2003). According to MacLeod and his

colleagues (2005), the lack of pre- to post-induction change in the number of negative events

may be attributed to the ceiling effect on the affective impact of the bad events in general.

Due to the limited affective impact of the individual positive events, the intensity of its

collective impact largely relies on its quantity. Unlike the positive events, the affective impact

of the individual negative events is more intense and thus, requiring less to achieve its

maximum possible effect or “impact threshold”. In line with this idea from MacLeod et al.

(2005), it is possible that the generation of negative future events did not vary between the

suicidal history group and the non-suicidal group across the mood conditions (pre- & post-

induction) simply because the “impact threshold” of the negative events has already been

reached prior to the mood challenge. It is also possible that the pre- and post-induction

number of negative events did not differ due to the aversive nature of the events overall.

Aversive events are likely to be perceived as more negative in terms of its affective impact. In

order to further explore this possibility, a thorough examination of the score sheets for the

negative future thinking task (pre- and post-sad MIP) was carried out. Based on careful

observation, issues about mental health emerged as the predominant theme of the negative

future events generated before and after the mood induction. Issues surrounding mental health

155
mainly included relapsing (e.g. being sectioned & hospitalisation), medication (e.g. ‘being on

it’ & ‘not being able to get off it’), worsening of other related symptoms (e.g. ‘becoming more

depressed, anxious, paranoid, or afraid’ & ‘voices getting worse’), not recovering from their

mental illness (e.g. ‘being stuck’, ‘being the same’, & ‘being just as I am now’), isolation (e.g.

‘losing contact with friends’ & ‘not having any friends’), and having the stigma (e.g. ‘being

seen as different’, or ‘not being normal’). Jackson and colleagues (2004) have indicated that

hospitalisation and treatment experiences during the initial episode of psychosis were

predictive of post-traumatic stress. It is possible that due to the aversive nature of these

negative events, the extent to which individuals can tolerate the emotional impact of these

events was already at its maximum prior to the mood challenge. It is also possible that the

groups did not differ in their fluency for negative events simply because they both shared

comparable worries and fears regarding their future mental health.

On the other hand, a number of studies have suggested that the impaired fluency for

positive events in the suicidal and depressed individuals were mainly due to the elevated

feelings of hopelessness (MacLeod et al., 1993; MacLeod, Tata, et al., 1997; MacLeod,

Pankhania, et al., 1997; Sidley et al., 1999; Hepburn, et al., 2006). Recalling the assumptions

of the DAH, low mood is believed to be linked together with feelings of hopelessness through

repeated episodes of depression. The stronger the link, the easier it becomes for low mood to

reactivate these feelings of hopelessness. While previous studies have already established the

link between the lack of positive future fluency and hopelessness in suicidal and depressed

individuals, the present study indicates that the fluency-hopelessness link is also evident in

FEP individuals with histories of suicidal attempts and DSH. The evidence of such a link

supports the assumptions of the DAH, which suggests that once hopelessness is already

embedded within the network of negative thinking process, even a slight dampening in mood

156
can impair the individual’s fluency for future events. More importantly, further analyses

showed that this observed impairment in fluency for positive events following the sad MIP

was not merely a marker of current generalised hopelessness (BHS), depression (CDSS), and

suicidal thinking (ISST). However, the fact that life circumstances following the initial

psychotic episode can be very traumatic for many individuals (Harrison & Fowler, 2004;

Jackson & Iqbal, 2000; Jackson, Trower, Reid et al. 2009; Riedesser, 2004; Tarrier et al.,

2007), there remains a possibility that these particularly distressing contextual factors

moderated the effect of the mood challenge on positive future fluency in this particular

sample. It is reasonable to speculate that individuals who are in distress are less likely to have

a positive view of their future. In a study conducted by O’Connor and Cassidy (2007), they

found that distress was strongly linked with reduced fluency for positive events in high-stress

optimists and low-stress pessimists.

Contrary to the hypothesis, the suicidal history group and the non-suicidal group did

not differ on the positive valence ratings of good events following the mood challenge.

Although the initial analysis indicated that the suicidal history group showed significantly

lower positive valence ratings of good events following the mood induction compared to the

non-suicidal group, controlling for generalised hopelessness reduced the initial finding to non-

significance. Intriguingly, controlling for generalised hopelessness also caused the overall

positive valence ratings of good events before and after the mood induction (within-subjects

main effect) not to differ. However, the positive valence ratings remained different between

the two groups (between-subjects main effect). This pattern of results seemed to suggest that

the perceived valence of future good events was not sensitive enough to the subtle changes in

transitory mood. According to the Hopelessness theory (Abramson, Metalsky, & Alloy,

1989), generalised hopelessness is sustained by a faulty thinking processes (i.e. magnification

157
& overgeneralization), which in this case has possibly caused the perceived valence appraisal

of good events to operate in an entirely systemic and stable mode as opposed to reactive. It is

worthy to note, however, that despite the lack of difference in the post-induction positive

valence ratings of good events between groups, the suicidal history group exhibited lower

positive valence ratings than the non-suicidal group. Consistent with the hypothesis, the

suicidal history group had considerably higher negative valence ratings of bad (or negative)

events following the sad MIP. Despite the lack of change in the number of bad events

following the mood induction, this finding seemed to suggest that the transitory shift in mood

caused the participants to perceive negative events as more unpleasant. In keeping with the

assumptions of the DAH, the subtle dampening in mood triggered a hopeless thinking style,

which in this case was a more negative perception of the affective impact of bad events.

Overall, this finding is consistent with the “impact threshold” that was discussed earlier. In

the initial speculation it was suggested that the number of post-induction bad events did not

differ due to the possibility that the perceived affective impact of bad events was already at its

maximum prior to the mood induction. The observed increase in the perceived negative

valence of bad events, however, suggest otherwise. This finding seems to suggest that the

“impact threshold” was only facilitated by the shift in transitory mood, which occurred as an

effect of the mood induction. The increase in the perceived negative valence of bad events

following the mood induction enhanced the perceived affective impact of the bad events

collectively, which justifies the lack of difference in the pre- and post-induction number of

bad events. In other words, the number of bad events generated across mood states (pre- &

post-induction) did not differ simply because the overall affective impact of bad events, albeit

the quantity was unchanged, was perceived to be a lot more unpleasant.

158
 As predicted, the suicidal history group had significantly lower likelihood ratings for

good events and higher likelihood ratings for bad events. This pattern of results was

consistent with the findings of previous studies that examined the link between the depressed

mood and future thinking in non-psychosis sample, which indicated that a depressed mood

was associated with lower likelihood of good/positive events (MacLeod & Cropley, 1995)

and higher likelihood of bad/negative events (Andersen, Spielman, & Bargh, 1992; MacLeod

et al., 1997). Similarly, a study on the previously suicidal but non-psychosis sample indicated

that lower likelihood ratings were significantly associated with generalised hopelessness as

measured by the BHS (MacLeod et al., 2005). One of the possible explanations for this mood-

linked perception of likelihood was explained in Tversky and Kahneman’s (1974) study on

heuristics and biases when judging the likelihood of uncertain events. According to Tversky

and Kahneman (1974), the likelihood of an uncertain event occurring is determined by the

perceived ease with which an individual can remember occurrences of similar events, a

judgemental heuristic that is also referred to as “availability”. If such ease of recollection is

facilitated by the mood congruence effect (Bower, 1981), it is therefore logical that the

suicidal history group, as being more vulnerable to the effect of the mood challenge,

remembered more events whose emotional content matches the sad emotional state that they

were in. In other words, individuals who are in a negative or sad mood will tend to rate the

likelihood of a good event as less likely simply because it is harder for them to recall

occurrences of similar events whose emotional content is in conflict with their current mood.

It is for this exact reason why the likelihood of bad events was greater in the suicidal history

group. Due to the negative shift in mood following the mood challenge, bad events were seen

as more likely simply because it was easier for the individuals to recall occurrences of similar

events whose emotional content matches their negative or sad mood state.

159
 As also hypothesised, following the mood challenge, the suicidal history group

exhibited lower levels of happiness ratings and higher levels of despondence ratings as

measured by the Visual Analogue Scale than the non-suicidal group. This pattern of results

were consistent with that of the mood priming study conducted by Hepburn and her

colleagues (2006), who found that the sad mood induction procedure was an effective method

to alter levels of happiness and despondence in a sample of non-depressed volunteers.

Interestingly, the suicidal history group and the non-suicidal group did not differ on their

levels of happiness and despondence prior to the mood challenge. The similarity in the

groups’ pre-induction mood ratings suggests that there were no pre-existing group differences

that could have biased the data in favour of the suicidal history group (e.g. as being more

despondent and less happy than the non-suicidal group). As the DAH focuses on the

individual’s cognitive vulnerability to hopelessness, it was crucial that the results of

independent t-tests have established that the greater degree of change in pre- to post-induction

mood ratings in the suicidal-history group was not simply due to the pre-existing vulnerability

to hopeless thinking during the pre-induction stage. Due to this, it was easier to determine that

the degree of change in the pre- to post-induction future fluency was mainly due to the

individual’s cognitive reactivity to hopelessness when in a sad mood and not simply due to

the worsening of a pre-existing vulnerability or mood state.

This pattern of data suggests that although the suicidal history group had significantly

higher levels of generalised hopelessness than the non-suicidal group on the whole, the

suicidal history group’s momentary feelings of despondence were more differentially active.

This finding is in keeping with the results of the ESM study in chapter 3, which revealed that

compared to the non-suicidal group, the suicidal history group had a greater fluctuation of

momentary hopelessness in response to the shifts in transitory mood (decrease of positive

160
affectivity & increase of negative affectivity) caused by the relatively stressful minor events

in everyday life.

As expected, the suicidal history group exhibited higher cognitive reactivity to

hopelessness as measured by the hopelessness subscale of the Leiden Index of Depression

Scale – revised version (LEIDS). Intriguingly, however, the CR to hopelessness as measured

by LEIDS’ hopelessness subscale did not correlate with the pre- to post-induction change in

positive future fluency. The lack of association between CR to hopelessness and change in

positive future fluency in the non-suicidal group, whose number of positive events was also

significantly altered following the mood challenge, is not particularly easy to explain. It is

possible that the lack of association was due to the fact that experimentally induced changes

fluency for positive events did not accurately represent suicidal vulnerability in real-life

situations thus, not showing a detectable link with CR to hopelessness as measured by the

LEIDS’ hopelessness subscale. It is also possible that the hopelessness subscale of the LEIDS

was simply not able to effectively capture the key elements that embody cognitive

vulnerability to hopelessness in this particular sample.

In summary, the results of this study indicate that overall, the suicidal history group

had significantly fewer positive events to look forward to than the non-suicidal group, which

is in agreement with previous studies. More importantly, the data from the present study also

suggest that the change in positive future fluency in the suicidal history group is a marker of

greater sensitivity to the subtle changes in mood following the mood challenge, which

confirms the assumption of the Differential Activation Hypothesis (DAH). As expected, the

mood challenge did not alter the negative future fluency in both groups, which was also

illustrated in the findings of a similar mood priming study (Hepburn et al., 2006). Whereas

the suicidal history group failed to exhibit less positive valence of good events, the group

161
exhibited greater negative valence of bad events as predicted. However, results on the

likelihood ratings showed a more consistent pattern of results with previous studies as the

suicidal history group illustrated lower likelihood ratings for good events and higher

likelihood ratings for bad events.

In keeping with the DAH, the suicidal history group exhibited notably reduced

momentary feelings of happiness and substantially elevated momentary feelings of

despondence in response to the sad mood induction procedure. Results on the use of the

LEIDS’ hopelessness subscale as a measure of cognitive vulnerability to hopelessness

produced a mixed pattern of results. As hypothesised, the suicidal history group had

significantly higher levels of CR to hopelessness than the non-suicidal group. The CR to

hopelessness as measured by LEIDS’ hopelessness subscale, however, did not correlate with

the pre- to post-induction change in positive future fluency as expected.

All in all, the results of this study extend the relevance of the DAH of suicidal relapse

from being a model of suicidal relapse in a previously depressed sample to a potentially

workable model of suicidal vulnerability in a sample whose diagnosis is psychosis. It also

adds an important contribution to the literature by illustrating the DAH as a valid cognitive

model of suicidal vulnerability in psychosis that can be tested via a concrete behavioural

marker (e.g. future fluency).

5.4.1. Strengths and limitations

The results of the present study are subject to a number of limitations. The fact that the

FT and the MEPS tasks were both conducted in one single testing session, it means that the

present study shares the same methodological limitations that were discussed in great detail in

the MEPS study (chapter 5). Whereas the absence of a neutral or a happy mood induction did

162
not have an unfavourable effect in the results of the present study, understanding the link

between hopelessness and different mood states might be of significant value for future

clinical work. The lack of difference between groups on the positive valence ratings of good

events was also not in agreement with the previous similar study (Hepburn et al., 2006).

Finally, given that mood challenge was intended to induce subtle changes in mood, the results

of this study must be interpreted with caution. Although the ESM study (chapter 3) suggests

that the DAH is ecologically valid and the results of the present study are consistent with

ESM data, a follow-up study will provide a valuable confirmation if the observed suicidal

vulnerability as measured by the lack of fluency for positive events following the mood

induction will predict a future suicidal behaviour in real life. Where there are a number of

limitations, there are also a number of strengths to this study. To the best of the author’s

knowledge, this is the first study to have examined the suicidal thinking process in psychosis

using the mood priming technique. The sample size of the study is also seen as one its

strengths. Compared to the previous studies (e.g. Hepburn et al., 2006 where N = 52 non-

depressed volunteers; Williams et al., 2007 where N = 32 volunteers with & without histories

of depression), the sample of 99 is relatively large, especially given a clinical group that is

often not easy to engage, let alone recruit for a study that can be potentially upsetting or

emotionally challenging. Also, the culturally diverse population of Birmingham made it

possible for this study to obtain a sample with a good mix of ethnicity and social backgrounds

(i.e. religion & family structures) underlining the generalizability of the findings.

Given that both problem solving impairment (as measured by the MEPS task) and

reduced fluency for positive events (as measured by the FT task) are considered as

behavioural outcomes closely linked with hopelessness, the clinical implications of the

present study are therefore very much comparable to the study in the chapter 4 (MEPS

163
study). It remains the case that the management of suicidal behaviour in young people with

psychosis is difficult and challenging for many clinicians. The results of this study, offer a

potentially effective way to explore the mechanism of suicidal relapse in psychosis.

Specifically, by employing the mood priming technique and the FT task to illustrate the

differences between the suicidal history group and the non-suicidal group, the present study

was able to demonstrate two important things: (1) that mood challenge is a safe and effective

mood priming technique even for a sample of individuals with psychosis and histories of

suicidal behaviour, and (2) that positive future fluency as a behavioural marker of

hopelessness may be of potentially useful value for future studies on suicidal behaviour in

early psychosis.

164
 CHAPTER 6

General Discussion

6.0. Introduction

The overarching aim of this thesis is to investigate the mechanism of suicidal thinking

in early psychosis when the suicide rate is at its highest. The core objective is to examine if

the recurrence of suicidal or hopeless thoughts over time can be understood within the

framework of the DAH of suicidal relapse (Lau et al., 2004). In order to achieve this, this

thesis employed two contrasting methodologies: (1) the ecological approach of the ESM, and

(2) the experimental approach of the sad mood induction procedure. The overall results from

this thesis support the DAH of suicidal relapse (Lau et al., 2004), and provide

recommendations for the application of the DAH framework in order to further explore

suicidal thinking in early psychosis.

6.1. Summary of findings

In the initial chapter it was noted that there was a lack of theoretical model to explain

suicidal behaviour in general psychotic disorders, especially in FEP when the risks of suicide

are greater (Brown, 1997; Harris & Barraclough, 1997; Palmer et al., 2005). Despite the

enormous amount of information about the risk factors of suicidal behaviour in early

psychosis, there is a limited amount of information about the underlying mechanisms of the

suicidal thinking process in this clinical group. In order to address this gap, the framework of

the DAH of suicidal relapse (Lau et al., 2004) was employed. The central idea of this

hypothesis suggests that once suicidal or hopeless thoughts are featured in an earlier

165
depressive episode, these hopeless thoughts form a link with the depressed mood, along with

the network of maladaptive cognition. Repeated depressive episodes strengthen this link such

that subsequent occurrences of depressed mood will trigger these hopeless/suicidal thoughts.

To empirically test the differential activation process, two contrasting methodologies were

employed. In chapter 3, the ESM (de Vries, 1992) was conducted in order to capture the

differential activation of hopeless or suicidal thoughts in the context of the individual’s

everyday life. The ESM is a systematic diary keeping method, which requires individuals to

fill in a self-report questionnaire at predetermined times of the day within his/her real-life

environment (de Vries, 1992). The key advantage of the ESM is that it measures key variables

of interest in real-life contexts as they occur. Overall, the findings of the ESM study were

largely in line with the hypotheses. In comparison to the non-suicidal group, the suicidal

history group exhibited higher levels of momentary hopelessness in everyday life as expected.

The suicidal group also exhibited greater momentary hopelessness linked to negative

affectivity (NA) and reduced hopelessness linked to positive affectivity (PA), compared to the

non-suicidal group. When confronted with unpleasant events, the suicidal group had a greater

increase in momentary hopelessness and NA, and a greater decrease in PA, compared to the

non-suicidal group. However, when confronted with challenging activities, the suicidal

history group exhibited greater NA than the non-suicidal group. However, the groups did not

differ in their momentary hopelessness and PA when faced with difficult activities. In the

discussion it was noted that there were a lack of structured activities in this group on a day to

day basis, and the main events of their typical weekly routines were face to face

conversations, telephone calls, or visits by family members or friends, which may explain

why event-related stress had more meaningful interactions with affectivity and momentary

hopelessness.

166
 As expected, measures of cognitive reactivity to hopelessness were found to be higher

in the suicidal history group than the non-suicidal group. Cognitive reactivity to hopelessness

was measured using the hopelessness subscale of the Leiden Index of Depression Scale –

revised version (Van der Does & Williams, 2003). Consistent with the hypothesis, higher CR

to hopelessness as measured by the LEIDS’ hopelessness subscale was found to be predictive

of the individual’s susceptibility to momentary hopelessness when affectivity is negative.

Similarly, higher CR to hopelessness as measured by the LEIDS was also found to be

predictive of the individual’s propensity to momentary hopelessness when faced with

unpleasant events. However, CR to hopelessness as measured by the LEIDS did not predict

momentary hopelessness when faced with stressful activities. Overall, the pattern of results

suggests that momentary hopelessness is more strongly linked with NA than PA, which is

consistent with the assumption of the DAH for suicidal relapse.

In chapters 4 and 5, the sad mood induction procedure was conducted in order to test

the differential activation of hopeless or suicidal thoughts by inducing individuals to certain

feelings of sadness, prior to being re-tested using the same sets of behavioural tasks from

baseline (prior to the mood challenge). The first task was the Means-Ends Problem Solving

task (MEPS; Platt & Spivack, 1975). The MEPS task is a verbal task, which was devised to

measure interpersonal problem solving ability. Previous studies have shown that an impaired

problem solving ability is a behavioural feature of hopelessness (Pollock & Williams, 2001;

Sadowsky & Kelly, 1993; Schotte & Clum, 1982). The purpose of the mood challenge was to

test if the change in mood will alter the interpersonal problem ability as suggested by the

DAH framework. The results of the study were consistent with this hypothesis as the suicidal

history group exhibited a more impaired problem solving performance following the mood

challenge. Further, compared to the non-suicidal group, the suicidal history group also

167
exhibited higher levels of despondence and lower levels of happiness as measured by the

Visual Analogue Scale (McCormick, Horne, & Sheather, 1988) following the sad mood

induction. However, contrary to the hypothesis, individuals’ CR to hopelessness as measured

by the LEIDS’ hopelessness subscale was not correlated with the pre- to post-induction

change in problem solving ability. Overall, the pattern of data from this problem solving study

replicated the results of the ESM, which indicated that negative affectivity reactivates

hopeless thoughts. Intriguingly however, the lack of association between vulnerability to

hopelessness (or the pre- to post-induction change in problem solving ability) and CR to

hopelessness (as measured by the LEIDS’ hopelessness subscale) is in conflict with the

apparent link between the LEIDS and momentary hopelessness when affectivity is negative.

Following the MEPS task, the Future Thinking (FT; MacLeod et al., 1993) task was

conducted. The FT task is also a verbal task, which was devised to measure fluency for future

expectations (also referred to as future events). Previous studies have illustrated that the lack

of fluency for positive events is also a behavioural feature of hopelessness (MacLeod et al.,

1993). As predicted, the suicidal history group had significantly fewer numbers of positive

events to look forward to than the non-suicidal group, which is in agreement with the findings

of previous studies. More importantly, the data from the present study also indicated that the

subtle downward shift in mood significantly reduced the positive future fluency in the suicidal

history group, which confirms the assumption of the DAH of suicidal relapse. As expected,

the downward shift in mood did not alter the negative future fluency in both groups, which

was also illustrated in the findings of a similar mood priming study (Hepburn et al., 2006).

Whereas the suicidal history group failed to exhibit less positive valence ratings for good

events, the suicidal history group exhibited greater negative valence ratings for negative

events as predicted. On the other hand, the data on the likelihood ratings showed a more

168
consistent pattern of results with previous studies as the suicidal history group illustrated

lower likelihood ratings for positive events and higher likelihood ratings for negative events

(Andersen et al., 1992; MacLeod et al., 1997; MacLeod & Cropley, 1995). Overall, the

pattern of data from this future thinking study was in keeping with the results of the problem

solving and ESM study, which indicated that the mechanism of suicidal/hopeless thoughts is

mood-dependent.

In summary, the results from the mood priming and ESM studies have both confirmed

that previously suicidal individuals are more “differentially active” to suicidal or hopeless

thoughts when in a low or negative mood, compared to the non-suicidal individuals.

However, whereas the CR to hopelessness as measured by the LEIDS’ hopelessness subscale

predicted vulnerability to momentary hopelessness when affectivity is negative in everyday

life, the vulnerability to hopelessness as measured by the MEPS and FT tasks did not correlate

with the CR to hopelessness as measured by the LEIDS. The reason for this remains unclear

and only further research will help establish the validity and reliability of the LEIDS’

hopelessness subscale as a measure of CR to hopelessness.

6.2. Limitations

The studies reported in this thesis are the first to have taken both the experimental and

ecological approach, to investigate the suicidal thinking process in FEP using the DAH of

suicidal relapse framework. For this reason, these studies only represent the starting point for

further investigation of the suicidal thinking mechanism in psychosis. Specifically, there are

three areas they could extend. First, future research could employ a follow-up study in order

to examine if the observed vulnerability to hopeless or suicidal thoughts (as measured from

either the ESM, or behavioural problem solving & future fluency tasks) will be predictive of

169
subsequent suicidal behaviours in real life. The studies conducted in this thesis were a

combination of cross-sectional (chapter 4 & 5) and longitudinal (chapter 3) methods, and the

inclusion of a follow-up was simply not feasible due to time constraints.

Second, future research could investigate the suicidal thinking process in other

psychosis populations. The studies in this thesis were restricted to FEP individuals only. Due

to the particularly high incidence of suicidal behaviour during this early stage of the illness,

the results may not be transferrable to individuals who are at a much later stage of the

psychotic illness.

Third, future research could explicitly investigate suicidal relapse via the ESM by

adding items that are specifically formulated to measure the severity of suicidal ideation and

intent. The questionnaire employed in the ESM study in this thesis was only limited to

assessing hopeless thoughts and feelings. The addition of items that specifically measures the

severity of suicidal thinking and intent could help uncover the extent of the relationship

between mood and hopelessness, and the contextual factors that can potentially trigger

suicidal relapse in everyday life.

6.3. Observations from the Research: Recommendations for future studies

on suicidality in psychosis

In this thesis it has been suggested that the interaction between the individual and

his/her natural context is crucial in understanding the underlying mechanism of suicidal

thinking. One of the important issues that arose from employing the ESM was that a number

of participants found the diary keeping task inconvenient and slightly irritating. The 6-day

duration of the ESM study and the daily frequency of sampling (total = 10) were perceived to

be quite intrusive and challenging. In the debriefing, when participants were asked if they

170
would take part again in a similar study, only a few were keen to do it. As noted in chapter 3,

the monetary incentive seemed to be the main motivation for taking part in the study. As

much as monetary incentives helped in the recruitment, there was no guarantee whether the

task was completed to the best standard possible, or simply to a standard that was acceptable

enough to earn the incentive. This speculation was mainly based on the fact the average

number of valid ESM reports per participant (59%) was slightly lower than the reported

compliance rate in other ESM studies of psychosis (66%; Oorschot et al., 2009). However,

there was also a possibility that due to the distressing and traumatic experiences following the

initial episode of psychosis (Harrison & Fowler, 2004; Jackson & Iqbal, 2000; Jackson et al.

2009; Riedesser, 2004; Tan et al., 2012; Tarrier et al., 2007), the participants were less able to

cope with a demanding task such as the ESM.

Having taken all of the methodological issues of the ESM into consideration, it is

possible that these issues will present potential ethical and practical difficulties for future

research. However, there are ways to minimise the difficulty of the ESM in this particular

group. First, the number of questions in the ESM diary could be simplified by focusing solely

on the mood, hopelessness, and contexts (i.e. people, places, & activities/events). A

questionnaire that is more straightforward and quicker to complete might reduce the “burden”

of doing it more frequently. Second, the use of electronic devices (i.e. PDA’s or smart

phones) could offer a more efficient way of filling in the ESM questionnaires. The option to

customise the sampling signals or prompts, from the irritating beeping sound of a digital

wristwatch to a more discrete mode in PDA’s, might present a more attractive diary keeping

method to the participant. Although previous electronic ESM studies have indicated that some

of the participants found the use of handheld devices slightly difficult (Kimhy et al., 2006),

the compliance rate was increased and overall feedback was positive (Graholm et al., 2008).

171
Finally, the irritation from frequent sampling could also be minimised by decreasing the

sampling frequency per day (e.g. 6 samplings per day instead of 10) and increasing the

duration of the ESM study (e.g. 10 days instead of 6). By doing this, equal number of ESM

reports (60) are generated at a less intense sampling rate, albeit over a longer period of time.

On the other hand, the only issue that arose from the mood priming studies was the

risk of residual feelings of unhappiness at the end of the testing session. Although it was

noted that only 3 of the 99 individuals who undertook the sad mood induction procedure

(chapter 4) reported some residual feelings of sadness at the end of the testing session, it is

possible that this could present a more serious issue in future research. This issue is

particularly crucial if the study involves individuals who are at higher suicidal risk (i.e.

previous attempters or self-harmers). Whereas it is difficult to predict the impact of the mood

challenge on an individual level, there are ways to minimise the risks of residual effects from

escalating into feelings of hopelessness. First, a happy mood induction could be offered to

counteract the effects of the sad mood induction procedure. Teasdale, Taylor, and Fogarty

(1980) have demonstrated the effectiveness of such procedure in inducing feelings of elation

to facilitate retrieval of happy memories. Second, frequent monitoring could be coordinated

with the participant’s care team. Third, as a responsible researcher, a leaflet with information

about agencies/organisations that could be contacted during out of working hours should be

given to the participants at the end of the session.

6.4. Clinical Implications

Results of the mood priming studies suggesting a link between an induced sad mood

and hopelessness is consistent with the pattern of data from the ESM study. The confirmation

of such link between the natural fluctuation of mood in everyday life and hopeless thoughts

172
conveys implications that are crucial to understanding the suicidal thinking process in FEP.

More specifically, it supports the assumptions of the DAH of suicidal relapse, which suggests

that the mechanism of hopeless/suicidal thinking is mood-dependent. From a clinical point of

view, the results of this study present potentially valuable inputs that will help manage and

prevent suicidal relapse in FEP more effectively. First, the application of the DAH for suicidal

relapse as a framework of suicidality in psychosis could provide clinicians a better

understanding of the suicidal thinking process, and a better insight for a more effective risk

assessment. Unlike the traditional suicidal risk assessment which mainly relies on historical

information (distal) and/or the immediate (proximal) risk factors, recognition of the

interaction between the distal and proximal risk factors as suggested by the DAH framework

could help establish a more effective way to assess suicidal vulnerability. Second, the

appraised “reactivity” to momentary hopelessness linked to changes in negative affectivity in

everyday life suggests that low-level hopelessness remains even though the individuals with a

history of suicidal behaviour were not “currently” suicidal. The absence of suicidal attempts

despite the activation of suicidal ideation supports the idea that attenuated hopelessness

persists on a day to day basis for those with histories of suicidal behaviour and this may be

speculated that if this was a target for intervention, this might act to interrupt the mood-

hopeless-suicidal attempt cycle. In view of this, the use of the ESM as a tool for assessing the

individual’s vulnerability to hopeless thoughts in everyday life could potentially offer a more

effective form of risk assessment. As the ESM was devised to sample data from the

individual’s natural environment, the data from the ESM could present a better understanding

of how hopelessness reacts to the natural fluctuations of mood in real life. For this reason, the

ESM could also function as an alternative measure of cognitive reactivity to hopelessness in

everyday life. Third, the ability of the ESM data to provide real life contexts (e.g. people,

173
places, activities, & events) to the interaction between mood and hopelessness could also

provide a better insight on the role of contextual factors in suicidal thinking. Further,

identification of good and problematic contexts could be a useful input in the development of

new interventions for suicidality. More importantly, the ESM could also be a valuable tool for

both the clinician and the individual by: (1) providing clinicians a way to assess the efficacy

of the interventions for suicidality, and (2) educating the individual to be more mindful of

his/her mood along with the context that he/she is in. Fourth, the impairment of problem

solving following a downward shift in mood suggests that the development of problem

solving abilities could be an important focus of interventions for suicidality. The development

of problem solving ability could facilitate a better coping mechanism and enhance the self-

esteem/confidence of the individual. A study on resilience to suicidality has indicated that

positive attributional style was one of the psychological factors that act as a “buffer” to

suicidality (Johnson, Wood, Gooding, & Tarrier, 2011). Finally, the decrease in fluency for

positive events following the sad mood induction suggests that the development of goal

specificity could be another important focus on interventions for suicidality. A study on the

effect of Mindfulness-Based Cognitive Therapy on the specificity of goals in a sample of

previously suicidal individuals with chronic depression has indicated that being mindful

facilitated identification of more specific goals (Crane, Winder, Hargus, Amarasinghe, &

Barnhofer, 2012). There is every reason to suppose that this may be successful in the early

phase of psychosis where suicide is at its highest and is a very positive avenue for further

research.

174
6.5. Conclusion

In conclusion, the results of the studies in this thesis illustrated that low mood triggers

the “differential activation” of hopeless/suicidal thoughts as proposed by the DAH of suicidal

relapse. Importantly, mood-linked impairments in problem solving and positive future

fluency, along with reactivity to momentary hopelessness in everyday life were all

demonstrated to be significantly more evident in the suicidal history group than the non-

suicidal group. Thus, it is indicated that mood-dependent problem solving deficit and

dysfluency for positive events, along with increased reactivity to momentary hopelessness

linked to negative affectivity are significant characteristics of a greater suicidal vulnerability

in FEP individuals with a history of suicidal behaviour. Therefore, the application of the DAH

as a framework for understanding the suicidal thinking in FEP warrants further studies, in

order to improve existing interventions for suicidality and reduce the likelihood of subsequent

suicidal relapse. Specifically, the use of the ESM as a potential tool for assessing suicidal

vulnerability also requires further research in order to improve existing risk assessment

procedures. Further, the mindfulness-based interventions used to prevent depression relapse in

MDD may well have utility in preventing escalation from momentary changes in hopelessness

linked to daily life experiences, in this most difficult and clinically challenging area of

psychosis.

175
 References

Abramson, L. Y., Metalsky, G. I., & Alloy, L. B. (1989). Hopelessness depression: A theory-

based subtype of depression. Psychological Review, 96, 358–372.

Abramson, L.Y., Alloy, A. B., Hogan, M. E., Whitehouse, W. G., Cornette, M., Akhavan, S.,

et al. (1998). Suicidality and cognitive vulnerability to depression among college

students: A prospective study. Journal of Adolescence, 21, 473–487.

Addington, D., Addington, J., and Maticka-Tyndale, D. (1993) Assessing depression in

schizophrenia: The Calgary Depression Scale. British Journal of Psychiatry, 163

(22), 39-44.

Addington, J., Williams, J., Young, J., Addington, D. (2004). Suicidal behaviour in early

psychosis. Acta Psychiatrica Scandinavica, 109 (2), 116–120.

American Psychiatric Association. (2000). Diagnostic and statistical manual of mental

disorders (4th ed.). Text Revision. Washington, DC.

Andersen, S.M., Spielman, L.A., & Bargh, J.A. (1992). Future-event schemas and certainty

about the future: Automaticity in depressives’ future-event predictions. Journal of

Personality and Social Psychology, 63, 711–723.

Arie, M., Apter, A., Orbach, I., Yefet, Y., & Zalzman, G. (2008). Autobiographical memory,

interpersonal problem solving, and suicidal behavior in adolescent inpatients.

Comprehensive Psychiatry, 49 (1), 22–29.

Bakst, S., Rabinowitz, J., Bromet, E.J. (2010). Is poor premorbid functioning a risk factor for

suicide attempts in first-admission psychosis? Schizophrenia Research, 116 (2–3),

210–216.

176
Barrett, E.A., Sundet, K., Faerden, A., Nesvag, R., Agartz, I., Fosse, R., Mork, E., Steen,

N.E., Andreassen, O.A., Melle, I. (2010). Suicidality before and in the early phases

of first episode psychosis. Schizophrenia Research. 119 (1–3), 11–17.

Barton K., & Jackson C. (2008). Reducing the impact of trauma amongst carers. Australia

and New Zealand Journal of Psychiatry, 42, 693-701.

Baumeister, R. E (1990). Anxiety and deconstruction: On escaping the self. In J. M. Olson &

M. P. Zanna. Self-inference processes: The Ontario Symposium (Vol. 6, pp. 259-

291). Hilisdale, NJ: Erlbaum.

Beck, A.T., Weissman, A., Lester, D., & Trexler, L. ( 1974). The measurement of pessimism:

The Hopelessness Scale. Journal of Consulting and Clinical Psychology, 42 (86), 1-

865.

Beck, A.T., Kovac, S.M., & Weissman, A. (1975). Hopelessness and suicidal behaviour: An

overview. Journal of the American Medical Association, 234, 1146-1 149.

Beck, A.T., Steer, R.A., Kovacs, M., et al. (1985). Hopelessness and eventual suicide: a 10-

year prospective study of patients hospitalized with suicidal ideation. American

Journal Psychiatry, 142, 559—563

Beck, A.T. & Steer, R.A. (1988). Manual for the Beck Hopelessness Scale. San Antonio,

Texas Psychological Corp.

Beck, A. T., Brown, G., Berchick, R. J., Stewart, B. L, & Steer, R. A. (1990). Relationship

between hopelessness and ultimate suicide: A replication with psychiatric patients.

American Journal of Psychiatry, 147, 190–195.

177
Beck, A. T., Steer, R. A., & Brown, G. (1993). Dysfunctional attitudes and suicidal ideation

in psychiatric outpatients. Suicide and Life Threatening Behavior, 23, 11–20.

Psychology, 109, 504–511.

Beck, A. T., Brown, G. K., Steer, R. A., Dahlsgaard, K. K., & Grisham, J. R. (1999).

Suicide ideation at its worst point: A predictor of eventual suicide in psychiatric

outpatients. Suicide and Life-Threatening Behavior, 29, 1-9.

Ben-Zeev, D., Morris, S., Swendsen, J., & Granholm, E. (2011). Predicting the occurrence,

conviction, distress, and disruption of different delusional experiences in the daily

life of people with schizophrenia. Schizophrenia Bulletin, E-pub ahead of print.

Bertelsen M, Jeppesen P, Petersen L. (2007). Suicidal behaviour and mortality in first-episode

psychosis: the OPUS trial. British Journal Psychiatry, 191 (Suppl. 51), s140–s146.

Biggam, F. H., & Power, K. G. (1999). Suicidality and the state-trait debate on problem-

solving deficits: A re-examination with incarcerated young offenders. Archives of

Suicide Research, 5, 27–42.

Birchwood, M., Smith, J., MacMillan, F. et al. (1989). Predicting relapse in schizophrenia:

the development and implementation of an early signs monitoring system using

patients and families as observers. Psychological Medicine, 19, 649–656.

Birchwood, M. J., Mason, R., MacMillan, F., & Healy J. (1993). Depression, demoralisation

and control over psychotic illness: A comparison of depressed and non-depressed

patients with a chronic psychosis. Psychological Medicine, 23, 387–395.

Birchwood, M., Fowler, D., & Jackson, C. (2001). Early intervention in psychosis.

Chichester: John Wiley and Son.

178
Birchwood, M., and MacMillan, J.F. (1993a). Early intervention in schizophrenia. Australian

and New Zealand Journal of Psychiatry, 27, 374-378.

Birchwood, M., Mason, R., MacMillan, F., & Healy, J. (1993b). Depression, demoralisation

and control over illness: A comparison of depressed and non-depressed patients with

a chronic psychosis. Psychological Medicine, 23, 387–395.

Birchwood, M., McGorry, P., and Jackson, H. (1997). Early Intervention in Schizophrenia.

British Journal of Psychiatry, 170, 2 – 5

Birchwood, M., Iqbal, Z., Trower, P., et al. (2000). Cognitive approach to depression and

suicidal thinking in psychosis, I. Ontogony of post psychotic depression. British

Journal of Psychiatry, 177, 516–521.

Birchwood, M., Fowler, D., & Jackson, C. (2001). Early intervention in psychosis.

Chichester: John Wiley and Son.

Bourgeois M, Swendsen K, Young F, Amador X, Pini S, Cassano GB, et al. (2004).

Awareness of disorder and suicide risk in the treatment of schizophrenia: Results of

the International Suicide Prevention Trial. American Journal of Psychiatry, 161,

1494–1496.

Bower, G. H. (1981). Mood and memory. American Psychologist, 36, 129-148.

Bresnahan M, Begg M, Brown AS, Schaefer C, Sohler N, Insel B, Vella L, Susser E (2007).

Race and Schizophrenia in a US birth cohort: Another example of health disparity?

International Journal of Epidemiology, 36 (4), 751-8.

Brown, S., 1997. Excess mortality of schizophrenia. A meta-analysis. British Journal of

Psychiatry 171, 502–508.

179
Brunet, K., Birchwood, M., Upthegrove, R., Michail, M., & Ross, K. (2012). A prospective

study of PTSD following recovery from first-episode psychosis: The threat from

persecutors, voices, and patienthood. British Journal of Clinical Psychology, 51,

418–433.

Carver, C. S., & Scheier, M. F. (1999). Themes and issues in the self-regulation of behavior.

In R. S. Wyer, Jr. (Ed.), Advances in social cognition, 12, (pp. 1-105). Mahwah,

NJ: Lawrence Erlbaum.

Ceskova, E., Prikryl, R., & Kasparek, T. (2011). Suicides in males after the first episode of

schizophrenia. Journal of Nervous Mental Disease, 199 (1), 62–64.

Christensen, T. C., Barrett, L. F., Bliss-Moreau, E., Lebo, K. & Kaschub, C. (2003). A

practical guide to experience-sampling procedures. Journal of Happiness Studies, 4,

53-78.

Clarke, M., Whitty, P., Browne, S., Mc Tigue, O., Kinsella, A., Waddington, J.L., Larkin, C.,

& O'Callaghan, E. (2006). Suicidality in first episode psychosis. Schizophrenia

Research, 86 (1–3), 221–225.

Cohen, S., Lavelle, J., Rich, C.L., & Bromet, E. (1994). Rates and correlates of suicide

attempts in first-admission psychotic patients. Acta Psychiatrica Scandinavica, 90

(3), 167–171.

Collins, A.A., Remington, G., Coulter, K., & Birkett, K. (1996). Depression in schizophrenia:

a comparison of three measures. Schizophrenia Research, 20, 205–209

Conaghan, S., & Davidson, K.M. (2002). Hopelessness and the anticipation of positive and

negative future experiences in older parasuicidal adults. British Journal of Clinical

Psychology, 41 (3). 233-242.

180
Conner, K.R., Duberstein, P.R., Conwell, Y., Seidlitz, L., & Caine, E.D. (2001).

Psychological vulnerability to completed suicide: A review of empirical studies.

Suicide Life Threatening Behaviour, 31, 367–385.

Cotton, S.M., Lambert, M., Schimmelmann, B.G., Foley, D.L., Morley, K.I., McGorry, P.D.,

& Conus, P. (2009). Gender differences in premorbid, entry, treatment, and outcome

characteristics in a treated epidemiological sample of 661 patients with first episode

psychosis. Schizophrenia Research, 114 (1–3), 17–24.

Craig, T.K., Garety P., & Power P. (2004). The Lambeth Early Onset (LEO) Team:

randomised controlled trial of the effectiveness of specialised care for early

psychosis. British Medical Journal, 329, 1067–1067.

Crumlish, N., Whitty, P., Kamali, M., Clarke, M., Browne, S., McTigue, O., Lane, A.,

Kinsella, A., Larkin, C., & O'Callaghan, E., (2005). Early insight predicts depression

and attempted suicide after 4 years in first-episode schizophrenia and

schizophreniform disorder. Acta Psychiatrica Scandinavica, 112 (6), 449–455.

Csikszentmihalyi, M., & Schneider, B. (2001). Conditions for optimal development in

adolescence: An experiential approach. Applied Developmental Science, 5, 122-124.

Curry, J. F., Miller, Y., Waugh, S., & Anderson, W. B. (1992). Coping responses in

depressed, socially maladjusted, and suicidal adolescents. Psychological Reports, 71,

80–82.

De Hert, M., McKenzie, K., & Peuskens, J. (2001). Risk factors for suicide in young people

suffering from schizophrenia: a long-term follow-up study. Schizophrenia Research,

47 (2–3), 127–134.

181
Delespaul, P., & deVries, M. W. (1987). The daily life of ambulatory chronic mental patients.

Journal of Nervous and Mental Disease, 175, 537–544.

Delespaul, P. (1995). Assessing Schizophrenia in Daily Life—The Experience Sampling

method. Maastricht, the Netherlands: Maastricht University Press

Delespaul, P., deVries, M., & van Os, J. (2002). Determinants of occurrence and recovery

from hallucinations in daily life. Social Psychiatry and Psychiatric Epidemiology,

37, 97–104.

DeVries, M. (1992) The Experience of Psychopathology. Investigating Mental Disorders in

Their Natural Settings. Cambridge: England, Cambridge University Press.

Dictionary.com Unabridged. Retrieved February 07, 2013, from Dictionary.com website:

http://dictionary.reference.com/browse/-less

Dyer J.A, & Kreitman, N. (1984). Hopelessness, depression, and suicide intent in parasuicide.

British Journal of Psychiatry, 144, 127—133.

Fialko, L., Freeman, D., Bebbington, P.E., Kuipers, E., Garety, P.A., Dunn, G., et al. (2006).

Understanding suicidal ideation in psychosis: Findings from the Psychological

Prevention of Relapse in Psychosis (PRP) Trial. Acta Psychiatrica Scandanavica,

114, 117–186.

Flanagan, P. & Compton, MT. (2012). A comparison of correlates of suicidal ideation prior to

initial hospitalization for first-episode psychosis with prior research on correlates of

suicide attempts prior to initial treatment seeking. Early Intervention Psychiatry, 6

(2), 138-44.

182
Foley, S., Jackson, D., McWilliams, S., Renwick, L., Sutton, M., Turner, N., Kinsella, A., &

O'Callaghan, E. (2008). Suicidality prior to presentation in first-episode psychosis.

Early Intervention Psychiatry, 2, 242–246.

Gaebel, W. and Woelwer, W. (1992) Facial expression and emotional face recognition in

schizophrenia and depression. European Archives of Psychiatry and Clinical

Neurosciences, 242, 46-52.

Gard, D. E., Kring, A. M., Gard, M., Horan, W. P., & Green, M. F. (2007). Anhedonia in

schizophrenia: Distinctions between anticipatory and consummatory pleasure.

Schizophrenia Research, 93, 253-360.

Goddard, L., Dritschel, B. & Burton, A. (1996). Role of autobiographical memory in social

problem solving and depression. Journal of Abnormal Psychology, 105, 609-616.

Granholm, E., Loh, C., & Swendsen, J. (2008). Feasibility and validity of computerized

ecological momentary assessment in schizophrenia. Schizophrenia Bulletin, 34, 507–

514.

Gumley, A., O’Grady, M., Power, K. & Schwannauer, M. (2004). Negative beliefs about self

and illness: A comparison of socially anxious and non-socially anxious individuals

with psychosis. Australian and New Zealand Journal of Psychiatry, 38, 960-964.

Gumley, A. I., Karatzias, A., Power, K. G., Reilly, J., McNay, L., & O'Grady, M. (2006).

Early inter-vention for relapse in schizophrenia: Impact of cognitive behavioural

therapy on negative beliefs about psychosis and self-esteem. British Journal of

Clinical Psychology, 45, 247–260.

183
Haahr, U., Friis, S., Larsen, T.K., Melle, I., Johannessen, J.O., Opjordsmoen, S., Simonsen,

E., Rund, B.R., Vaglum, P., & McGlashan, T. (2008). First-episode psychosis:

diagnostic stability over one and two years. Psychopathology 41, (5), 322–329.

Hafner, H., Maurer, K., Loffler, W., & Riecher-Rossler, A. (1993). The influence of age and

sex on the onset and early course of schizophrenia. British Journal of Psychiatry,

162, 80-86.

Hafner, H., Riecher-Rossler, A., Maurer, K., Fatkenheuer, B., & Loffler, W. (1992). First

onset and early symptomatology of schizophrenia: A chapter of epidemiological and

neurobiological research into age and sex differences. European Archives of

Psychiatry and Clinical Neurosciences, 242, 109-118.

Hafner, H., Maurer, K., Loffler, W., an der Heiden, W., Munk-J0rgensen, P., Hambrecht, M.,

& Riecher-Rb'ssler, A. (1998). The ABC schizophrenia study: A preliminary

overview of the results. Social Psychiatry and Psychiatric Epidemiology, 33, 380-

386.

Harkavy-Friedman, J.M. (2006). Can early detection of psychosis prevent suicidal behavior?

American Journal of Psychiatry, 163, 768-70.

Harkavy-Friedman, J.M., Restifo, K., Malaspina, D., Kaufmann, C.A., Amador, X.F., Yale,

S.A., & Gorman, J.M. (1999). Suicidal behavior in schizophrenia: characteristics of

individuals who had and had not attempted suicide. American Journal of Psychiatry,

156 (8), 1276–1278.

Harris, E.C., & Barraclough, B. (1997). Suicide as an outcome for mental disorders. A meta-

analysis. British Journal of Psychiatry 170, 205–228.

184
Harrison, C. L., & Fowler, D. (2004). Negative symptoms, trauma, and autobiographical

memory: An investigation of individuals recovering from psychosis. The Journal of

Nervous and Mental Disease, 192, 745-753.

Harvey, S.B., Dean, K., Morgan, C., Walsh, E., Demjaha, A., Dazzan, P., Morgan, K., Lloyd,

T., Fearon, P., Jones, P.B., & Murray, R.M. (2008). Self-harm in first-episode

psychosis. British Journal of Psychiatry, 192 (3), 178–184.

Hawton, K. (1997). Attempted suicide. In D.M. Clarke & C.G. Fairburn (Eds.). Science

and Practice of Cognitive Behaviour Therapy (pp. 285 – 312). Oxford, Oxford

University Press.

Hawton, K., Sutton, L., Haw, C., Sinclair, J., & Deeks, J.J. (2005). Schizophrenia and suicide:

systematic review of risk factors. British Journal of Psychiatry, 187, 9–20.

Hawton, K., Sutton, L., Haw, C., Sinclair, J., & Deeks, J.J. (2005). Schizophrenia and suicide:

Systematic review of risk factors. British Journal of Psychiatry, 187, 9–20.

Hawton, K. & van Heeringen, K. (2009). Suicide. Lancet, 373, 1372–1381.

Heinrichs, D.W. & Carpenter, W.T. Jr. (1985). Prospective study of prodromal symptoms in

schizophrenic relapse. American Journal of Psychiatry, 142, 371-373.

Heila, H., Isometsa, E.T.,Henriksson,M.M. et al (1997) Suicide and schizophrenia: a

nationwide psychological autopsy study on age- and sex-specific clinical

characteristics of 92 suicide victims with schizophrenia. American Journal of

Psychiatry, 154, 1235-1242.

Hektner, J. M., Schmidt, J. A., & Csikszentmihalyi, M. (2007). Experience sampling method:

Measuring the quality of everyday life. Thousand Oaks, California, Sage

Publications.

185
Henquet, C., Rosa, A., Delespaul, P., Papiol, S., Fananas, L., van Os, J., & Myin-Germeys, I.

(2009). COMT ValMet moderation of cannabis-induced psychosis: a momentary

assessment study of 'switching on' hallucinations in the flow of daily life. Acta

Psychiatrica Scandanavica, 119, 156-60.

Henquet, C., van Os, J., Kuepper, R., Delespaul, P., Smits M., Campo, J.A., & Myin-

Germeys, I. (2010). Psychosis reactivity to cannabis use in daily life: an experience

sampling study. British Journal of Psychiatry, 196, 447-53.

Hepburn, S.R., Barnhofer, T., & Williams, J.M.G. (2006). Effects of mood on how future

events are generated and perceived. Personality and Individual Differences, 41, 801–

811.

Hor, K. & Taylor, M. (2010). Suicide and schizophrenia: a systematic review of rates and risk

factors. Journal Psychopharmacol, 24 (4 Suppl), 81–90.

Hunter, E.C., O'Connor, R.C., 2003. Hopelessness and future thinking in parasuicide: the role

of perfectionism. British Journal of Clinical Psychology, 42, 355–365.

Hurlburt, R. T. & Melancon, S. M. (1987). “Goofed-up” images: Thought sampling with a

schizophrenic woman. Journal of Nervous and Mental Disease, 175, 575–578.

Iqbal, Z., Birchwood, M., Chadwick, P., & Trower P. (2000). Cognitive approach to

depression and suicidal thinking in psychosis. 2. Testing the validity of a social

ranking model. British Journal of Psychiatry, 177, 522–528.

Isometsa, E., Heikkinen, M., Henriksson, M., Aro, H. & Lonqvist, J. (1995). Recent life

events and completed suicide in bipolar affective disorder: A comparison with Major

Depressive Disorder in Finland. Journal of Affective Disorder, 33, 99-106.

186
Ivanoff, A., Smyth, N. J., Grochowski, S., Jang, S. J., & Klein, K. E. (1992). Problem solving

and suicidality among prison inmates: Another look at state versus trait. Journal of

Consulting and Clinical Psychology, 60, 970-973.

Jackson, H.J., McGorry, P.D., & McKenzie, D. (1994). The reliability of DSM-III prodromal

symptoms in first episode psychotic patients. Acta Psychiatrica Scandinavica, 90,

375-378.

Jackson, C. & Iqbal, Z. (2000). Psychological adjustment to early psychosis. In M.

Birchwood, D. Fowler, & C. Jackson (Eds.), Early intervention in psychosis: A

guide to concepts, evidence and interventions (pp.64-99). England: John Wiley &

Sons.

Jackson, C., Knott, C., Skeate, A., et al. (2004). The trauma of first episode psychosis: The

role of cognitive mediation. Australian and New Zealand Journal of Psychiatry, 38,

327–333.

Jackson, C., Trower, P., Reid, I., Smith, J., Hall, M., Townend, M., et al. (2009). Improving

psychological adjustment following a first episode of psychosis: A randomised

controlled trial of cognitive therapy to reduce post psychotic trauma symptoms.

Behaviour Research and Therapy, 47, 454-462.

Johnson, J., Tarrier, N., & Gooding, P.A. (2008). An investigation of aspects of the cry of

pain model of suicide risk: the role of defeat in impairing memory. Behaviour

Research and Therapy, 46 (8), 968-975.

Kikuchi, H., Yoshiuchi, K., Miyasaka, N., Ohashi, K., Yamamoto, Y., Kumano, H., Kuboki,

T. & Akabayashi, A. (2006). Reliability of recalled self-report on headache intensity:

187
 Investigation using ecological momentary assessment technique. Cephalalgi, 26,

1335-1343.

Kim, C.H., Jayathilake, K., & Meltzer, H.Y. (2002). Hopelessness, neurocognitive function,

and insight in schizophrenia: Relation to suicidal behavior. Schizophrenia Research,

60, 71–80.

Kimhy, D., Delespaul, P., Corcoran, C., Ahn, H., Yale, S., & Malaspina, D. (2006).

Computerized experience sampling method (ESMc): Assessing feasibility and

validity among individuals with schizophrenia. Journal of Psychiatric Research, 40,

221–230.

Kimhy, D., Delespaul, P., Ahn, H., Cai, S., Shikhman, M., Lieberman, J.A., Malaspina, D., &

Sloan, R.P. (2010). Concurrent measurement of "real-world" stress and arousal in

individuals with psychosis: assessing the feasibility and validity of a novel

methodology. Schizophrenia Bulletin, 36, 1131-1139.

King, E. A., Baldwin, D. S., Sinclair, J. M., et al. (2001). TheWessex Recent In-Patient

Suicide Study, 1. Case-control study of 234 recently discharged psychiatric patient

suicides. British Journal of Psychiatry, 178, 531-536.

Kirch, D.G., Lieberman, J.A., & Matthews, S.M. (1992). First-episode psychosis: Part I.

Editors' introduction. Schizophrenia Bulletin, 18 (2), 177-178.

Kiviniemi, M. T. & Rothman, A. J. (2006). Selective memory biases in individuals’ memory

for health-related information and behavioural recommendations. Psychology &

Health, 21 (2), 247-272.

188
Klonsky, E. D., Kotov, R., Bakst, S., Rabinowitz, J., & Bromet, E. J. (2012). Hopelessness as

a predictor of attempted suicide among first admission patients with psychosis: A 10-

year cohort study. Suicide & Life Threatening Behavior, 42 (1), 1-10.

Kreitman, N. (1977). Parasuicide. London: Wiley.

Kring, A. M., Kerr, S. L., Smith, D. A. & Neale, J. M. (1993). Flat affect in schizophrenia

does not reflect diminished subjective experience of emotion. Journal of Abnormal

Psychology, 102 (4), 507-517.

Kring, A. M. & Neale, J. M. (1996). Do schizophrenic patients show a disjunctive

relationship among expressive, experiential, and psychophysiological components of

emotion? Journal of Abnormal Psychology, 105, 249-257.

Lam, D., Schuck, N., Smith, N., Farmer, A. & Checkley, S. (2003). Response style,

interpersonal difficulties and social functioning in major depressive disorder. Journal

of Affective Disorders, 75, 279-283.

Lamph, G. (2010) Early psychosis: raising awareness among non-mental health nurses.

Nursing Standard. 24, 47, 35-40.

Lancon, C., Auquier, P., Reine, G., Toumi, M., & Addington, D. (1999). Evaluation of

depression in schizophrenia: psychometric properties of a French version of the

calgary depression scale. Psychiatry Research., 89, 123–132

Lardinois, M., Myin-Germeys, I., Bak, M., Mengelers, R., van Os J., & Delespaul, P.A.

(2003). The dynamics of symptomatic and non-symptomatic coping with psychotic

symptoms in the flow of daily life. Acta Psychiatrica Scandanavica, 116, 71-5.

189
Lardinois, M., Lataster, T., Mengelers, R., Van Os J., & Myin-Germeys, I. (2011). Childhood

trauma and increased stress sensitivity in psychosis. Acta Psychiatrica Scandanavica,

123, 28-35.

Lardinois, M., Lataster, T., Mengelers, R., Van Os J., & Myin-Germeys, I. (2011). Childhood

trauma and increased stress sensitivity in psychosis. Acta Psychiatrica Scandanavica,

123, 28-35.

Lataster, T., Collip, D., Lardinois, M., van Os J., & Myin-Germeys, I. (2010). Evidence for a

familial correlation between increased reactivity to stress and positive psychotic

symptoms. Acta Psychiatrica Scandanavica, 122, 395-404.

Lau, M. A., Segal, Z. V., & Williams, J. M. G. (2004). Teasdale’s differential activation

hypothesis: implications for mechanisms of depressive relapse and suicidal

behaviour. Behaviour Research and Therapy, 42, 1001–1017.

Leverich, G.S., Altshuler Frye, M.A., Suppes, T., Keck, P.E., McElroy, S.L., Denicoff, K.D.,

Obrocea, G., Nolen, W.A., Kupka, R., Walden, J., Grunze, H., Perez, S.,

Luckenbaugh, D.A., & Post, R.M. (2003). Factors associated in suicide attempts in

684 patients with bipolar disorder in the Stanley Foundation Bipolar Network.

Journal of Clinical Psychiatry, 64, 506 - 515.

Levine, S.Z., Bakst, S., & Rabinowitz, J. (2010). Suicide attempts at the time of first

admission and during early course schizophrenia: a population based study.

Psychiatry Research, 177 (1–2), 55–59.

Lezak, M. D. (2004) Neuropsychological assessment (4th ed.). New York: Oxford University

Press.

190
Lindenmayer, J.P., Czobor, P., Alphs, R., Anand, R., Islam, Z., & Pestreich, L. (2001). The

InterSept Scale for Suicidal Thinking (ISST): A new assessment instrument for

suicidal patients with schizophrenia. Schizophrenia Research, 49 (suppl. 1-2), 5.

Lindenmayer JP., Czobor P., Alphs L., Nathan AM., Anand R., Islam Z., & Chou JC. (2003).

The InterSePT scale for suicidal thinking reliability and validity. Schizophrenua

Research, 63, 161-170

Linehan, M. M., Goodstein, J. L., Nielsen, S. L., & Chiles, J. A. (1983). Reasons for

staying alive when you are thinking of killing yourself: The Reasons for Living

Inventory. Journal of Consulting and Clinical Psychology, 51 (2), 276-286.

Limosin F, Loze JY, Philippe A, Casadebaig F, et al. (2007). Ten-year prospective follow-up

study of the mortality by suicide in schizophrenic patients. Schizophrenia Research

94, (1-3), 23-28.

MacLeod, A. K., Rose, G., & Williams, J. M. G. (1993). Components of hopelessness about

the future in parasuicide. Cognitive Therapy and Research, 17, 441–455.

MacLeod, A. K., & Cropley, M. L. (1995). Depressive future-thinking: The role of valence

and specificity. Cognitive Therapy and Research, 19, 35–50.

MacLeod, A. K., Byrne, A., & Valentine, J. D. (1996). Affect, emotional disorder, and future-

directed thinking. Cognition and Emotion, 10, 69–85.

MacLeod, A.K., Tata, P., Kentish, J., & Jacobsen, H. (1997). Retrospective and prospective

cognitions in anxiety and depression. Cognition and Emotion, 11, 467-479.

191
MacLeod, A. K., Pankhania, B., Lee, M. & Mitchell, N. (1997). Parasuicide, depression and

the anticipation of positive and negative future experiences. Psychological Medicine,

27, 973-977.

MacLeod, A. K., Tata, P., Evans, K., Tyrer, P., Schmidt, U., Davidson, K., Thornton, S., &

Catalan, J. (1998). Recovery of positive future thinking within a high-risk

parasuicide group: Results from a pilot randomized controlled trial. British Journal

of Clinical Psychology, 37, 371–379.

MacLeod, A. K., & Salaminiou, E. (2001). Reduced positive future-thinking in depression:

Cognitive and affective factors. Cognition and Emotion, 15, 99–107.

MacLeod, A.K., Tata, P., Tyrer, P., Schmidt, U., Davidson, K., & Thompson, S., (2005).

Hopelessness and positive and negative future thinking in parasuicide. The British

Journal of Clinical Psychology 44, 495–504.

MacLeod, A. K., Pankhania, B., Lee, M., & Mitchell, D. (1997). Parasuicide, depression and

the anticipation of positive and negative experiences. Psychological Medicine, 27,

973–977.

Malla, A.K. & Norman, R.M. Prodromal symptoms in schizophrenia. (1994). British Journal

of Psychiatry, 164, 487-493.

Malone, K. M., Oquendo, M. A., Haas, G. L., Ellis, S. P., Li, S. H., & Mann, J. J. (2000).

Protective factors against suicidal acts in major depression: Reasons for living.

American Journal of Psychiatry, 157, 1084–1088.

Mangalore, R. & Knapp, M. (2007). Cost of schizophrenia in England. Journal of Mental

Health Policy and Economics, 10, 23-41.

192
Mann, J. J., Waternaux, C., Haas, G. L., & Malone, K. M. (1999). Towards a clinical model

of suicidal behaviour in psychiatric patients. American Journal of Psychiatry, 156,

181–189.

Marshall, M., Lewis, S., Lockwood, A., Drake, R., Jones, P., & Croudace, T. (2005).

Association between duration of untreated psychosis and outcome in cohorts of first-

episode patients: a systematic review. Archives of General Psychiatry, 62 (9), 975-

983.

McCormack, H.M., Horne, D.J., & Sheather, S. (1988). Clinical applications of visual

analogue scales: a critical review. Psychological Medicine, 18, 1007-1019.

McGorry, P. D., Singh, B. S., Copolov, D. L. et al. (1990). Royal Park Multidiagnostic

Instrument for Psychosis: Part II. Development, Reliability, and Validity.

Schizophrenia Bulletin, 16, 517-536.

McGorry, P., & Jackson. H. (1999). Recognition and management of early psychosis. A

Preventative Approach. Cambridge. Cambridge University Press.

McGorry, P.D., Chanen, A., McCarthy, E., Van Riel, R. et al. (1991). Posttraumatic stress

disorder following recent-onset psychosis: An unrecognized postpsychotic

syndrome. Journal of Nervous and Mental Disease, 179 (5), 253-258.

Melle, I., Johannesen, J.O., Friis, S., Haahr, U., Joa, I., Larsen, T.K., Opjordsmoen, S., Rund,

B.R., Simonsen, E., Vaglum, P., & McGlashan, T. (2006). Early detection of the first

episode of schizophrenia and suicidal behavior. American Journal of Psychiatry, 163

(5), 800–804.

193
Miles, H., MacLeod, A. K., & Pote, H. (2004). Retrospective and prospective cognitions in

adolescents: Anxiety, depression, and positive and negative affect. Journal of

Adolescence, 27, 691–701.

Minkoff, K., Bergman, E., Beck, A.T., et al. (1973). Hopelessness, depression, and attempted

suicide. American Journal of Psychiatry, 130, 455—459

Montross, L., Kasckow, J., Golshan, S., Solorzano, E., Lehman, D., & Zisook, S. (2008).

Suicidal ideation and suicide attempts among middle-aged and older patients with

schizophrenia spectrum disorders and concurrent sybsyndromal depression. The

Journal of Nervous and Mental Disease, 196 (12), 884-890.

Morrens, M., Krabbendam, L., Bak, M., Delespaul, P., Mengelers, R., Sabbe, B., Hulstijn, W.,

van Os J., & Myin-Germeys I. (2007). The relationship between cognitive

dysfunction and stress sensitivity in schizophrenia: a replication study. Social

Psychiatry & Psychiatric Epidemiology, 42, 284-7.

Myin-Germeys, I., Delespaul, P.A., & deVries, M.W. (2000). Schizophrenia patients are more

emotionally active than is assumed based on their behavior. Schizophrenia Bulletin,

26, 847-54.

Myin-Germeys, I., Nicolson, N.A., & Delespaul, P.A. (2001). The context of delusional

experiences in the daily life of patients with schizophrenia. Psychological Medicine,

31, 489-98.

Myin-Germeys, I., Krabbendam, L., Jolles, J., Delespaul, P.A., & van Os, J. (2002). Are

cognitive impairments associated with sensitivity to stress in schizophrenia? An

experience sampling study. American Journal of Psychiatry, 159, 443-9.

194
Myin-Germeys, I., Krabbendam, L., Delespaul, P., & van Os, J. (2003). Can cognitive deficits

explain differential sensitivity to life events in psychosis? Social Psychiatry &

Psychiatric Epidemiology, 38, 262-8.

Myin-Germeys, I., Krabbendam, L., Delespaul, P.A., & van Os, J (2004). Sex differences in

emotional reactivity to daily life stress in psychosis. Journal of Clinical Psychiatry,

65, 805-9.

Myin-Germeys, I., Delespaul, P., & van Os, J. (2005). Behavioural sensitization to daily life

stress in psychosis. Psychological Medicine, 35, 733-41.

Myin-Germeys, I., Oorschot, M., Collip, D., Lataster, J., Delespaul, P., & van Os, J (2009).

Experience sampling research in psychopathology: opening the black box of daily

life. Psychological Medicine, 39, 1533-47.

National Institute for Health and Clinical Excellence. (2009). Schizophrenia: Core

Interventions in the Treatment and Management of Schizophrenia. London: National

Institute for Health and Clinical Excellence.

Nekanda-Trepka, C.J.S., Bishop, S., & Blackburn, M. (1983). Hopelessness and depression.

British Journal of Clinical Psychiatry, 132, 954—956

Nordentoft, M., Jeppesen, P., Abel, M., Kassow, P., Petersen, L., Thorup, A., Krarup, G.,

Hemmingsen, R., & Jorgensen, P. (2002). OPUS study: suicidal behaviour, suicidal

ideation and hopelessness among patients with first-episode psychosis. One-year

follow-up of a randomised controlled trial. British Journal of Psychiatry Supplement,

43, 98–106.

195
Nordentoft, M., Laursen, T.M., Agerbo, E., Qin, P., Høyer, E.H. & Mortensen, P.B. (2004).

Change in suicide rates for patients with schizophrenia in Denmark, 1981–97: Nested

case-control study. British Medical Journal, 329, 261.

O'Connor, R.C., O'Connor, D.B., O'Connor, S.M., Smallwood, J. & Miles, J. (2004).

Hopelessness, stress and perfectionism: the moderating effects of future thinking.

Cognition and Emotion, 18, 1099-1120.

O’Connor, R.C. (2011). Towards a Motivational–Volitional Model of Suicidal Behaviour. In

R.C. O’Connor, S. Platt, & J. Gordon (Eds.), The International Handbook of

Suicide Prevention: Research, Policy and Practice (p.181-198). Chichester: John

Wiley & Sons.

O'Connor, R.C. & Cassidy, C. (2007). Predicting hopelessness: the interaction between

optimism/pessimism and specific future expectancies. Cognition and Emotion, 21,

596–613.

O'Connor, R.C., Connery, H. & Cheyne,W. (2000). Hopelessness: The role of depression,

future directed thinking and cognitive vulnerability. Psychology, Health and

Medicine, 5, 155-161.

Office of National Statistics. (2000). Psychiatric morbidity among adults living in private

households. London.

Office of National Statistics. Suicide Rate in the United Kingdom 2011. Retrieved on 15 Jan

2012, from http://www.ons.gov.uk/ons/dcp171778_295718.pdf.

Oorschot, M., Kwapil, T., Delespaul, P., & Myin-Germeys, I. (2009).Momentary assessment

research in psychosis. Psychological Assessment, 21, 498-505.

196
Oquendo, M. A., Halberstam, B., & Mann, J. J. (2003). Risk factors of suicidal behavior: The

utility and limitations of research instruments. In M. B. First (Ed.), Standardized

evaluation in clinical practice (pp. 103-130). Washington, DC: American Psychiatric

Publishing.

Oquendo, M., Bongiovi-Garcia, M., Galfalvy, H., Goldberg, P., Grunebaum, M., Burke, A.,

& Mann, J. (2007). Sex Differences in Clinical Predictors of Suicidal Acts After

Major Depression: A Prospective Study. American Journal of Psychiatry, 164, 134-

141.

Orbach, I., Bar-Joseph, H., & Dror, N. (1990). Styles of problem solving in suicidal

individuals. Suicide and Life-Threatening Behaviour, 20, 56 – 64.

Palmer, B.A., Pankratz, V.S., & Bostwick, J.M. (2005). The lifetime risk of suicide in

schizophrenia: a re-examination. Archive of General Psychiatry, 62 (3), 247–253.

Palmier-Claus, J., Taylor, P.J., Gooding, P., Dunn. G., & Lewis, S. (2011). Affective

variability predicts suicidal ideation in individuals at ultra-high risk of developing

psychosis: An experience sampling study. British Journal of Clinical Psychology, in

press.

Pinto, A. & Whisman, M. A. (1996). Negative affect and cognitive biases in suicidal and

nonsuicidal hospitalized adolescents. Journal of the American Academy of Child and

Adolescent Psychiatry, 35, 158–165.

Pirkis, J. & Burgess, P. (1998). Suicide and recency of health care contacts. A systematic

review. The British Journal of Psychiatry, 173, 462–74.

Platt, J.J., & Spivack, G. (1972). Problem-solving thinking of psychiatric patients. Journal of

Consulting and Clinical Psychology, 39, 148-151.

197
Platt, J.J., & Spivack, G. (1975). Manual for the means-ends problem solving procedure.

Philadelphia: Department of Mental Health Services, Hahnemann Community

Mental Health/Mental Retardation Center.

Pollock, L. R., & Williams, J. M. G. (2001). Effective problem solving in suicide attempters

depends on specific autobiographical recall. Suicide and Life-Threatening Behavior,

31, 386–396.

Pollock, L. & Williams, J.M.G. (2004). Problem solving in suicide attempters. Psychological

Medicine, 34, 163-167.

Pompili, M., Lester, D., Grispini, A., Innamorati, M., Calandro, F., Iliceto, P. et al. (2009).

Completed suicide in schizophrenia: Evidence from a case-control study. Psychiatry

Research, 167, 251-257.

Power, P. (2010). Suicide prevention in early psychosis. In French, P., Smith, J., Shiers, D.,

Reed, M. and Rayne, M. (Eds.) Promoting Recovery in Early Psychosis: A Practice

Manual. London: Wiley Blackwell

Power, P. & Robinson, J. (2009). Suicide prevention in first-episode psychosis. In McGorry,

P. & H. Jackson (Eds), The Recognition and Management of Early Psychosis: A

Preventative Approach. Cambridge University Press.

Reinecke, M.A., DuBois, D.L., & Schultz, T.M. (2001). Social problem solving, mood, and

suicidality among inpatient adolescents. Cognitive Therapy Research, 25 (6),743–

756.

Rihmer, Z. (2005). Prediction and prevention of suicide in bipolar disorder. Clinical

Neuropsychiatry, 2, 48 - 54.

Rihmer, Z. (2007). Suicide in mood disorders. Current Opinion in Psychiatry, 20, 17 – 22.

198
Riedesser, P. (2004). Psychosis as a traumatic event. In D. Burgin & Heinermeng (Eds.),

Childhood and adolescent psychosis (pp.61-66). Switzerland: Karger.

Robinson, J., Harris, M.G., Harrigan, S.M., Henry, L.P., Farrelly, S., Prosser, A., Schwartz,

O., Jackson, H., & McGorry, P.D. (2010). Suicide attempt in first-episode psychosis:

A 7.4 year follow-up study. Schizophrenia Research, 116 (1), 1–8.

Rooke, O. & Birchwood. M. (1998). Loss, humiliation and entrapment as appraisals of

schizophrenic illness: A prospective study of depressed and non-depressed patients.

British Journal of Clinical Psychology, 37, 259–268.

Rush Jr, J., First, M.B., & Blacker, D. (2008). Handbook of Psychiatric Measures. (2nd ed.).

Washington, DC. American Psychiatric Publishing, Inc

Sadowski, C. & Kelly, M. (1993). Social problem solving in suicidal adolescents. Journal of

Consulting and Clinical Psychology, 61,121–127.

Saha, S., Chant, D., & McGrath, J. (2007). A systematic review of mortality in schizophrenia:

is the differential mortality gap worsen over time? Archive of General Psychiatry,

64, 1123-1131.

Salter, D. & Platt, S. (1990). Suicidal intent, hopelessness, and depression in a parasuicide

population: The influence of social desirability and elapsed time. British Journal of

Clinical Psychology, 27, 247-258.

Schotte, D. E. & Clum, G. A. (1982). Suicide ideation in a college population: A test of a

model. Journal of Consulting and Clinical Psychology, 50, 690-696.

Schotte, D. E. & Clum, G. A. (1987). Problem-solving skills in suicidal psychiatric patients.

Journal of Consulting and Clinical Psychology, 55, 49–54.

199
Sharpley, M., Hutchinson, G., McKenzie, K., & Murray, R.M. (2001) Understanding the

excess of psychosis among the African–Caribbean population in England. Review of

current hypotheses. British Journal of Psychiatry Supplement, 40, 60–8.

Sidley G. L, Calam, R., Wells, A., Hughes, T., & Whitaker, K. (1999). The prediction of

parasuicide repetition in a high-risk group. British Journal of Clinical Psychology,

38, 375-386.

Stone, A. A., & Shiffman, S. (1994). Ecological momentary assessment (EMA) in behavioral

medicine. Annals of Behavioral Medicine, 16, 199-202.

Swendsen, J., Ben-Zeev, D., & Granholm, E. (2011). Real-time electronic ambulatory

monitoring of substance use and symptom expression in schizophrenia. American

Journal of Psychiatry, 168, 202-9.

Tarrier, N., Khan, S., Cater, J., & Picken, A. (2007). The subjective consequences of suffering

a first episode psychosis: trauma and suicide behaviour. Social Psychiatry and

Psychiatric Epidemiology, 42, 29–35.

Teasdale, J. D. (1988). Cognitive vulnerability to persistent depression. Cognition and

Emotion, 2, 247–274.

Teasdale, J.D. & Barnard, P.J. (1993). Affect, cognition and change: Re-modelling depressive

thought. Hove, UK: Lawrence Erlbaum Associates Ltd.

Thewissen, V., Bentall, R.P., Oorschot, M., Campo, J.A., van Lierop, T., van Os, J., & Myin-

Germeys, I. (2011). Emotions, self-esteem, and paranoid episodes: an experience

sampling study. British Journal of Clinical Psychology, 50, 178-195.

Tiihonen, J., Wahlbeck, K., Lonnqvist, J. et al. (2006). Effectiveness of antipsychotic

treatments in a nationwide cohort of 2230 patients in community care after first

200
 hospitalisation due to schizophrenia and schizoaffective disorder: Observational

follow up study. British Medical Journal, 333, 224.

Tversky, A. & Kahneman, D. (1974). Judgment under Uncertainty: Heuristics and Biases.

Science, New Series, 185 (4157), 1124-1131.

Upthegrove, R., Birchwood, M., Ross, K., Brunett, R., McCollum, R., & Jones, L. (2010).

The evolution of depression and suicidality in first episode psychosis. Acta

Psychiatrica Scandinavica 122 (3), 211-218.

Van der Does, A.J.W. & Williams, J.M.G. (2003). The Leiden Index of Depression

Sensitivity, Revised version (LEIDS-R). http://www.dousa.nl/publications

Varese, F., Udachina, A., Myin-Germeys, I., Oorschot, M., & Bentall,R. (2011). The

relationship between dissociation and auditory verbal hallucinations in the flow of

daily life of patients with psychosis. Psychosis, 3, (1), 14-28.

Velten, E. (1968). A laboratory task for induction of mood states. Behaviour Research and

Therapy, 6, 473–482.

Verdoux, H., Liraud, F., Gonzales, B., Assens, F., Abalan, F., & van Os, J., (2001). Predictors

and outcome characteristics associated with suicidal behaviour in early psychosis: A

two-year follow-up of first-admitted subjects. Acta Psychiatrica Scandinavica, 103

(5), 347–354.

Verma, S., Subramaniam, M., Abdin, E., Poon, L. Y., & Chong, S. A. (2012). Symptomatic

and functional remission in patients with first-episode psychosis. Acta Psychiatrica

Scandinavica, 126, 282–289.

201
Walsh, E., Harvey, K., White, I., et al. (2001) Suicidal behaviour in psychosis: prevalence and

predictors from a randomised controlled trial of case management: report from the

UK700 trial. British Journal of Psychiatry, 178, 255-260.

Watkins, E. & Baracaia, S. (2002). Rumination and social problem-solving in depression.

Behaviour Research and Therapy, 40, 1179-1189.

Wetzel, K.D., Margulies, T., Davis, R. et al. (1980). Hopelessness, depression, and suicide

intent. Journal of Clinical Psychiatry, 41, 159—160.

Williams, J. M. G. (1996). Depression and the specificity of autobiographical memory. In

Remembering Our Past: Studies in Autobiographical Memory (ed. D. Rubin) (pp.

244-267). Cambridge University Press: Cambridge.

Williams, J. M. G. (2001). Suicide and attempted suicide. Understanding the cry of pain.

London: Penguin.

Williams, J.M.G., Barnhofer, T., Crane, C., & Beck A.T. (2005). Problem solving deteriorates

following mood challenge in formerly depressed patients with a history of suicidal

ideation. Journal of Abnormal Psychology, 114 (3), 421–431.

Williams, J. M., Crane, C., Barnhofer, T., Van der Does, A. J., & Segal, Z. V. 2006.

Recurrence of suicidal ideation across depressive episodes. Journal of Affective

Disorders, 912 (3), 189-194.

Williams, J. M. G., Van der Does, A. J.W., Barnhofer, T., Crane, C., & Segal, Z. V. (2008).

Cognitive reactivity, suicidal ideation and future fluency: investigating a differential

activation theory of suicidality. Cognitive Therapy and Research, 32, 83–104.

World Health Organization. (2012). Public Health Action For the Prevention of Suicide.

Retrieved on January 20, 2012, from http://apps.who.int/iris/bitstream/10665/75166/

202
 1/9789241503570_eng.pdf.

Wunderink, L., Sytema, S., Nienhuis, F.J., & Wiersma, D. (2009). Clinical recovery in first-

episode psychosis. Schizophrenia Bulletin, 35 (2), 362-369.

Zahl, D.L. & Hawton, K. (2004). Repetition of deliberate self-harm and subsequent suicicde

risk: Long-term follow-up study of 11,583 patients. British Journal of Psychiatry,

185, 70-75.

203
Appendices

204
 APPENDIX 1. PARTICIPANT INFORMATION SHEET

Participant Information Sheet

Study Title

The association between daily hassles, low mood, and hopelessness in patients with psychosis: A move
towards validating the Differential Activation Hypothesis of suicidal relapse and recurrence using the
Experience Sampling Method

You are being invited to take part in a research study. Before you decide whether or not to take part, it
is important to understand why this research study is being carried out and also what it involves.
Please take time to read the following information carefully and discuss it with others if you wish. You
are welcome to ask us any questions and our contact details are available at the end of this information
sheet.

What is the purpose of the study?

There are two main aims to this study. First, is to examine the effect of mood on the way we look into
our future and how we solve common day to day problems. Second, is to look at the effect of daily life
hassles on your day to day mood and thoughts.

Do I have to take part?

Participation in this study is completely voluntary. You are free to – (a) decline to participate, (b)
refuse to answer any individual question, or (c) withdraw your participation at any time without giving
a reason, and without my medical care or legal rights being affected.

What will I have to do, if I take part?

This research involves two separate studies. You may choose to take part only in study 1, study 2, or
both. If you do not wish to take part at all then there is no need to return the reply form from your
letter of invitation.

Study 1 – Future Thinking (FT) task and Means Ends Problem Solving (MEPS) tasks

Assessments: Before the main part of the experiment starts you will undergo a series of assessments.
These will measure your current level of depression, suicidal thinking, hopelessness and future
thinking. The assessments are conducted in the form of questionnaires and interview. Each test will
take between 10-15 minutes to complete.

Future Thinking Task: You will be asked to think of possible future experiences that will occur over 3
different time periods (next week including today, next year, & next 5 to 10 years). You will then be
asked to think of future experiences under two different conditions (negative and positive). You will
be given 1 minute to generate as many responses as you can for each of the time period and
conditions. (Total duration: 15 minutes)

205
Means-Ends Problem Solving Task: You will be presented with problem scenarios on cards which will
be simultaneously read aloud by the experimenter. Each scenario will outline an initial situation in
which there is a problem to be solved and a desired endpoint. You will be given 2 minutes to describe
the most effective strategy for solving the problem. (Total duration: 20 minutes)

Note: There are no right and wrong answers for the FT & MEPS tasks, they are both relatively easy to
complete.

Musical Mood Induction Procedure: After completing the FT and MEPS tasks, a musical mood
induction procedure will then be performed. This will consist of listening to a sad music and reading
cards containing sad statements. The purpose of the procedure is to induce a sad mood. (Total
duration: 8-10 minutes)

Following this procedure, you will be asked to complete the same tasks that you did prior to the mood
induction procedure.

Study 2 - Experience Sampling Method (ESM)

Assessments: Before the main part of the experiment starts you will undergo a series of assessments.
These will measure your current level of depression, suicidal thinking, hopelessness and future
thinking, mood, rumination, response style, and life events. The assessments are conducted in the form
of questionnaires and interview. Each test will take between 10-15 minutes to complete. (Note: Some
of the assessments completed on study 1 will not be repeated.)

ESM procedure: You will be asked to assess your mood, thoughts, and activities in your day to day
environment (e.g. at home, at work). You will be given a digital wrist watch and 6 ESM questionnaire
diaries; one diary for each day of the study. Over the 6 days of study, the digital watch will emit 10
randomised beeps between the hours of 07.30am – 10.30pm, to inform you to fill out the
corresponding page in the ESM questionnaire diary. It takes about 2 minutes to complete each
questionnaire. The questionnaire assesses your current thought, mood, self-worth, future thinking,
psychotic experience, location, activity, physical needs and substance use. (Total duration: Each diary
questionnaires = 2 minutes; 10 questionnaire per day (10 x 2min) = 20 minutes; 6 days of diary
assessments (6 X 20min) = 120 minutes).

Note: You will only be expected to complete the diary assessments whenever it is possible & safe to
do so. For instance, if you usually get up at around 10am and do not want to be disturbed when the
watch emits a beep sound, you could put the watch in another room, or hide it in a drawer. The same
thing applies should you wish to go to bed earlier than 10.30pm. Also, you will not be expected to
pause from your day to day activity to fill in your diary unless it is safe and possible to do so (e.g.
cycling & driving).

What about my expenses?

If you take part in the experiment your transportation costs to and from the pre-assessment venue will
be met. Unfortunately, we cannot refund petrol costs for your own personal vehicle, but can reimburse
you for public transport cost (on provision of a receipt/bus or train tickets).

This sounds really complicated, will I get any help?

Yes, when you have expressed an interest in taking part in the study, we will contact you, allowing
you to ask any questions and address any concerns or worries you have about the study. You will be
given a brief session on how you take part in the Experience Sampling Method study. A researcher

206
will guide you through a sample questionnaire, briefing you in how to use the watch, and provide all
the guidance you will require. In the case of an emergency during the experiment, the researcher will
be contactable by phone (e.g. problems with watch, diary loss)
What are the possible benefits of taking part?

You will be paid £20 upon completion of Study 1 (MEPS & FT study) and £30 upon completion of
Study 2 (ESM study) in appreciation of your time and effort.

When your participation is complete, you will be given an opportunity to learn about this research,
which may be useful in understanding yourself and others. By keeping a diary for 6 days (study 2),
you may gain some insight on how your thoughts, activities, and events could make you feel a certain
way (e.g. sad, cheerful, etc.). We do not, however, guarantee that everyone would benefit from the
study as the daily life experiences of each individual will vary from person to person.

All in all, whilst we cannot promise that this study will help you, we hope that our results will add to
the knowledge about daily life hassles, hopelessness, and low mood.

What are the possible risks of taking part?

When filling out questionnaires (for study 1 & 2), you may come across a question(s) that you find
unpleasant. For instance, for study 1 you will be asked to think about possible negative events in your
future. A couple of our questionnaires have questions about past events or occasions when you were
feeling quite low, which you might also find uncomfortable.

Some of you may find study 2 slightly challenging or demanding as you will be need to complete
your diary assessments at random times of the day. The diary assessments are especially challenging
during the first day but once you get used to doing it, each questionnaire should only take no more
than 2 minutes to complete.

If you want to seek help or wish to discuss your concerns further, support and assistance will be
available via your Early Intervention Service (EIS) key worker. Counselling from an appropriate
professional will also be offered if required.

What will happen once I have finished taking part in the study?

You will firstly be given an in depth debrief of the aims of the study, and when the data analysis is
completed, a copy of the report will be issued to you.

Will my taking part keep confidential?

We recommend that your GP should know that you are taking part in this research. If you are happy
for us to tell them, we will write them a letter. Nobody else will know about your participation in the
study, and all results will be made anonymous (that is, your name will not be on them).

You will be assigned a code number which will protect your identity. All data will be kept in secured
files, in accord with the standards of the NHS Research Ethics. Only the researchers involved in this
study and those responsible for research oversight will have access to the information you provide.
There will no identifying information (e.g. name, address, & telephone number) in your questionnaires
so no one will be able to know how you did in your assessments. Your signed consent form will be
kept completely separate from your paper-based assessments.

207
Finally, it is no individual person's responses that interest us; we are studying the association between
low mood, hopelessness, and daily life hassles between clinical groups in general, so your name and
any other identifying information will not appear on the final report.

What will happen to the results of the study?

The results of the study will be published as an internal and external report, being made available to
the educational supervisors of the student conducting this research at the University of Birmingham.
The study may also be external published through publication to a scientific journal. However, your
anonymity will be preserved at all stages of this process.

Who is organising and funding the research?

This research is organised jointly by the University of Birmingham and Birmingham and Solihull
Mental Health Trust: Early Intervention Service. The study has been reviewed by Birmingham South
Research Ethics Committee and has been ethically approved (Insert Ethics Approval/Ref Number).

What happens now?

Think about all the information on this sheet and tell your Early Intervention Service (EIS) key worker
or the person who sent you the sheet (please see reply form on the letter of invitation) whether you
want to learn more about the research. If you do, we will telephone you at home and offer you an
appointment for an assessment visit. If you are not sure about anything, you can ask questions at that
first visit. At the first visit, we will go through all the information on this sheet to make sure that you
understand it. We will then ask you to sign a consent form to agree to the research.

Where can I get more information?

If you have any further questions about this study, please feel free to contact any members of the
researcher team below:

 Donna Back
PhD Psychology Student
School of Psychology
University of Birmingham, Edgbaston, B15 2TT
0121 414 7209
 07985 882 878
 dbb759@bham.ac.uk

 Dr Chris Jackson
Consultant Clinical Psychologist
Early Intervention Services
1 Miller Street, Aston Birmingham, B6 4NF
 0121 301 1850 Fax: 0121 301 1851
 Chris.Jackson@bsmht.nhs.uk

 Prof Max Birchwood

Director of Early Intervention Services
1 Miller Street, Aston Birmingham, B6 4NF,
 0121 301 1850, Fax: 0121 301 1851
 M.j.birchwood.20@bham.ac.uk

208
Important contact points DURING the study:

If you have any concerns about the conduct of this study please contact:

 Dr Paul McDonald
Manager of Research and Development Unit
Birmingham & Solihull Mental Health NHS Trust
Suite P, Radclyffe House, 66/68 Hagley Road
Birmingham, B16 8PF
 0121 678 4326
 paul.mcdonald@bsmht.nhs.uk

For ESM diary study-related enquiries/concerns (e.g. faulty watch, missing diaries, etc.), please
contact:

 Donna Back (between 9am – 5pm only)

 0121 414 7209
 (Work mobile – tbc)
 dbb759@bham.ac.uk

For support and assistance (should you feel upset, unhappy, or have any concerns about your mental
health during and after the study), please contact your Early Intervention Service (EIS) key worker on:

 Early Intervention Services (between 9am – 5pm only)

 0121 301 1850

For out of hours or 24 hours advice and support, please contact:

 24 hours - Mental Health Services Switchboards

North or Heart of Birmingham

 0121 685 7300 or 0121 623 5500

QEPH and South Birmingham

 0121 678 2000

Solihull Patients
 0121 424 2000

 PALS (Patient Advice & Liaison Service)

 0800 953 0045

 Birmingham Focus Line

 0800 027 2127

THANK YOU for taking time to read this.

209
 APPENDIX 2. PARTICIPANT’S LETTER OF INVITATION

School of Psychology
513 Frankland Bldg
Edgbaston
Birmingham
B15 2TT

Tel No. 0121 414 7209

 dbb759@bham.ac.uk

>Patient’s Name<
>Patient’s Address<

>Date<

Dear __________________,

My name is Donna Back. I am a postgraduate student at the University of Birmingham. I am conducting a

research study on “The association between daily hassles, low mood, and hopelessness” as part of the
requirements of my PhD in Psychology, and I would like to invite you to participate. This study is jointly
sponsored by the University of Birmingham and Birmingham and Solihull Mental Health Trust. The main aim of
the study is to explore the relationship between daily life hassles and mood. If you do decide to participate, you
will be asked to complete several questionnaires and participate in the experiments explained in the attached
information sheet.

Your participation is confidential. The data collected from this study will be kept in a secure location at the
University of Birmingham, which only the research team has access to. The results of the study may be
published or presented at professional meetings, but your identity will not be revealed. Your participation is also
anonymous, which means that no one (not even the research team) will know what your answers are. So, please
do not write your name or any other identifying information on any of the questionnaires.

You will receive a total of £50.00 as an appreciation for your time and participation (Study 1 = £20.00 & Study 2
= £30.00). Your travel expenses will also be reimbursed upon proof of travel receipts or tickets (bus or train). If
you withdraw from the study prior to the conclusion, your reimbursement will be pro-rata (total amount due will
be divided by the number of hours spent).

Taking part in this study is voluntary. You are free to withdraw at anytime during the study without giving any
reason.

If you would like to participate, please read the attached information sheet for full details of the study. If you
have any questions about it or would like to discuss participating, please contact me using the details listed
below. Alternatively, you may complete and sign the attached reply form and hand it back to your key worker or
send it to the address given below. You do not need to reply if you do not want to participate in the study.

Thank you for your consideration.

With kind regards,

Donna Back

210
 APPENDIX 3. LETTER TO THE PARTICIPANT’S GP

School of Psychology
513 Frankland Bldg
Edgbaston
Birmingham
B15 2TT

Tel No. 0121 414 7209

>GP’s Name<
>GP’s Address<

Dear Dr ………………….,

Re: Patient’s Name______________________

Date of Birth ______________________
NHS No ______________________

Study Title: “The association between daily life hassles, low mood, and hopelessness in
patients with psychosis”

Your patient is participating in the above study on daily life hassles, low mood, and hopelessness. The
study will involve completion of a number of questionnaires, simple problem-solving and future
directed thinking tasks, but will not involve any changes in their treatment. A copy of the participant
information sheet is enclosed for your reference.

If you require any further information, please contact me on the numbers above, or the Principal
Investigator, Prof Max Birchwood on 0121 301 1850.

Yours sincerely,

Donna Back
Chief Investigator/PhD Student

(On behalf of the study investigators)

Cc: >Patient’s name<

211
 APPENDIX 4. PARTICIPANT CONSENT FORM (MEPS & FT STUDY)

Participant Consent Form

Study title: The association between daily hassles, low mood, and hopelessness in patients with psychosis:
A move towards validating the Differential Activation Hypothesis of suicidal relapse and recurrence using
the Experience Sampling Method

By signing this informed consent form you are indicating that you understand the nature of the research study
and that you agree to participate in the research.

1. I confirm that I have read and understood the information sheet dated 25th March 2009
(version 2) for the above study and have had the opportunity to discuss the details with
………………………. and ask questions.

2. I understand that my participation is voluntary and that I am free to withdraw at any time,
without giving any reason, and without my medical care or legal rights being affected.

3. I understand that my participation will be anonymous (that is, my name will not be linked
with any data I give) and that all information I provide will remain confidential.

4. I also understand that relevant sections of my medical notes and data collected during the
study may be looked at by members of the research team, from regulatory authorities or
from the NHS Trust where it is relevant to my taking part in this research. I give permission
for these individuals to have access to my records.

5. I agree to communication with my GP about my participation in the research.

5.a. I would like to be copied in to all such correspondence

6. I hereby fully and freely consent to participate in the above Study 1- Future Thinking
Task and Means Ends Problem Solving task, which has been fully explained to me.

I have read and understood the statements above, and voluntarily sign this form. I further acknowledge
that I have received an offer of a copy of this consent form.

Volunteer _____________________________
Signature _____________________________ Date ___________________________
*If you wish to be told the results of this research, please tick here and provide your contact details at the
back of this form.

Investigator Donna Back___________________

Signature _____________________________ Date ___________________________

Person taking consent _______________________

Signature _____________________________ Date ___________________________

212
 APPENDIX 5. PARTICIPANT CONSENT FORM (ESM STUDY)

Participant Consent Form

Study title: The association between daily hassles, low mood, and hopelessness in patients with psychosis:
A move towards validating the Differential Activation Hypothesis of suicidal relapse and recurrence using
the Experience Sampling Method

By signing this informed consent form you are indicating that you understand the nature of the research study
and that you agree to participate in the research.

1. I confirm that I have read and understood the information sheet dated 25 th March 2009
(version 2) for the above study and have had the opportunity to discuss the details with
………………………. and ask questions.

2. I understand that my participation is voluntary and that I am free to withdraw at any time,
without giving any reason, and without my medical care or legal rights being affected.

3. I understand that my participation will be anonymous (that is, my name will not be linked
with any data I give) and that all information I provide will remain confidential.

4. I also understand that relevant sections of my medical notes and data collected during the
study may be looked at by members of the research team, from regulatory authorities or from
the NHS Trust where it is relevant to my taking part in this research. I give permission for these
individuals to have access to my records.

5. I agree to communication with my GP about my participation in the research.

5.a. I would like to be copied in to all such correspondence

6. I hereby fully and freely consent to participate in the above Study 2- The Experience Sampling
Method, which has been fully explained to me.

I have read and understood the statements above, and voluntarily sign this form. I further acknowledge
that I have received an offer of a copy of this consent form.

Volunteer _____________________________
Signature _____________________________ Date ___________________________

*If you wish to be told the results of this research, please tick here and provide your contact details at the
back of this form.

Investigator Donna Back___________________

Signature _____________________________ Date ___________________________

Person taking consent _______________________

Signature _____________________________ Date ___________________________

213
APPENDIX 6. COLUMBIA SUICIDE HISTORY FORM

214
215
216
217
 APPENDIX 7. LEIDEN INDEX OF DEPRESSION SCALE – REVISED

__________________________________________________________________________________

Instructions
Below are a number of statement that may apply to you to a lesser or greater extent.
Almost every statement concerns your thoughts about a certain matter at time when you feel down or
when you are in a low mood. This does not mean a seriously depressed mood or true depression. Your
task is to indicate the extent to which the statements apply to you when you feel somewhat sad.

Try to imagine the following situation when filling out this questionnaire.
It is certainly not a good day, but you don’t truly feel down or depressed.
Perhaps your mood is an early sign of something worse, but things might improve in the next
day or two.
On a scale of 0 to 10 (0 = not at all; 10 = extremely sad; 6 and above = a truly depressed
mood), you would choose 3 or 4 to describe your mood.
This scale looks like this:

1 2 3 4 5 6 7 8 9 10
not at all sad somewhat sad depressed extremely sad

Please try to imagine yourself in the above situation, for instance by thinking back to the last time you
felt somewhat sad (score 3 or 4).

{Now take some time to imagine such a situation}

To what extent are you able to imagine such a situation? O well

O somewhat
O not at all

Now proceed to the next question (even if you find it difficult to imagine yourself in such a situation).
_________________________________________________________________________________
Behaviour Research & Therapy 40; 105-120 (2002)
Revised version © 2003, Willem Van der Does & Mark Williams
__________________________________________________________________________________

218
This applies to me....... (please circle)

not a bit mode- strongly very

at all rately strongly
1 I can only think positive when I am in a good 0 1 2 3 4
mood.
2 When in a low mood, I take fewer risks. 0 1 2 3 4
3 When I feel sad, I spend more time thinking about 0 1 2 3 4
what my moods reveal about me as a person.
4 When in a sad mood, I am more creative than 0 1 2 3 4
usual.
5 When I feel down, I more often feel hopeless about 0 1 2 3 4
everything.
6 When I feel down, I am more busy trying to keep 0 1 2 3 4
images and thoughts at bay.
7 In a sad mood, I do more things that I will later 0 1 2 3 4
regret.
8 When I feel sad, I go out and do more pleasurable 0 1 2 3 4
things.
9 When I feel sad, I feel as if I care less if I lived or 0 1 2 3 4
died.
10 When I feel sad, I am more helpful. 0 1 2 3 4
11 When I feel sad, I am less inclined to express 0 1 2 3 4
disagreement with someone else.
12 When I feel somewhat depressed, I think I can 0 1 2 3 4
permit myself fewer mistakes.
13 When I feel down, I more often feel overwhelmed 0 1 2 3 4
about things.
14 When in a low mood, I am more inclined to avoid 0 1 2 3 4
difficulties or conflicts.
15 When I feel down, I have a better intuitive feeling 0 1 2 3 4
for what people really mean.
16 When in a sad mood, I become more bothered by 0 1 2 3 4
perfectionism.
17 When I feel sad, I more often think that I can make 0 1 2 3 4
no one happy.
not a bit Mode- strongly very
at all rately strongly

Please continue to the next page.

219
 This applies to me....... (please circle)
not a bit mode- strongly very
at all rately strongly
18 When I feel bad, I feel more like breaking things. 0 1 2 3 4
19 I work harder when I feel down. 0 1 2 3 4
20 When I feel sad, I feel less able to cope with 0 1 2 3 4
everyday tasks and interests.
21 In a sad mood, I am bothered more by aggressive 0 1 2 3 4
thoughts.
22 When I feel down, I more easily become cynical 0 1 2 3 4
(blunt) or sarcastic.
23 When I feel down, I feel more like escaping 0 1 2 3 4
everything.
24 When I feel sad, I feel more like myself. 0 1 2 3 4
25 When I feel down, I more often neglect things. 0 1 2 3 4
26 When I feel sad, I do more risky things. 0 1 2 3 4
27 When I am sad, I have more problems 0 1 2 3 4
concentrating.
28 When in a low mood, I am nicer than usual. 0 1 2 3 4
29 When I feel down, I lose my temper more easily. 0 1 2 3 4
30 When I feel sad, I feel more that people would be 0 1 2 3 4
better off if I were dead.
31 When I feel down, I am more inclined to want to 0 1 2 3 4
keep everything under control.
32 When I feel sad, I spend more time thinking about 0 1 2 3 4
the possible causes of my mood.
33 When in a sad mood, I more often think about how 0 1 2 3 4
my life could have been different.
34 When I feel sad, more thoughts of dying or 0 1 2 3 4
harming myself go through my mind.
not a bit Mode- strongly very
at all rately strongly

Please check whether all items are answered. Thank you.

220
 APPENDIX 8. CALGARY DEPRESSION SCALE FOR SCHIZOPHRENIA

CALGARY DEPRESSION SCALE

1. DEPRESSION.
How would you describe your mood over the last two weeks?
Do you keep reasonably cheerful or have you been very depressed or low spirited recently?
In the last two weeks how often have you (own words) every day? all day?
0 Absent
1 Mild Expressed some sadness or discouragement on questioning.
2 Moderate Distinct depressed mood persisting up to half the time over last two weeks, present
daily.
3 Severe Markedly depressed mood persisting daily over half the time, interfering with
normal motor and social functioning.

2. HOPELESSNESS.
How do you see the future for yourself?
Can you see any future or has life seemed quite hopeless?
Have you given up or does there still seem some reason for trying?
0 Absent
1 Mild Has at times felt hopeless over the last week but still has some degree of hope for the
future.
2 Moderate Persistent, moderate sense of hopelessness over last week. Can be persuaded to
acknowledge possibility of things being better.
3 Severe Persisting and distressing sense of hopelessness.

3. SELF-DEPRECIATION.
What is your opinion of yourself compared to other people?
Do you feel better or not as good or about the same as most?
Do you feel inferior or even worthless?
0 Absent
1 Mild Some inferiority; not amounting to feelings of worthlessness.
2 Moderate Subject feels worthless, but less than 50% of the time.
3 Severe Subject feel worthless more than 50% of the time. May be challenged to
acknowledge otherwise.

4. GUILTY IDEAS OF REFERENCE.

Do you have the feeling that you are being blamed for something or even wrongly accused?
What about ? (Do not include justifiable blame or accusations; exclude delusions of guilt).
0 Absent
1 Mild Subject feels blamed but not accused less than 50% of the time.
2 Moderate Persisting sense of being blamed, and/or occasional sense of being accused.
3 Severe Persistent sense of being accused. When challenged acknowledges that it is not so.

221
5. PATHOLOGICAL GUILT.
Do you tend to blame yourself for little things you may have done in the past?
Do you think you deserve to be so concerned about this?
0 Absent
1 Mild Subject sometimes feels over guilty about some minor peccadillo, but less than 50%
of the time.
2 Moderate Subject usually, (over 50% of time) feels guilty about past, actions, the significance
of which he/she exaggerates.
3 Severe Subject usually feels he/she is to blame for everything that has gone wrong, even
when not his/her fault.

6. MORNING DEPRESSION.
When you have felt depressed over the last two weeks, have you noticed the depression being worse at
any particular time of day?
0 Absent No depression.
1 Mild Depression present but no diurnal variation.
2 Moderate Depression spontaneously mentioned to be worse in the morning.
3 Severe Depression markedly worse in morning, with impaired functioning which improved
in afternoon.

7. EARLY WAKENING.
Do you wake earlier in the morning than is normal for you?
How many times a week does this happen?
0 Absent No early wakening.
1 Mild Occasionally wakes (up to twice weekly) one hour or more before normal time to
wake or alarm time.
2 Moderate Often wakes early (up to five times weekly) one hour or more before normal time to
wake or alarm
3 Severe Daily wakes one hour or more before normal time.

8. SUICIDE.
Have you felt that life wasn’t worth living?
Did you ever feel like ending it all?
What did you think you might do?
Did you actually try?
0 Absent
1 Mild Frequently thought of being better of dead, or occasional thoughts of suicide.
2 Moderate Deliberately considered suicide with a plan, but made no attempt.
3 Severe Suicidal attempt apparently designed to end in death (i.e. accidental discovery or
inefficient means).

9. OBSERVED DEPRESSION.
Based on interviewer’s observations during the entire interview.
The question “do you feel like crying?” used at an appropriate point in the interview, may elicit
information useful to this observation.
0 Absent
1 Mild Subject appears sad and mournful even during parts of the interview involving
effectively neutral discussion.
2 Moderate Subject appears sad and mournful throughout the interview, with gloomy
monotonous voice and is tearful or close to tears at times.
3 Severe Subject chokes on distressing topics, frequently sighs deeply and cries openly, or is
persistently in a state of frozen misery.

222
 APPENDIX 9. BECK HOPELESSNESS SCALE

This questionnaire consists of 20 statements. Please read the statements carefully one by
one. If the statement described your attitude for the past week including today, mark the “T”
indicating TRUE in the column next to the statement. If the statement does not describe your attitude,
mark the “F” indicating FALSE in the column next to this statement. Please be sure to read each
statement carefully.

Please be sure to read each statement carefully.

True False
1. I look forward to the future with hope and enthusiasm. T F
2. I might as well give up because there is nothing I can do about making T F
things better for myself.
3. When things are going badly, I am helped by knowing that they can’t stay T F
that way forever.
4. I can’t imagine what my life would be like in ten years. T F
5. I have enough time to accomplish the thing I most want to do. T F
6. In the future, I expect to succeed in what concerns me most. T F
7. My future seems dark to me. T F
8. I happen to be particularly lucky and I expect to get more of the good T F
things in life than an average person.
9. I just don’t get the breaks, and there’s no reason to believe that I will in the T F
future.
10. My past experiences have prepared me well for my future. T F
11. All I can see ahead is unpleasantness rather than pleasantness. T F
12. I don’t expect to get what I really want. T F
13. When I look ahead to the future, I expect to be happier than I am now. T F
14. Things just won’t work out the way I want them to. T F
15. I have great faith in the future. T F
16. I never get what I want, so it’s foolish to want anything. T F
17. It is very unlikely that I will get any real satisfaction in the future. T F
18. The future seems vague and uncertain to me. T F
19. I can look forward to more good times than bad times. T F
20. There’s no use in really trying to get something I want because I probably T F
won’t get it.

223
 APPENDIX 10. INTERSEPT SCALE FOR SUICIDAL THINKING

Score
0 1 2 (1 to 2)
1. Wish to die None Weak Moderate to strong
2. Reasons for living vs. dying For living outweigh for About equal For dying outweigh for
dying living
3. Desire to make active suicide attempt None Weak Moderate to strong
4. Passive suicidal desire Would take precautions to Would leave life/death to Would avoid steps
save lives chance necessary to save or
maintain life
5. Frequency of suicidal ideation Rare or occasional Intemittent Accepting
6. Attitude towards ideation/wish Rejecting Ambivalent or indifferent Has no ability to control
impulses
7. Control over suicidal/acting out or delusions/ Has complete ability to Unsure of ability to control Has no ability to control
hallucinations of self-harm control impulses impulses impulses
8. Deterrents to active attempt (e.g. religious values, Would not attempt Some concerns about Minimal or no deterrents
family) because of deterrents deterrents
9. Reason for contemplating attempt To maintain the Combination of 0 and 2 Escape, solve problems
environment, revenge; get
attention
10. Method: Specificity/planning of contemplated Not considered or not Considered but details not Details worked out; well
attempt applicable worked out formulated plan
11. Expectancy/anticipation by patient of actual None Uncertain Yes
attempt
12. Delusions/Hallucinations of self-harm (including None Occasional Frequent
command hallucinations)
TOTAL SCORE

224
 Semi-Structured Interview

InterSePT Scale for Suicical Thinking

1. The items assess the extent of suicidal thoughts and their characteristics as well as the patient’s attitude towards them.
2. The scale should be rated on the basis of all information available to the rater.
3. Depending on the psychiatric status of the patients as well as the degree to which he/she is articulate, the rater has the option to follow different lines f
inquiry than those suggested by the questions provided below.
4. The general time frame for rating each of the items is the last 7 days.
5. If there is ambiguity, rate the highest rating for the week.
__________________________________________________________________________________________________________________

Possible Questions:
1. How are you feeling this week?
2. In the past week, have you ever thought about taking your life?
3. If so, how strong have these thoughts been?
4. How frequently have you had this thoughts this past week?
5. How strong would you say your wish to die is?
6. This past week, have you looked forward to taking your life?
__________________________________________________________________________________________________________________

7. Which has been stronger this past week – your reasons for living or your reasons for dying?
8. If you had been in a dangerous or life-threatening situation this past week, what actions would you have taken to save your life?
9. In this past week, have you been able to control your suicidal thinking or might you have made an attempt at any time?
10. Is there anything in your life that would have made taking your life this past week seem like a bad idea, for example, your religion, family,
etc.?
11. (IF PATIENT WAS SUICIDAL IN PAST WEEK) What have reasons been for thinking about taking your life during this past week? Do you
think there reasons are good ones?
12. If you have committed sucide this past week, how would you have done it?
13. In the past week, have you heard voices, commands or others telling you to take your life?

225
 APPENDIX 11. THE ESM TIME SAMPLING SCHEDULE

DAY No. TS1 TS2 TS3 DAY No. TS1 TS2 TS3
DAY 1 08:46:00 08:13:00 08:39:00 DAY 4 08:24:00 08:51:00 07:57:00
09:45:00 10:20:00 09:56:00 09:31:00 09:58:00 10:30:00
11:21:00 10:56:00 11:49:00 11:52:00 10:56:00 11:58:00
13:25:00 13:09:00 12:20:00 12:19:00 13:04:00 12:36:00
14:55:00 14:54:00 14:54:00 14:27:00 14:02:00 14:34:00
15:23:00 16:29:00 15:26:00 15:42:00 15:18:00 16:22:00
16:49:00 17:02:00 17:39:00 16:55:00 17:28:00 17:59:00
19:14:00 18:58:00 18:27:00 18:25:00 18:58:00 18:53:00
20:04:00 19:56:00 20:30:00 20:46:00 20:48:00 20:15:00
21:40:00 21:20:00 21:49:00 21:49:00 21:40:00 21:41:00

DAY 2 08:43:00 07:59:00 08:53:00 DAY 5 08:16:00 08:33:00 08:07:00

10:06:00 09:50:00 09:17:00 10:13:00 09:18:00 09:17:00
11:41:00 11:38:00 11:55:00 11:38:00 11:36:00 11:03:00
13:23:00 13:15:00 13:14:00 13:26:00 12:57:00 13:20:00
14:14:00 14:57:00 14:35:00 14:40:00 14:13:00 14:14:00
16:26:00 16:28:00 15:50:00 16:14:00 15:55:00 15:24:00
17:53:00 17:13:00 17:45:00 16:47:00 16:52:00 17:43:00
18:46:00 19:22:00 19:15:00 18:30:00 19:23:00 18:42:00
20:20:00 19:46:00 20:32:00 19:50:00 20:13:00 20:17:00
21:41:00 22:01:00 21:54:00 21:56:00 21:40:00 21:24:00

DAY 3 08:09:00 08:39:00 08:37:00 DAY 6 08:03:00 08:48:00 08:06:00

10:24:00 09:18:00 09:18:00 10:00:00 09:45:00 09:27:00
11:08:00 11:57:00 11:52:00 11:18:00 11:31:00 11:58:00
12:38:00 13:22:00 13:06:00 12:34:00 12:59:00 12:19:00
13:45:00 14:27:00 14:48:00 14:16:00 14:56:00 14:06:00
16:16:00 15:59:00 15:17:00 16:15:00 15:51:00 15:32:00
17:19:00 17:36:00 17:00:00 16:47:00 17:26:00 17:11:00
18:22:00 19:22:00 18:18:00 19:01:00 19:28:00 18:48:00
20:02:00 20:05:00 20:27:00 20:16:00 20:56:00 20:16:00
21:37:00 21:49:00 21:56:00 21:22:00 22:06:00 21:47:00

226
 APPENDIX 12. THE ESM DEBRIEFING FORM

Participant Number

Date: _____________________________
Interviewer: _____________________________
___________________________________________________________________________

1. Did the Experience Sampling influence…

Your mood Yes/No
If so, how?

Your activities Yes/No

If so, how?

Your thoughts Yes/No

If so, how?

Your contact with other people Yes/No

If so, how?

2. Did the Experience Sampling disturb you Yes/No

If so, how?

3. Was this an ordinary week (with respect to the complaints) Yes/No

If not, what was different?

4. Were there special events or problems during this week? Yes/No

If so, what?

227
5. Were there difficult items in the booklets? Yes/No
If so, which item/s?

6. Could you give a good reflection of your experiences?

Yes/No
If not, why not?

7. Is there anything you missed in the booklets? Yes/No

If so, what?

8. Did you take your medication during this period?

Yes/No
If not, what medication and why not?

Remarks:

___________________________________________________________________________

___________________________________________________________________________

___________________________________________________________________________

___________________________________________________________________________

___________________________________________________________________________

228
 APPENDIX 13. VISUAL ANALOGUE SCALE

VAS Mood Rating

Participant Number

Date

Place a vertical mark on the line below to indicate how you feel right now.

At this moment, I feel...

not at all extremely

Happy

not at all extremely

Despondent

229
 APPENDIX 14. THE MEPS TASK – PROBLEM SCENARIOS

Item no. Problem scenario

2 Heather loved her boyfriend very much, but they had many arguments. One day,
he (the boyfriend) left her. Heather wanted things to be better. The story ends
with everything fine between her and her boyfriend. Please begin your story
when her boyfriend left after an argument.

3 Mrs Philips came home after shopping and found that she had lost her watch.
She was very upset about it. The story ends with Mrs Philips finding her watch
and feeling good about it. Please begin your story when Mrs Philips realised that
she had lost her watch.

4 Caroline had just moved in that day to a new neighbourhood and didn’t know
anyone. Caroline wanted to have friends in this new neighbourhood. The story
ends with Caroline having many good friends and feeling at home in the
neighbourhood. Please begin your story with Caroline in her room, unpacking
boxes.

6 One day, Alice saw a beautiful man she had never seen before while eating in a
restaurant. She was immediately attracted to him. The story ends when they get
married. Please begin your story when Alice first sees the man in the restaurant.

8 Jane noticed that her friends seemed to be avoiding her. Jane wanted to have
friends and be liked. The story ends when Jane’s friends like her again. Please
begin your story when Jane first notices her friends avoiding her.

10 Jenny is having trouble getting along with her supervisor at work. Jenny is very
unhappy about this. The story ends with Jenny’s supervisor liking her. Please
begin your story when Jenny wasn’t getting along well with her supervisor at
work.

230
 APPENDIX 15. VELTEN NEGATIVE STATEMENTS

1. It seems such an effort to do anything.

2. I feel pessimistic about the future.
3. I have too many bad things in my life.
4. I have very little to look forward to.
5. I’m drained of energy, worn out.
6. I’m not as successful as other people.
7. Everything seems futile, pointless.
8. I just want to curl up and go to sleep.
9. There are things about me that I don’t like.
10. It’s too much of an effort even to move.
11. I’m absolutely exhausted.
12. The future seems just one string of problems.
13. My thoughts keep drifting away.
14. I get no satisfaction from the things I do.
15. I’ve made so many mistakes in the past.
16. I’ve got to really concentrate just to keep my eyes open.
17. Everything I do turns out badly.
18. My whole body has slowed down.
19. I regret some of the things I’ve done.
20. I can’t make the effort to liven myself up.
21. I feel depressed with the way things are going.
22. I haven’t any real friends anymore.
23. I do have a number of problems.
24. There’s no one I can really feel close to.
25. I wish I were somebody else.
26. I’m annoyed at myself for being so bad at making decisions.
27. I don’t make a good impression on other people.
28. The future looks hopeless.
29. I don’t get the same satisfaction out of things these days.
30. I wish something would happen to make me feel better.

231