Screening: Berhe Beyene (Email:)
Screening: Berhe Beyene (Email:)
Screening: Berhe Beyene (Email:)
Berhe B 1
What is screening?
Berhe B 2
Definition
Search for unrecognized disease or defect by means
of rapidly applied tests, examinations or other
procedures in apparently healthy individuals.
Berhe B 3
Screening Test Vs Diagnostic Test
Berhe B 4
Some Common Screening Tests
„ Pap smear for cervical dysplasia or cervical cancer
Fasting blood cholesterol for heart disease
Fasting blood sugar for diabetes
Blood pressure for hypertension
„ Mammography for breast cancer
PSA test for prostate cancer
Fecal occult blood for colon cancer
Ocular pressure for glaucoma
PKU test for phenolketonuria in newborns
Thyroid stimulating hormone(TSH) for hypothyroid and
hyperthyroid
Berhe B 5
Diseases Appropriate for Screening
1. Important public health problem – frequent or serious
Berhe B 7
Criteria for instituting a screening programme
Disease -Serious
-High prevalence of preclinical stage
-Natural history understood
-Long period between first signs and overt disease
Diagnostic test -sensitive and specific
-Simple and cheap
-safe and acceptable
-Reliable
Diagnostic and -Facilities are adequate
treatment -Effective, acceptable, and safe treatment
available
Berhe B 8
WHO Criterion for Screening
Is it a health problem?
Is there treatment?
Are there facilities in place?
Is it detectable pre-clinically?
Is there a suitable screening test?
Is the screening test acceptable to people?
Is the natural history of disease understood?
Are the costs acceptable?
Berhe B 9
Types of screening
There are different types of screening, each
with specific aims
a) Mass screening – involves screening of the
whole population
b) Multiple or multiphase screening – involves a
variety of screening tests on the same
occasion.
Berhe B 10
Types of screening cont’d
c) Targeted screening of groups with specific
exposures – is often used in environmental
and occupational health
d) Case-finding or opportunistic screening – is
restricted to patients who consult a health
practitioner for some other purpose
Berhe B 11
Validity of a Screening Test
Validity of a test is the ability to differentiate
accurately between those who have the disease and
those who do not(indicate which individuals have the
disease and which do not).
Sensitivity and Specificity are two measures of the
validity of a screening test.
Sensitivity and specificity can also be taken as
measures of validity for other tests which are applied
for diagnostic purposes.
Berhe B 12
Sensitivity
The ability of a test to correctly identify those
who have the disease
Berhe B 14
Two-by –two table for the calculation of validity of screening
tests
Definitive Diagnosis
Berhe B 15
Validity cont’d
Sensitivity = TP X 100 a
TP+FN 100
=Pr(T+|D+)
ac
Specificity = TN X100
d
TN + FP 100
=Pr(T−|D−) bd
Berhe B 16
Validity cont’d
E.g. A group women aged 40-64 assigned into two
groups, i.e. screened vs. comparison. The women in
the screened group were offered an initial screening
examination for breast cancer (mammography and
physical examination) followed by three follow-up
examinations at yearly intervals. The women in the
comparison group were not offered the screening
examination but rather continued to receive their
usual care.
Berhe B 17
Validity cont’d
Total of 64,810 screening examinations were performed
among the study populations. During the first 5 years of
observation, 132 breast cancers were diagnosed among
1,115 biopsies or aspirations that were recommended on
the basis of the results of the screening procedures.
Berhe B 18
Group discussion
Group of 3 to 4 members
Prepare a two by two table and calculate
measures of validity of the test?
Berhe B 19
Validity cont’d
Table 1: Sensitivity and specificity of breast cancer screening examination (Health Insurance
Plan of Great New York (HIP Program)
Breast cancer
Cancer cancer not Total
Confirmed confirmed
Berhe B 20
Validity cont’d
Sensitivity: Among women diagnosed with breast
cancer during the study period, approximately 75 %
tested positive on the screening procedure
Berhe B 22
Balance in choosing a screening test
• Disease serious & definitive • High risk with further
treatment exists diagnostic techniques
Sensitivity Specificity
Berhe B 23
Sensitivity and specificity at different levels
Cutoff points
Persons
A B C
True Negatives
50 80 120 Diastolic BP
Berhe B 24
Sensitivity and specificity...
Berhe B 25
Cut-points of sensitivities and specificities
Berhe B 28
Parallel testing
When screening tests are given in parallel, all are
administered at the same time, and persons with
positive results on any test are considered positive.
Berhe B 29
Sequential testing
When screening tests are given in series, an
initial screening test is administered, and only
persons with positive test are evaluated with
an additional screening procedure.
Berhe B 30
Sequential testing: Re-screen the Positives
from the first test
Subject is disease positive when test positive
in both tests
Berhe B 31
Predictive Value of a Screening Test
Predictive value is the ability of a test to predict the
presence or absence of disease from test results.
1. Predictive Value of a Positive Test (PVPT) or
Positive Predictive Value
probability that the person tested positive by this
specific test truly has the disease.
PVPT = TP X 100 a
TP+FP PV 100
ab
Berhe B 32
Predictive Value cont’d
2. Predictive Value of a Negative Test (PVNT) or
Negative Predictive Value
Berhe B 33
Predictive Value cont’d
Example: Refer table 1.
PVPT = 11.8 % means the probability that a woman who tested
positive on the screen actually had breast cancer is 11.8%.
PVNT = 99.9% means the probability that a woman who tested
negative truly did not have breast cancer is 99.9%.
Berhe B 34
Predictive Value cont’d
The higher the prevalence, the more likely it
is that a positive test is predictive of the
diseases i.e PVPT will be high.
Berhe B 35
Predictive Value cont’d
The more specific the test, the less likely an
individual with a positive test will be free from the
disease ( FP) and the greater the predictive value
positive.
Berhe B 36
Example 1 : The table below shows relation between results of liver
scan and diagnosis in 344 patients
Abnormal Normal
Normal 27 54 81
Berhe B 37
Based on the given conditions in the table calculate:
a) Sensitivity?
b) Specificity?
Berhe B 38
Solutions
Berhe B 39
c) PV+ = True positives X 100
Berhe B 40
Example 2:
90% = Tp/20
Tp = 0.9 x 20
Tp = 18
FN = 20-18
=2
Berhe B 42
Specificity = TN/total non- disease
95% = TN/980
TN = 0.95 x 980
TN = 931
FP = 980-931
= 49
Predictive value a positive test-
(PV+) = 18/18+49 x 100
= 27%
Predictive value a negative test-
(PV-) = 931/2+931 x 100
= 99.8%%
Berhe B 43
Example :
Disease Non-disease
Positive 9 49 58
Berhe B 44
Predictive value a positive test-
= 15.5%
= 99.89%%
Berhe B 45
Exercise 1:
Berhe B 46
Exercise 2
In a group of 500 children, 50 have a genuine hearing problem.
Of these, 45 fail the school hearing test, as do 30 of the children
with normal hearing (perhaps they had a cold on the day of the
test). Based on this scenario
Develop the two- by-two table
What percentage of children with a real hearing problem failed
the school test?
What percentage of children with normal hearing passed the
school test?
What percentage of children who failed the school hearing test
had a real hearing problem?
What percentage of children who passed the school hearing test
really did have normal hearing?
Berhe B 47
Reliability(Precision)
Berhe B 48
1. Percent agreement
„
Is the ability of a screening program to correctly classify
individuals either as truly affected or truly unaffected
It is proportion of correctly categorizing of
individuals among the total tested individuals
A perfect agreement occurs when:
b=0
c=0
Berhe B 49
Percent agreement…
Berhe B 50
Reliability…
1. Cohen’s Kappa
Difference between observed and expected
agreement expressed as a fraction of the
maximum difference and ranges between -1 to
1.
Is an appropriate reliability measure (or
measure of agreement) for a screening test
which gives a categorical result
It considers agreement that may occur by
chance alone
Berhe B 51
Measures of Reliability - Kappa
Berhe B 52
Cohen’s Kappa
Berhe B 53
Measures of Reliability – Kappa...
• Cohen’s kappa is thus the agreement adjusted for that expected by
chance
Self Interview Total
-administered Yes No
questionnaire Yes 61 2 63
No 6 25 31
an nt
Total 67 27 94
ch me
.
ce
by ree
te ag
ec he
d
xp s t
t e hu
ha is t
fo ppa
te k a
rt
jus ’s
ad hen
d
Co
Berhe B 54
Measures of Reliability - Kappa
Berhe B 55
Reliability…..
The source of variability that can affect the
reproducibility of results of screening test include:
Biological variation –varies considerably for a given
individual with time and other
Related to instrument- which relates to the reliability of
the instrument itself, such as standard mercury
sphygmomanometer for BP
Berhe B 56
Reliability…..
Intra-observer variability – differences in repeated
measurements by the same screener. The difference is due to
the changes (such as physiological, environmental, etc.)
occurring to that individual over that time period).
Berhe B 58
Evaluation of screening program
Two issues: Feasible and effective program . Both must be
considered carefully.
Berhe B 59
Feasibility
The feasibility of a screening program is
determined by a number of factors related to
program performance which measure:
The acceptability of the program to the
potential screenees, measured by
number of persons examined
proportion of the target population that is
screened.
Berhe B 60
Feasibility cont’d
cost-effectiveness
The costs of the screening program must be considered in terms
of
total costs
resources expended per detected case of the disease.
the subsequent diagnosis and treatment. Screening program must also
include provision for follow up of persons whose screening tests are
positive. This can be measured by:
Proportions of those with positive tests who are followed, diagnosed, and treated.
Berhe B 61
Feasibility cont’d
The yield (number of cases detected by the
screening program).
Persons with the disease
Yield = detected by the test
X 100
Total screened
TP
Yield = X 100
TP + FN + TN + FP
Berhe B 62
Strategies to Increase Yield of Screening
Select a test with high specificity
– High sensitivity >> Low false - > > High PV-
– High specificity > > Low false + > > High PV+
Select disease with high prevalence of pre-
clinical stage
Target high risk group for screening
Berhe B 63
Effectiveness
Evaluation of effectiveness of a screening program
must be based on measures that reflect the impact
of a program on the course of a disease.
Berhe B 64
Study designs used to evaluate screening programs
1) Correlational studies
• Have been used to examine trends in disease
rates in relation to screening frequencies
within a population, or to compare the
relationship between the frequencies of
screening and disease rates for different
populations
• These studies can suggest the possibility of a
benefit of a screening program, but they can
not test that hypothesis
Berhe B 65
Study designs cont…
2) Case control studies
Individuals with and without the disease are
compared with respect to their past exposure
(screening)
Berhe B 66
Study designs cont…
3) Cohort studies
• the case-fatality rate of those who chose to be
screened is compared with the comparative rate
among those whose diagnoses were symptom-
related.
• is the most frequently used approach.
• in interpreting the results of such studies, the
potential effects of the self-selection of participants
as well as lead-time and length bias must be taken
into account.
Berhe B 67
NB:
Lead-time bias: The period between when disease is
detected by screening and when it would have
become symptomatic and been diagnosed in the
usual way.
Berhe B 68
Study designs cont…
4) Randomized trial
• Since no comparability of screened and symptom-
diagnosed cases is the chief threat to validate, the
optimal assessment of the efficacy of screening
program derives from randomized trials
• The screening program is allocated at random by the
investigators after individuals have agreed to
participate in the trial
Berhe B 69
Study designs cont…
• The Health Insurance Plan (HIP) breast cancer screening
project, New York, was a randomized trial designed to
evaluate whether periodic breast cancer screening with
mammography and physical examination would result in
reduced breast cancer mortality among women aged 40 to 64
years.
Berhe B 70
Questions ?
Berhe B 71