Czac094
Czac094
Czac094
DOI: https://doi.org/10.1093/heapol/czac094
Advance access publication date: 3 November 2022
Original Article
Abstract
Vertical global health programmes often evaluate success with a narrow focus on programmatic outcomes. However, evaluation of broader
patient-centred and unintended outcomes is critical to assess impacts on patient choice and autonomy. Here, we evaluate the effects of a post-
partum intrauterine device (PPIUD) intervention on outcomes related to contraceptive method choice. The stepped-wedge cluster randomized
contolled trial (RCT) took place in five Tanzanian hospitals. Hospitals were randomized to receive immediate (Group 1; n = 11 483 participants)
or delayed (Group 2; n = 8148 participants) intervention. The intervention trained providers on PPIUD insertion and counselling. The evaluation
surveyed eligible women (18+, resided in Tanzania, gave birth at a study hospital) on provider postpartum contraceptive counselling during preg-
nancy or immediately postpartum. In our completed study, participants were considered exposed (n = 9786) or unexposed (n = 10 145) to the
intervention based on the location and timing of their birth (no blinding). Our secondary analysis examined differences by intervention exposure
on the likelihood of being counselled on IUD only, multiple methods, multiple method durations, a broad method mix; and on the number of
methods women were counselled across two samples: all eligible women, and only women who reported receiving any contraceptive coun-
selling. Among all eligible women, counselling on the IUD alone was 7% points higher among the exposed (95% confidence interal (CI): 0.02,
0.12). Among women who received any counselling, those exposed to the intervention were counselled on 1.12 fewer contraceptive methods
(95% CI: 0.10, 2.34). The likelihood of receiving counselling on any non-IUD method decreased among those exposed, while the likelihood of
being counselled on an IUD alone was 14% points higher among the exposed (95% CI: 0.06, 0.22), suggesting this intervention increased
IUD-specific counselling but reduced informed contraceptive choice. These findings underscore the importance of broad metrics that capture
autonomy and rights (in addition to programmatic goals) at all stages of health programme planning and implementation.
Keywords: Vertical health programmes, programme evaluation, Tanzania, family planning, quality of care, sexual and reproductive health and rights
© The Author(s) 2022. Published by Oxford University Press in association with The London School of Hygiene and Tropical Medicine.
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial License (https://creativecommons.org/
licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For
commercial re-use, please contact journals.permissions@oup.com
Health Policy and Planning, 2023, Vol. 38, No. 1 39
methods, and the methods offered should represent a broad clinics whose patients deliver at the referral hospitals. Hos-
range of contraceptive attributes so that women can choose pital doctors were trained on PPIUD insertion and removal,
the method that meets their individual needs (Senderowicz, while nurses and midwives in the satellite clinics were trained
2020). These attributes may include duration of use (short- to integrate PPIUD counselling into routine antenatal family
acting, medium-acting, long-acting and permanent), presence planning counselling. The FIGO/AGOTA initiative identi-
of hormones (hormonal and non-hormonal), provider depen- fied and trained a cascade of master trainers to carry out
dence (provider dependent and independent) and locus of a ‘training the trainer’ approach for counselling and inser-
control (female- and male-controlled) (Festin et al., 2016; tion. Counselling training sessions included ‘information on
Senderowicz, 2020). the advantages of PPIUD’, presentation of visual aids and
In this article, we seek to evaluate a vertical family planning role-playing of potential counselling scenarios (de Caestecker
programme implemented in Tanzania, looking beyond the et al., 2018). The training was also designed to include
intended programmatic outcomes of interest—PPIUD coun- content on method removal. Trained nurses and midwives
selling and uptake—to examine the programme’s impact on were then expected to provide family planning counselling
other reproductive health outcomes related to informed, full to women seeking antenatal care that included information
contraceptive choice. about the PPIUD as part of a wide range of contraceptive
methods.
The central FIGO team in London maintained a data ‘dash-
board’ to provide real-time feedback to clinicians and project
Materials and methods
leaders from the hospitals included in the PPIUD Project in
Overview of study and description of the Tanzania and the other intervention countries. The dashboard
intervention reported numbers of: (1) deliveries; (2) PPIUD insertions; (3)
We performed a secondary analysis of data from a large women counselled on PPIUD; (4) PPIUD removals; (5) women
cluster-randomized stepped-wedge PPIUD trial in Tanzania followed-up; (6) PPIUD expulsions; (7) providers trained to
to evaluate the impact of the PPIUD intervention on out- insert PPIUD; and (8) providers trained to counsel on PPIUD.
comes related to freedom of informed contraceptive method FIGO coordinators in London regularly communicated dash-
choice. Since the intervention was primarily focused on train- board statistics with each other, Tanzanian affiliate staff and
ing providers in hospitals, it was cluster randomized to avoid providers. Providers were given real-time feedback about their
contamination, while the stepped-wedge design allowed the progress towards meeting project goals, all tied to PPIUD
intervention to eventually reach all study hospitals. counselling and insertion, with the primary marker of success
The research described here is part of a broader evaluation being a calculation from dashboard numbers of the percent of
of the PPIUD Project implemented by the International Fed- all deliveries leading to a PPIUD insertion. A detailed descrip-
eration of Gynacology and Obstetrics (FIGO) (de Caestecker tion of the intervention can be found in de Caestecker et al.
et al., 2018). This initiative began in 2013, with the goal ‘to (2018).
address the gap in the continuum of maternal health care and This analysis is part of a mixed-methods sequence of learn-
to provide for the postpartum contraceptive needs of women ing motivated by the qualitative portion nested within this
by increasing the capacity of healthcare professionals to offer larger RCT. As results from in-depth interviews with women
PPIUDs by training community midwives, health workers, who received antenatal care under PPIUD intervention condi-
doctors, and delivery unit staff, as appropriate, in counselling tions began to reveal a tendency for counselling to focus on
and insertion of PPIUD’ (de Caestecker et al., 2018). In 2016, the IUD to the exclusion of other methods (Senderowicz et al.,
FIGO brought this programme to Tanzania via their national 2021), we turned to the large quantitative dataset to explore
affiliate, the Association of Gynaecologists and Obstetricians how widespread this phenomenon was among the broader
of Tanzania (AGOTA), with the goal of introducing and insti- study population. Our objective was to estimate the impact
tutionalizing PPIUD for those seeking antenatal care and of the PPIUD intervention on individual-level overall access
facility-based births. AGOTA implemented the FIGO PPIUD to information about contraceptive methods and a broad
Project at six public referral hospitals throughout Tanzania contraceptive method mix during antenatal, peripartum and
(Dodoma General Hospital in Dodoma, Muhimbili National immediate postpartum contraceptive counselling.
Hospital in Dar es Salaam, Mt. Meru Hospital in Arusha,
Tumbi-Pwani Regional Referral Hospital in Kibaha, Mbeya Research ethics
Zonal Referral Hospital in Mbeya and Sekou-Toure Regional
The study received human subjects research approval from the
Referral Hospital in Mwanza). Doctors and nurses were
National Institute of Medical Research (NIMR) in Tanzania
trained in counselling and insertion. Hospitals were selected
(protocol number: NIMR/HQ/R.8a/Vol.IX/2006), and ethical
to receive the intervention by AGOTA, the implementer of the
approval as exempt by the institutional review board at Har-
intervention, in order to provide coverage of PPIUD services
vard University (protocol number: IRB15–1605). Respon-
for different geographic regions of Tanzania among tertiary
dents provided written informed consent to be interviewed, or
care facilities. The size of facilities varied, with smaller referral
thumbprints and a witness’s signature if they could not sign
hospitals such as Mbeya having fewer providers (58 com-
their names.
bined junior doctors and trained Ob/Gyns) than the larger
Muhimbili National Hospitals (240 combined junior doctors
and trained Ob/Gyns). Trial design
The FIGO/AGOTA programme trained providers on cadre- A cluster-randomized stepped-wedge trial approach with a 1:1
appropriate skills and knowledge to support the implementa- allocation ratio for clusters was designed to evaluate FIGO
tion of PPIUD. AGOTA organized a series of trainings in the intervention sites in Nepal, Sri Lanka and Tanzania. A detailed
six referral hospitals and surrounding satellite antenatal care description of that overarching trial design can be found in
Health Policy and Planning, 2023, Vol. 38, No. 1 41
Canning et al., 2016. The trial is registered with clinical- provided informed consent. Consent to participate in the eval-
trials.gov NCT02718222. Because of substantial differences uation was sought at the individual level after cluster-level
between the implementation of both intervention and study randomization. The survey included questions on fertility
procedures between countries, results are presented here for desires; experiences with family planning counselling dur-
Tanzania alone. ing the antenatal period, peripartum and immediate post-
partum periods; perceptions of PPIUD and contraceptive
intentions.
Cluster selection, randomization, and steps
In Tanzania, AGOTA selected six large public referral hospi-
tals across different regions to receive the PPIUD intervention.
Outcomes of interest
Each hospital served as a cluster. There was no blinding.
Given the range of hospital sizes, we employed a strategy The predefined primary outcome of interest to this evaluation
of block randomization, in which clusters were matched on was the percent uptake of PPIUD, defined as the proportion
annual obstetric caseload into blocks of n = 2 clusters. Using of all women who received a PPIUD divided by the number
Stata v14, EP generated a random number for each cluster of women who delivered in one of the study hospitals over
and assigned the lower number within each pair to Group 1 the course of the study period. No subgroups were excluded
(Dodoma General Hospital in Dodoma, Muhimbili National from this end-point. Key predefined secondary outcomes as
Hospital in Dar es Salaam, and Mbeya Zonal Referral Hos- defined at the outset of the study were also focused on PPIUD-
pital in Mbeya) and the higher number to Group 2 (Mt. related outcomes (e.g. the percentage of women who receive
Meru Hospital in Arusha, Sekou-Toure Referral Hospital PPIUD counselling and the percentage of PPIUD acceptors
in Mwanza and Tumbi-Pwani Regional Referral Hospital in who have PPIUD expulsions). Predefined primary and sec-
Kibaha). According to the stepped-wedge design, Group 1 ondary outcomes pertained to the cluster level. Analyses of
hospitals received the early intervention, and Group 2 hospi- these outcomes have been performed and reported on else-
tals received the late intervention (Figure 1). After randomiza- where (Huber-Krum et al., 2019; Hackett et al., 2020; Pearson
tion, evaluators learned of a pre-existing PPIUD intervention et al., 2020).
at Sekou-Toure hospital in Mwanza (Group 2), and this The current secondary data analysis builds on the results of
hospital was dropped from the study. our concurrent nested qualitative study to expand the scope
Data collection began on 15 January 2016. Group 1 hospi- of this inquiry for this study. Results from semi-structured in-
tals were scheduled to begin PPIUD trainings after 3 months of depth interviews with women at antenatal clinics exposed to
baseline data collection, while Group 2 hospitals were sched- the PPIUD intervention suggested that antenatal family plan-
uled to receive the intervention after 9 months. Due to delays ning counselling was directive and biased to focus on the IUD
in training implementation and other logistical challenges, the to the exclusion of other methods. Here, we have developed a
actual timing of the rollout to each hospital varied slightly, set of quantitative measures to test whether these qualitative
and these delays are accounted for in the data analysis. The results can be expanded to the study population more gener-
intervention began in Group 1 on 15 April 2016. Group 2 ally. As such, the purpose of this analysis is to examine the
was due to begin on 15 September 2016, but project imple- effect of the intervention on a set of person-centred family
mentation actually began on 17 November 2016, approx- planning outcomes related to method mix and availability of
imately 2 months later than planned. The present analysis choice (World Health Organization, 2021). The five outcomes
uses quantitative data from the first contact with respon- of interest are measured on the individual level among women
dents in the immediate period following delivery of their index who reported having received any perinatal family planning
pregnancy. counselling and include: (1) likelihood of being counselled
on the IUD alone (and no other contraceptive methods);
(2) likelihood of being counselled on multiple (more than
Data collection and participants one) methods; (3) likelihood of being counselled on multiple
Women were eligible to participate if they had given birth method durations of use; (4) likelihood of being counselled on
at a study hospital during data collection, resided in Tanza- a broad contraceptive method mix; and (5) number of meth-
nia, and were over age 18. All eligible women were invited ods counselled on. Additionally, to assess whether losses to the
to participate. Trained data collectors collected survey data number of methods counselled were compensated for by gains
in the postnatal wards of study hospitals, administering pre- in counselling under the PPIUD intervention, we explored
programmed tablet-based questionnaires to all women who the total number of contraceptive methods on which women
42 Health Policy and Planning, 2023, Vol. 38, No. 1
received counselling among all women in the study, includ- (a premium service at some hospitals that offers patients
ing those who reported not receiving any family planning better amenities and a lower provider-to-patient ratio for
counselling. a higher cost) as fixed effects. Results are presented with
For all outcomes, women reported on their contraceptive P-values associated with standard errors adjusted using
counselling throughout the antenatal, perinatal and imme- cluster wild bootstrapping with Rademacher weights. This
diate postpartum continuum. Women reported whether they method is designed to correct for the inflation of preci-
had received counselling on each of the following contracep- sion associated with replications based on a small num-
tive methods at any point throughout the perinatal period: ber of clusters. We present intraclass correlation coeffi-
female or male sterilization, injectables, implants, oral contra- cients (ICC) with associated standard errors from logistic
ceptives, condoms, emergency contraceptives, diaphragm, cer- models.
vical mucus observation, calendar-based methods, lactational To assess sensitivity to model specification, we also con-
amenorrhoea, withdrawal, or another method. ducted multilevel mixed-effects regression models on the same
We calculated Outcome 1 as a binary variable (received outcomes of interest. We present the detailed methods and
counselling at any time on any method other than or in addi- results of these analyses in Appendix C.
tion to the IUD = 0, received counselling on only the IUD
and no other method = 1). Outcome 2 was a binary vari-
able (received counselling on only one method = 0, received Analytic sample
counselling on more than one method = 1). Outcome 3 was A total of 22 691 women who delivered during the study
calculated as a binary variable [received counselling on meth- period (15 January 2016–15 January 2017) in five hospitals
ods from one duration group (long-acting, only short-acting, were screened for eligibility (Figure 2). Of these, 21 033 met
only medium-acting or only-permanent) = 0, received coun- the inclusion criteria. Of those eligible, 1031 women (4.9%)
selling on methods from two or more durations groups = 1]. declined to participate. Among eligible women who agreed to
Outcome 4 was calculated as a binary variable (not receiving participate, 371 (1.9%) were missing data on primary study
counselling on at least one method from each contracep- variables and were dropped from analyses in the full analytic
tive attribute group = 0, receiving counselling on a method sample. A total of N = 19 631 women were included in the
from each attribute group = 1) (Senderowicz, 2020). A more full analytic sample. In the full analytic sample, 9305 (47.4%)
detailed explanation of the derivation of Outcomes 3 and 4 is reported that they did not receive any contraceptive coun-
presented in Appendix A. Finally, Outcome 5 was calculated selling at all during their antenatal and perinatal care and
as an ordinal count variable, which was the sum of the number were dropped for analyses to create the counselled analytic
of methods on which each woman received counselling. sample. The counselled analytic sample included N = 10 078
women. For Group 1, n = 5771 (n = 1874 at Dodoma Gen-
Analytic approach eral Hospital in Dodoma; n = 1220 at Muhimbili National
The outcomes of interest for this analysis are focused on the Hospital in Dar es Salaam and n = 2478 at Mbeya Zonal
content of the contraceptive counselling received. In addi- Referral Hospital in Mbeya), and for Group 2, n = 4555
tion to affecting the content of counselling, however, we (n = 3153 at Mt. Meru Hospital in Arusha and n = 1353
also expected the intervention to affect the proportion of at Tumbi-Pwani Regional Referral Hospital in Kibaha). In
women who receive counselling at all. As a result, except the counselled analytic sample, 5198 women (51.6%) were
where otherwise noted, we ran all analyses among two sam- exposed to the intervention, while the remaining 4880 con-
ples: (1) the sample of all respondents; and (2) the sub- stitute our control group. The trial was planned for 1 year
sample of respondents who reported receiving any coun- and was stopped when the expected duration was com-
selling. We used difference-in-difference linear probability plete. Sample size calculations were performed for primary
models to estimate the effect of the intervention on the out- study end-points but were not performed post hoc for the
comes of interest, controlling the time period as a fixed secondary outcomes employed in this analysis. Given that
effect and the hospital as a random effect. Coefficients for our analytic samples were quite large (full analytic sam-
binary outcomes (IUD only; counselled on >2 methods; coun- ple = 19 631 and counselled analytic sample= 10 078 obser-
selled on >2 methods durations; and counselled on a broad vations), we operated under the assumption that these sam-
mix of methods) can be interpreted as the percentage point ple sizes were pragmatic for the purposes of our secondary
increase or decrease in the probability of the outcome asso- analysis.
ciated with the intervention. Coefficients for the models for Table 1 describes the characteristics of study participants in
which the outcome was number of methods counselled can the full analytic sample. Examining only women who reported
be interpreted as the difference in the number of methods ever receiving any contraceptive counselling perinatally, the
counselled on associated with the intervention. We use an intervention and control groups remained similar in terms
intent-to-treat approach, classifying women who received any of their sociodemographic characteristics and were not per-
maternity services at hospitals where the intervention had fectly balanced. Those in the control group were, on average,
taken place as exposed, and women who received services 0.30 years younger, more likely to have at least a primary
when the hospitals had not yet received the intervention as education, less likely to be married/cohabitating, more likely
unexposed. to be Catholic, Muslim or Protestant and less likely to be
We present results for unadjusted and adjusted models. Evangelical Christian compared to their counterparts in the
Adjusted models controlled for sociodemographic character- intervention group. There were no statistically significant dif-
istics including women’s age, educational attainment, par- ferences between the respondents in Groups 1 and 2 (Pearson
ity, marital status, religion and ‘fast track’ hospital service et al., 2020).
Health Policy and Planning, 2023, Vol. 38, No. 1 43
Role of the funding source Table 2), a greater proportion of the control group received
The funder played no role in the study design; the collection, counselling on multiple contraceptive methods (94.8% vs
analysis or interpretation of data; the writing of the report; 79.0%), multiple methods durations (93.77% vs 77.34%)
or the decision to submit for publication. All authors had full and a broad method mix (59.39% vs 33.55%). Including
access to all the data in the study and accept the responsibility participants who did not receive any counselling at all, both
to submit for publication. groups received counselling on approximately the same aver-
age number of methods (2.09 in the control group, 2.07 in
the intervention group) and methods durations (1.36 vs 1.31)
(Table 1). However, among only participants who received
Results any contraceptive counselling, the control group received
Descriptive results counselling on a greater number of methods (4.29 meth-
Table 1 shows means and percentages for control variables ods vs 3.35 methods) but approximated the same number
and key outcomes between the intervention and control group of methods durations (2.8 durations vs 2.4 durations) (Sup-
in the full analytic sample. Some differences in education, plementary Appendix B Table 2). In the counselled group,
religion and hospital track were observed. Overall, a larger the two groups were counselled on approximately the same
proportion of the intervention group received any contracep- number of methods durations.
tive counselling (54.7% vs 49.0%). However, this appears to Figure 3 shows that a greater proportion of women
have been largely driven by counselling on IUD only (8.7% in the intervention group received counselling on the IUD
of participants who received any counselling in the interven- than in the control group (76.47% vs 61.07%; P < 0.0001),
tion group vs 0.9% in the control group). A larger proportion meeting the PPIUD Project’s aim of increasing counselling
of participants in the control group received counselling on on PPIUD (Figure 1). However, women exposed to the
multiple contraception methods (46.1% vs 42.8%), multi- PPIUD intervention had lower rates of counselling on every
ple contraceptive method durations (45.7% vs 41.9%) and other method, including oral contraceptive pills (67.95%
a broad method mix (28.9% vs 18.2%). Among participants in the intervention group vs 88.59% in the control group;
who received any counselling (Supplementary Appendix B P < 0.0001), condoms (33.28% vs 59.73%; P < 0.0001),
44 Health Policy and Planning, 2023, Vol. 38, No. 1
fertility awareness-based methods, (5.68% vs 30.02%; and unadjusted models, we observed a 16% point decrease in
P < 0.0001), injectables (74.03% vs 90.45%; P < 0.0001) and the probability of receiving counselling on multiple contracep-
implants (67.74% vs 88.59%; P < 0.0001). tive methods associated with the intervention (P < 0.001). The
intervention also decreased counselling on multiple methods
Difference-in-difference estimates of the effect of durations in this sample by 16% points in the unadjusted
the intervention on counselling model (P < 0.001) and 17% points in the adjusted model
Table 2 shows the results from a series of difference-in- (P = 0.01). For the measure of broad contraceptive method
difference regression analyses of the effect of the interven- mix, the size and the direction of the effect are consistent
tion on contraceptive counselling and method mix among with other results (around a 30% point reduction associ-
women who received contraceptive counselling, with wild ated with the intervention), however this relationship was
cluster bootstrapped P-values. Among women who received not statistically significant at alpha level 0.05 (P = 0.11 and
any counselling, those exposed to the intervention saw a sig- 0.13 in adjusted and unadjusted models, respectively). In the
nificant reduction in the number of methods on which they full sample, there were no differences between the control
were counselled. Women in the control group reported receiv- and intervention groups on receiving counselling on multi-
ing counselling on an average of 4.29 methods overall. In ple methods, multiple methods durations or a broad mix of
our unadjusted model, this number is reduced by 19% in the methods in unadjusted models. However, in adjusted mod-
intervention group, to 3.48 methods. In our adjusted model, els, in the full sample, the intervention was associated with a
women’s choice set is reduced by 26% in the intervention 16% point reduction in counselling on multiple methods, and
group, to 3.17 methods. a 17% point reduction on counselling on multiple methods
In the full sample, the intervention increased exclusive durations.
counselling on the IUD (and no other method) by 6% points Results from the mixed-effects models shown in Appendix
in the unadjusted model, and seven in the adjusted model. C yielded similar results, with, for example, women exposed
Among women who received any counselling, the interven- to the intervention having more than five times the odds
tion increased exclusive counselling on the IUD (and no other of receiving counselling exclusively on the IUD and no
method) by 12% points in the unadjusted model and 14 in other method compared to women in the control group
the adjusted model (P < 0.001). In this sample, in the adjusted (Supplementary Appendix C Table 3) among women who
Health Policy and Planning, 2023, Vol. 38, No. 1 45
reported receiving any counselling. In the full sample, women greater in the intervention group compared to the control.
exposed to the intervention having nearly five times the odds Women exposed to the intervention were also substantially
of receiving counselling exclusively on the IUD and no other less likely to receive counselling on multiple methods, multi-
method compared to women in the control group. ple method durations, or a broad contraceptive mix with a
variety of contraceptive attributes.
Strengths of this study include a large sample size and a
Discussion rigorous cluster-randomized stepped-wedge design as part of
Dedicated PPIUD programmes have repeatedly been demon- a broader, mixed-methods approach (Hussey and Hughes,
strated to increase PPIUD uptake, expanding access to this 2007). Our large and diverse sample across five geographic
highly effective method at a crucial period in the reproductive regions in Tanzania provides strong support for generalizabil-
life-course (Pleah et al., 2016; Karra et al., 2019; Pradhan ity to the postpartum population of Tanzania with facility-
et al., 2019; Pearson et al., 2020). In this study, we explore based births. Results were robust to multiple model spec-
the effects of a targeted PPIUD intervention on a broader set ifications and similar in adjusted and unadjusted models.
of person-centred outcomes related to family planning and The fact that qualitative in-depth interviews were conducted
health. We find strong evidence that intervention resulted in concurrently to the quantitative data collection allows us to
increased postpartum counselling on the IUD but reduced draw on the strengths of both qualitative and quantitative
counselling on all non-IUD related methods, and reduced methods. Here, we used qualitative results for hypothesis
the set of contraceptive methods from which women could generation, complementarity, expansion and triangulation of
choose. While the intervention increased contraceptive coun- findings (Greene et al., 1989).
selling overall, this increase was attributable nearly entirely to The study also had several limitations. The dataset was
counselling on IUD only, as indicated by analyses performed not originally intended to focus on person-centred outcome
among all women (including those who reported receiving no or contraceptive decision-making and thus does not compre-
counselling) showing no differences in the number of methods hensively measure all dimensions of contraceptive decision-
or methods durations counselled on. making, including measures of access. Recall bias is a concern
Our results indicate that the PPIUD intervention resulted when relying on retrospective reporting via survey methods,
in a stark reduction in access to information about a wide and here it may be possible that respondents may be more
method mix in contraceptive counselling. Women exposed likely to recall their counselling on the PPIUD than other
to the intervention experienced between an 18% and 26% methods. Data on the implementation of the intervention
reduction in the number of methods they were counselled is only available at the cluster level, leaving us unable to
on, and the proportion receiving counselling exclusively on confirm that individual respondents were exposed to interven-
the IUD (and no other methods) was more than nine times tion activities in their respective satellite clinics. There were
46 Health Policy and Planning, 2023, Vol. 38, No. 1
P-valueb
<0.001
0.002
0.007
0.134
0.072
on certain demographic variables; however, randomization
ensures exchangeability of treatment and control groups, and
any differences are therefore attributable to random chance.
Counselled sample
Given that randomization was performed by computer soft-
(−0.27, −0.06)
(−0.29, −0.05)
(0.06, 0.22)
(−0.70, 0.09)
(−2.34, 0.10)
ware at the cluster level, is extremely unlikely to have been
altered, either systematically or otherwise, by these charac-
95% CI
teristics. Finally, while the small number of clusters (5) may
have distorted estimates of precision, we were able to produce
robust standard errors using cluster wild bootstrapping.
0.14
−0.16
−0.17
−0.30
−1.12
The PPIUD intervention was designed with the goal of
Coef
Adjusteda
2020). Our analysis suggests that despite this success, the pro-
gramme did not increase women’s access to comprehensive
Full sample
who were counselled were less were less likely to have been
Adjusted models controlled for maternal marital/cohabiting status, parity, educational attainment, religion, hospital track and age as fixed effects.
(−0.69, 0.07)
(−1.96, 0.34)
(0.05, 0.20)
PPIUD introduction.
Findings from this analysis corroborate the findings from
95% CI
<0.001
0.999
0.947
0.591
0.702
Taken with the present findings, these data suggest that this
intervention led to a reduction in the breadth and quality of
contraceptive counselling.
Full sample
≥2 methods
autonomy, rather than on statistics that may drive providers and three of us are from the United States. Two of us are early
to encourage use of any given method. career scholars, two of us are mid-career scholars, and two of
Quantitative measurement, like any other form of knowl- us are senior scholars in leadership positions at our respective
edge production, is socially mediated and constructed, and as institutions. Our mix of training and research skills includes
such, can hold and promote the tacit ideologies and biases epidemiology, biostatistics, clinical medicine, gender analysis,
of those who create and deploy it (Merry, 2011). The quan- and overall experience in global health.
titative metrics used for programme evaluation often serve
as goal posts and benchmarks for success, especially when
delivered to providers in real time. Given this important feed- Meeting presentations
back loop, it is critical that measurement and evaluation take Preliminary versions of this work have been presented to
a holistic view of health and rights, and a narrow focus on the Population Association of America and to the African
any given outcome should be avoided. It is as crucial to mea- Population Conference.
sure and monitor the potential unintended consequences of
our programmes as it is to measure and monitor the ones we
intend. Clinical trial registration
This trial was registered at clinicaltrials.gov: NCT02718222.
Supplementary data
Supplementary data is available at Health Policy and Planning Trial registration
online.
The trial is registered with clinicaltrials.gov NCT02718222.
Data availability Ethical approval. The study received ethical approval from
Due to the terms of our ethics approval, the individual par- the National Institute of Medical Research (NIMR) in Tan-
ticipant data used in this analysis are not currently publicly zania (protocol number: NIMR/HQ/R.8a/Vol.IX/2006). The
available. study received a human subjects exemption from the institu-
tional review board at Harvard University (protocol number
IRB15–1605)
Funding
This study was funded by an anonymous donor. LS’ contri- Conflict of interest. The authors declare that they have no
bution was supported by the Ruth L. Kirschstein National conflicts of interest to disclose.
Research Service Award (T32HD049302) and a Population
Research Infrastructure grant (P2C HD047873). NAS’s con-
tribution was supported by NIAAA T32 AA007459. The References
content is solely the responsibility of the authors and does not Barot S, Cohen S, Darroch J et al. 2015. Sexual and reproductive
necessarily represent the official views of the Eunice Kennedy health and rights indicators for the SDGs. https://www.guttmacher.
National Institute of Child Health and Human Development, org/pubs/SRHR-Indicators-Post-2015-Recommendations.pdf.
the National Institute on Alcohol Abuse and Alcoholism, or Bashford A, Levine P. 2010. The Oxford Handbook of the History of
Eugenics. Oxford: Oxford University Press.
the National Institutes of Health. Funding bodies played no
Britton LE, Williams CR, Onyango D et al. 2021. “When it comes to
role in the study design, data analysis or preparation of this time of removal, nothing is straightforward”: a qualitative study
manuscript for publication. of experiences with barriers to removal of long-acting reversible
contraception in Western Kenya. Contraception: X 3: 100063.
Brown W, Druce N, Bunting J et al. 2014. Developing the “120 by 20”
Acknowledgements goal for the Global FP2020 Initiative. Studies in Family Planning 45:
The authors are grateful to Russell Diamond at the Social Sci- 73–84.
ence Computing Collaborative at the University of Wisconsin- Cahill N, Sonneveldt E, Stover J et al. 2018. Modern contraceptive use,
Madison for his support of this work. The authors are also unmet need, and demand satisfied among women of reproductive
grateful to our colleagues at FIGO and AGOTA for their age who are married or in a union in the focus countries of the Fam-
support and collegiality throughout this evaluation. ily Planning 2020 initiative: a systematic analysis using the Family
Planning Estimation Tool. The Lancet 391: 870–82.
Cairncross S, Periès H, Cutts F. 1997. Vertical health programmes. The
Author’s contributions Lancet 349: S20–1.
Canning D, Shah IH, Pearson E et al. 2016. Institutionalizing postpar-
L.S., E.P., J.F. and T.B. contributed to the conception and tum intrauterine device (IUD) services in Sri Lanka, Tanzania, and
design of the study. L.S., E.P., J.F. and N.U. contributed to Nepal: study protocol for a cluster-randomized stepped-wedge trial..
project administration and the collection/acquisition of data. BMC Pregnancy and Childbirth 16: 1–11.
L.S. and N.S. contributed to the data analysis, data inter- Christopherson S. 2016. “NWHN-SisterSong Joint Statement of Princi-
pretation, and manuscript writing. All authors reviewed and ples on LARCs.” National Women’s Health Network. https://www.
commented on the manuscript and approved the final version nwhn.org/nwhn-joins-statement-principles-larcs/.
Cleland J, Conde-Agudelo A, Peterson H et al. 2012. Contraception and
health. The Lancet 380: 149–56.
Reflexivity statement Cleland J, Shah IH, Benova L, Fresh A. 2015. Look at the level of
unmet need for family planning in the postpartum period, its causes
The authorship team for this study consists of three women and program implications. International Perspectives on Sexual and
and three men. Two of us are Tanzanian, one of us is German, Reproductive Health 41: 155–62.
48 Health Policy and Planning, 2023, Vol. 38, No. 1
Connelly MJ. 2008. Fatal Misconception: The Struggle to Control Pleah T, Hyjazi Y, Austin S et al. 2016. Increasing use of postpar-
World Population. Cambridge: Belknap Press of Harvard University tum family planning and the postpartum IUD: early experiences in
Press. West and Central Africa. Global Health: Science and Practice 4:
de Caestecker L, Banks L, Bell E et al. 2018. Planning and implemen- S140–52.
tation of a FIGO postpartum intrauterine device initiative in six Pradhan E, Canning D, Shah IH et al. 2019. Integrating postpartum
countries. International Journal of Gynecology & Obstetrics 143: contraceptive counseling and IUD insertion services into maternity
4–12. care in Nepal: results from stepped-wedge randomized controlled
Dehlendorf C, Henderson JT, Vittinghoff E et al. 2018. Development trial. Reproductive Health 16: 69.
of a patient-reported measure of the interpersonal quality of family PSI. 2015. Enabling the healthy spacing and limiting of pregnancies:
planning care. Contraception 97: 34–40. programmatic approaches to expand postpartum IUD access.
Festin MPR, Kiarie J, Solo J et al. 2016. Moving towards the goals of https://www.psi.org/publication/enabling-the-healthy-spacing-of-
FP2020—classifying contraceptives. Contraception 94: 289–94. pregnancy-programmatic-approaches-to-expand-postpartum-iud-
Greene JC, Caracelli VJ, Graham WF. 1989. Toward a conceptual frame- ppiud-access/.
work for mixed-method evaluation designs. Educational Evaluation Secura GM, Allsworth JE, Madden T et al. 2010. The contraceptive
and Policy Analysis 11: 255–74. CHOICE project: reducing barriers to long-acting reversible con-
Hackett K, Huber-Krum S, Francis JM et al. 2020. Evaluating the imple- traception. American Journal of Obstetrics and Gynecology 203:
mentation of an intervention to improve postpartum contraception 115.e1–7.
in Tanzania: a qualitative study of provider and client perspectives. Senderowicz L. 2019. “I was obligated to accept”: a qualitative explo-
Global Health: Science and Practice 8: 270–89. ration of contraceptive coercion. Social Science & Medicine 239:
Higgins JA, Kramer RD, Ryder KM. 2016. Provider bias in Long-Acting 112531.
Reversible Contraception (LARC) promotion and removal: percep- Senderowicz L. 2020. Contraceptive autonomy: conceptions and mea-
tions of young adult women. American Journal of Public Health surement of a novel family planning indicator. Studies in Family
106: 1932–7. Planning 51: 161–76.
Holt K, Zavala I, Quintero X et al. 2019. Development and validation of Senderowicz L, Kolenda A. 2022. “She told me no, that you cannot
the client-reported quality of contraceptive counseling scale to mea- change”: understanding provider refusal to remove contraceptive
sure quality and fulfillment of rights in family planning programs. implants. SSM—Qualitative Research in Health 2: 100154.
Studies in Family Planning 50: 137–58. Senderowicz L, Pearson E, Hackett K et al. 2021. “I haven’t heard much
Huber-Krum S, Hackett K, Senderowicz L et al. 2019. Women’s per- about other methods”: quality of care and person-centeredness in a
spectives on postpartum intrauterine devices in Tanzania. Studies in program to promote the postpartum intrauterine device in Tanzania.
Family Planning 50: 317–36. BMJ Global Health 6: e005775.
Hussey MA, Hughes JP. 2007. Design and analysis of stepped wedge Speizer IS, Corroon M, Calhoun L et al. 2014. Demand generation activ-
cluster randomized trials. Contemporary Clinical Trials 28: 182–91. ities and modern contraceptive use in urban areas of four countries:
Jayaraman S, Vermund SH. 2015. Trauma research—a field without a a longitudinal evaluation. Global Health: Science and Practice 2:
home base. Science Translational Medicine 7: 302ed11. 410–26.
Karra M, Pearson E, Pradhan E et al. 2019. The effect of a postpartum Sridhar A, Salcedo J. 2017. Optimizing maternal and neonatal outcomes
IUD intervention on counseling and choice: evidence from a cluster- with postpartum contraception: impact on breastfeeding and birth
randomized stepped-wedge trial in Sri Lanka. Trials 20: 407. spacing. Maternal Health, Neonatology and Perinatology 3: 1–10.
Keshavjee S, Farmer PE. 2012. Tuberculosis, drug resistance, and the Stanback J, Steiner M, Dorflinger L et al. 2015. WHO tiered-
history of modern medicine. New England Journal of Medicine 367: effectiveness counseling is rights-based family planning. Global
931–6. Health: Science and Practice 3: 352–7.
Lindegren Lou M, Kennedy CE, Bain-Brickley D et al. 2012. Integra- Sudhinaraset M, Afulani PA, Diamond-Smith N et al. 2018. Develop-
tion of HIV/AIDS services with maternal, neonatal and child health, ment of a person-centered family planning scale in India and Kenya.
nutrition, and family planning services. The Cochrane Database of Studies in Family Planning 49: 237–5.
Systematic Reviews 9: CD010119. Suh S. 2019. Metrics of survival: post-abortion care and reproductive
Makins A, Arulkumaran S. 2018. Institutionalization of postpartum rights in Senegal. Medical Anthropology 38: 152–66.
intrauterine devices. International Journal of Gynecology & Obstet- Tran NT, Seuc A, Coulibaly A et al. 2019. Post-partum family plan-
rics 143: 1–3. ning in Burkina Faso (Yam Daabo): a two group, multi-intervention,
Merry SE. 2011. Measuring the world: indicators, human rights, and single-blinded, cluster-randomised controlled trial. The Lancet
global governance. Current Anthropology 52: S83–95. Global Health 7: e1109–17.
Mogeni R, Mokua JA, Mwaliko E et al. 2019. Predictors of con- USAID. 2014. USAID family planning program timeline: before 1965
traceptive implant uptake in the immediate postpartum period: a to the present.
cross-sectional study. The European Journal of Contraception & van de Ruit C. 2019. Unintended consequences of community health
Reproductive Health Care 24: 438–43. worker programs in South Africa. Qualitative Health Research 29:
Morroni C, Glasier A. 2020. Increasing the use of effective postpartum 1535–48.
contraception: urgent and possible. The Lancet Global Health 8: World Health Organization. 2012. Programming Strategies for Postpar-
e316–7. tum Family Planning. Geneva.
Panter-Brick C, Eggerman M, Tomlinson M. 2014. How might global World Health Organization. 2021. WHO Guideline on Health Work-
health master deadly sins and strive for greater virtues?. Global force Development, Attraction, Recruitment and Retention in Rural
Health Action 7: 23411. and Remote Areas. Geneva: World Health Organization.
Pearson E, Senderowicz L, Pradhan E et al. 2020. Effect of a postpar- Yamin AE, Boulanger VM. 2013. Embedding sexual and reproductive
tum family planning intervention on postpartum intrauterine device health and rights in a transformational development framework:
counseling and choice: evidence from a cluster-randomized trial in lessons learned from the MDG targets and indicators. Reproductive
Tanzania. BMC Women’s Health 20: 1–3. Health Matters 21: 74–85.
Pfitzer A, Mackenzie D, Blanchard H et al. 2015. A facility birth can Yirgu R, Wood SN, Karp C et al. 2020. “You better use the safer one…
be the time to start family planning: postpartum intrauterine device leave this one”: the role of health providers in women’s pursuit of
experiences from six countries. International Journal of Gynecology their preferred family planning methods. BMC Women’s Health 20:
& Obstetrics 130: S54–61. 1–9.
Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.
Alternative Proxies: