Cohort Profile: The Performance Monitoring for Action (PMA) panel surveys

Key Features.

• The Performance Monitoring for Action (PMA) platform is an open cohort with embedded cross-sectional surveys that was established to capture contraceptive use dynamics behaviours and contraception supply-demand factors.

• PMA data are nationally representative in Kenya, Niger, Uganda, Burkina Faso and Côte d’Ivoire; and representative of subnational geographies in the Democratic Republic of Congo (Kongo Central and Kinshasa provinces), Nigeria (Kano and Lagos states) and India (Rajasthan state).

• PMA cohorts were enrolled in late 2019 (Phase 1), with two annual follow-up surveys conducted in each site (Phases 2 and 3). Phase 4 data are forthcoming in Burkina Faso, Côte d’Ivoire, Niger, Nigeria, Uganda and Kinshasa.

• Across the eight countries in PMA, a total of 40 986 women aged 15–49 years participated in Phase 1, 49 184 women participated in Phase 2 and 53 865 women in Phase 3.

• PMA female surveys measure key family planning and reproductive health indicators, along with gender- and other health-related measures. The Service Delivery Point surveys capture information on service quality and contraceptive availability and stock.

• To request access to PMA data and discuss potential collaboration, visit (https://www.pmadata.org/data/request-access-datasets) or email the lead investigator, Dr Philip Anglewicz, at (panglew1@jhu.edu).

Why was the cohort established?

The Performance Monitoring for Action (PMA) platform was established in 2013, with the initial goal of tracking progress towards the ‘120 × 20 goal’, or 120 million additional contraceptive users by the year 2020¹ through annual cross-sectional surveys. In 2019, PMA entered a new stage and broadened its goals to measure and explain contraceptive use dynamics, and gender-related influences on family planning (FP). Given the scarcity of large-scale representative panel studies on FP in low- and middle-income countries, PMA addresses a notable data gap in the field. It complements other large-scale, cross-sectional surveys such as the Demographic and Health Surveys (DHS) and Multiple Indicator Cluster Surveys (MICS) by (i) focusing on contraceptive dynamics, (ii) generating novel measures of FP and (iii) connecting supply- and demand-side features of the FP environment by linking population-based surveys to facility-based surveys. PMA’s panel approach is also well-suited for evaluating the effects of FP policies and programs, such as the introduction of new contraceptive methods, or implementation of interventions to increase access to FP services, and the impact of external factors, such as the COVID-19 pandemic.

PMA’s design was structured in two stages. The first stage, referred to as ‘PMA2020’, took place between 2013 and 2019 and consisted of repeated annual cross-sectional surveys among households, women of reproductive age and health facilities (referred to as ‘service delivery points’). Data were collected to track national and/or regional FP indicators in Burkina Faso, Côte d’Ivoire, Ethiopia, Ghana, Niger, Nigeria, Democratic Republic of Congo (DRC), Kenya, Uganda, India and Indonesia.²

The second stage is the focus of this manuscript. It started in 2019 when PMA shifted its study design to a longitudinal panel with embedded cross-sections to (i) more accurately track changes in FP indicators over time, (ii) explore contraceptive dynamics and associated factors (including gender-related measures) and (iii) study supply and demand factors that underlie contraceptive service utilization. This stage of PMA has been carried out at the national level in Kenya, Uganda, Burkina Faso, Côte d’Ivoire and Niger, and at subnational levels in Nigeria (Kano and Lagos states), DRC (Kinshasa and Kongo Central provinces) and India (Rajasthan state). In all geographies, PMA has conducted a baseline and two follow-up surveys, referred to as Phase 1, Phase 2 and Phase 3 surveys, respectively. A subset of geographies will have a fourth phase of data collection, including Burkina Faso, Niger, Côte d’Ivoire, Lagos, Kano, Kinshasa and Uganda.

Who is in the cohort?

PMA uses a multi-stage stratified cluster sampling approach to select the household and female samples. The sample size determination was based on estimating the modern contraceptive prevalence rate among all women aged 15–49 years old with a minimum 3% margin of error at the national level, and 5% at the subnational level (where applicable), assuming a response rate of 95% at the household and 95% at the female levels.

Sampling started with a selection of primary sample units called enumeration areas defined by the most recent census in each country. PMA chose enumeration areas with the probability of selection proportionate to population size after stratification by region, county, and/or area of residence (rural/urban). Within each selected enumeration area, PMA field staff conducted a mapping and listing of households and health facilities, after which 35 households were randomly selected for interview (40 households were selected in Lagos to compensate for lower response rate). In the household survey, local interviewers enumerated all occupants in the sampled dwellings, and all women ages 15–49 years who were usual members of the household or slept in the household the night before were selected for the female survey. At the end of the female interview, women were consented and enrolled in the panel. Before each female interview, women were asked to provide either verbal or written consent, depending on local IRB procedures (Supplementary Table S2, available as Supplementary data at IJE online). Consent was obtained from the woman at each phase, regardless of whether the woman consented to be enrolled in the panel and re-contacted at the previous phase.

Service Delivery Points surveys were administered after the female survey. The primary, secondary and tertiary public facilities serving the selected enumeration area were automatically included in the Service Delivery Point panel sample, and up to three randomly selected private Service Delivery Points within the enumeration area were selected. Client Exit Interviews were also conducted with clients of sampled Service Delivery Points, but are not the focus of this cohort profile (Supplementary Text S1, available as Supplementary data at IJE online).

Mode of interview

Surveys were conducted on smartphones using an ODK (open data kit) app. Most large-scale surveys are conducted by ‘outsider’ interviewers who are not residents of the communities surveyed. One of the innovations of PMA is that data collection model is conducted by women of reproductive age from within or nearby surveyed communities, who serve as ‘Resident Enumerators’, an approach that yields better data quality than outsider interviewers.^3,4

Response rate

Across the eight countries, 38 585 households completed the Phase 1 household survey (average response rate: 98.0%) and 40 333 women completed the Phase 1 female survey (average response rate: 97.2%) (Table 1). Table 2 provides descriptive statistics for the PMA female panel sample at Phase 1 across geographies. The response rates for cross-sectional survey varied over time but generally remained above 93%. Comparisons of the PMA Phase 1 samples and DHS show that the PMA samples closely match the main sociodemographic and FP characteristics of other nationally representative surveys (e.g. DHS and MICS); and PMA samples remained stable across Phases (Supplementary Text S2, available as Supplementary data at IJE online). Other studies that compared PMA and DHS also found similar characteristics between the two surveys.^5,6

Table 1.

Performance Monitoring for Action household and female response rates among cross-section dwelling units, 2019–23

Geography phase a	Households selected	Households occupied	Households interviewed	Household response rate	Total eligible women b	Eligible women interviewed	Women response rate	Consented to follow-up	Women participation rate
Burkina Faso Phase 1	5844	5764	5695	98.8%	6881	6590	95.8%	6418	93.3%
Burkina Faso Phase 2	5881	5647	5530	97.9%	6837	6388	93.6%	6257	91.5%
Burkina Faso Phase 3	5848	5536	5377	97.1%	6530	6078	93.1%	5957	91.2%
Côte d’Ivoire Phase 1	4274	4154	3988	96.0%	4263	4135	97.0%	3989	93.6%
Côte d’Ivoire Phase 2	4311	4001	3830	95.7%	4112	3949	96.0%	3804	92.5%
Côte d’Ivoire Phase 3	4309	3930	3774	96.0%	4065	3873	95.3%	3643	89.6%
DRC, Kinshasa Phase 1	2030	2009	1968	98.0%	2740	2611	95.3%	2553	93.2%
DRC, Kongo Central Phase 1	1995	1986	1965	98.9%	1975	1950	98.7%	1900	96.2%
DRC Phase 1	4025	3995	3933	98.4%	4715	4561	96.8%	4453	94.4%
DRC, Kinshasa Phase 2	2072	1962	1866	95.1%	2547	2369	93.0%	2288	89.8%
DRC, Kongo Central Phase 2	2026	1918	1891	98.6%	1962	1929	98.3%	1713	87.3%
DRC Phase 2	4098	3880	3757	96.8%	4509	4298	95.3%	4001	88.7%
DRC, Kinshasa Phase 3	2072	1921	1828	95.2%	2475	2326	94.0%	2245	90.7%
DRC, Kongo Central Phase 3	2054	1887	1861	98.6%	1896	1856	97.9%	1664	87.8%
DRC Phase 3	4126	3808	3689	96.9%	4371	4182	95.7%	3909	89.4%
India, Rajasthan Phase 1	4690	4633	4577	98.8%	5513	5408	98.1%	5058	91.7%
India, Rajasthan Phase 2	4690	4501	4421	98.2%	5544	5428	97.9%	5217	94.1%
India, Rajasthan Phase 3	4690	4459	4364	97.9%	5620	5481	97.5%	5371	95.6%
Kenya Phase 1	10 780	10 581	10 378	98.1%	9605	9477	98.7%	8825	91.9%
Kenya Phase 2	10 803	10 068	9727	96.6%	9455	9323	98.6%	9201	97.3%
Kenya Phase 3	10 867	9966	9565	96.0%	9629	9489	98.5%	9351	97.1%
Niger Phase 1	3601	3557	3515	98.8%	3809	3633	95.4%	3360	88.2%
Niger Phase 2	3646	3466	3428	98.9%	3837	3696	96.3%	3365	87.7%
Niger Phase 3	3596	3374	3334	98.8%	3909	3794	97.1%	3411	87.3%
Nigeria, Lagos Phase 1	1822	1738	1619	93.2%	1521	1469	96.6%	1394	91.7%
Nigeria, Kano Phase 1	875	864	857	99.2%	1128	1122	99.5%	1094	97.0%
Nigeria Phase 1	2697	2602	2476	95.2%	2649	2591	97.8%	2488	93.9%
Nigeria, Lagos Phase 2	1873	1712	1598	93.3%	1553	1483	95.5%	1447	93.2%
Nigeria, Kano Phase 2	879	858	854	99.5%	1142	1136	99.5%	1038	90.9%
Nigeria Phase 2	2752	2570	2452	95.4%	2695	2619	97.2%	2485	92.2%
Nigeria, Lagos Phase 3	1843	1703	1588	93.2%	1498	1426	95.2%	1372	91.6%
Nigeria, Kano Phase 3	875	838	834	99.5%	1156	1144	99.0%	1112	96.2%
Nigeria Phase 3	2718	2541	2422	95.3%	2654	2570	96.8%	2484	93.6%
Uganda Phase 1	4269	4148	4023	97.0%	4069	3938	96.8%	3820	93.9%
Uganda Phase 2	4939	4566	4399	96.3%	4537	4346	95.8%	4202	92.6%
Uganda Phase 3	5090	4597	4430	96.4%	4386	4227	96.4%	4041	92.1%
All Countries and Phases	122 544	116 344	113 084	97.2%	124 194	120 074	96.7%	115 110	92.7%

DRC, Democratic Republic of Congo.

^aPhase 1 dates: 2019–20; Phase 2 dates: 2020–21; Phase 3 dates: 2021–23.

^bEligible women interviewed is defined as the number of women with completed household and female questionnaires, and who slept at the house the night before the survey.

Table 2.

Descriptive data of the Performance Monitoring for Action female panel sample at Phase 1, 2019–20

	Burkina Faso	Cote D’Ivoire	Democratic Republic of Congo		Kenya	Niger	Nigeria		Rajasthan	Uganda
			Kinshasa	Kongo Central			Lagos	Kano
N	6532	4074	2549	1902	8797	3376	1399	1085	5071	3843
Age, years
15–19	21.3	21.1	22.4	21.0	21.6	19.4	13.9	24.8	17.5	24.3
20–24	17.5	18.3	21.4	16.5	17.9	20.9	14.0	16.4	18.1	20.0
25–29	15.1	16.0	16.5	16.3	16.6	19.9	16.1	18.0	17.6	17.3
30–34	15.9	16.3	12.5	14.9	15.8	15.4	18.2	14.8	14.9	13.9
35–39	13.0	12.9	10.5	13.8	11.8	11.1	18.3	11.9	13.0	11.0
40–44	11.2	10.4	10.5	10.4	9.6	9.5	13.1	8.0	11.5	8.0
45–49	6.1	5.0	6.2	7.1	6.7	3.9	7.0	6.1	7.5	5.6
Education
None/primary	77.1	66.9	7.7	40.0	50.0	84.4	11.6	65.6	50.6	61.7
Secondary	21.2	27.4	72.9	56.8	38.0	13.4	50.6	29.6	31.9	30.2
Tertiary	1.7	5.7	19.4	3.2	12.0	2.2	37.8	4.8	17.4	8.0
Marital status
Not married	20.7	32.5	50.8	25.6	31.7	12.3	32.2	20.1	22.8	27.2
In union	75.9	62.6	42.4	64.1	59.8	84.1	61.8	76.1	75.1	59.6
Divorced/widowed	3.4	4.9	6.9	10.3	8.5	3.6	6.0	3.8	2.2	13.3
Wealth
Tertile 1	33.8	29.4	30.3	29.4	36.1	31.7	28.8	29.1	27.6	30.9
Tertile 2	32.9	31.3	33.7	32.5	35.4	31.9	32.4	36.4	34.4	33.3
Tertile 3	33.3	39.4	36.1	38.1	28.5	36.4	38.8	34.6	38.0	35.8
Residence, urban	23.0	60.7			28.7	18.9		35.5	24.1	28.8
Contraceptive use and demand among all women
CPR	27.9	29.3	44.0	36.6	46.5	12.4	39.1	9.5	50.8	35.5
mCPR	26.0	23.2	24.6	23.3	44.1	11.1	26.1	8.2	44.9	29.9
Unmet need for FP	21.1	20.4	10.8	23.7	12.3	18.6	10.7	21.6	7.6	17.5

CPR, contraceptive prevalence; FP, family planning; mCPR: modern contraceptive prevalence.

The panel sample includes all age-eligible women who completed or partially completed the female questionnaire at Phase 1 and consented to follow-up. Percentages are weighted proportions.

How often have they been followed up?

Physical structured households form the basis of the panel, and all eligible women within selected dwelling units were followed over time. When an initially sampled dwelling was vacant or demolished, a new dwelling was randomly selected from the new listing to replace the lost one. At the annual follow-ups (Phase 2, 2020–21; and Phase 3, 2021–23), the interviewers returned to the enumeration areas and dwellings selected at Phase 1 (2019–20) to interview all women 15–49 years. This includes women who participated in the Phase 1 survey and were still eligible at follow-up, adolescents who turned 15 and new female residents aged 15–49 years who had moved into the selected households since Phase 1. Women aged out of the sample when reaching 50 years. In addition, enrolled women were interviewed if they resided in the study area, even if they moved to a different housing structure/unit, in line with the follow-up protocol (Supplementary Text S3, available as Supplementary data at IJE online). Service Delivery Point baseline data were collected at Phase 1 and follow-up data were collected annually at the same facilities, with replacement facilities selected for those that closed over time. PMA surveys timelines are available in the supplementary materials (Supplementary Figure S1, available as Supplementary data at IJE online).

Attrition

Of the 40 986 respondents who completed or partially completed the Phase 1 female questionnaire, 31 714 were followed up at Phase 2—attrition ranged from 8% in Kano, Nigeria to 23% in Côte d’Ivoire—and 12 799 new participants were enrolled in Phase 2. Of the 49 184 who completed or partially completed the Phase 2 female questionnaire, 35 857 completed the Phase 3 panel questionnaire—attrition ranged from 11% in Kano, Nigeria to 28% in Uganda—and 10 227 new participants were enrolled in Phase 3 (Table 3). In most countries, both at Phase 2 and Phase 3, women lost to follow-up were younger, more educated, more often living in urban areas, with a partner, and had a lower parity (Supplementary Tables S8 and S9, available as Supplementary data at IJE online).

Table 3.

Performance Monitoring for Action sample flow, 2019–23

	Burkina Faso	Cote D’Ivoire	Democratic Republic of Congo		Kenya	Niger	Nigeria		Rajasthan	Uganda
			Kinshasa	Kongo Central			Lagos	Kano
Phase 1, 2019–20
Full samplea	6790	4268	2639	1970	9558	3666	1507	1127	5469	3992
Phase 2, 2020–21
Ineligibleb	258	194	90	68	761	290	108	42	398	149
Permanently lost to follow-up	262	238	257	152	602	154	106	32	158	283
Temporarily lost to follow-up	779	719	286	216	1177	391	163	52	369	518
Attrition (%)	16	23	21	19	20	16	19	8	10	21
Sample additionsc	1764	1267	972	628	3134	1189	473	170	1082	2120
Full sample	8034	5104	3264	2378	11 329	4411	1766	1223	5995	5680
Phase 3, 2021–23
Ineligibleb	595	355	274	292	335	499	110	121	397	435
Permanently lost to follow-up	278	260	184	173	410	174	120	32	148	346
Temporarily lost to follow-up	1516	1069	442	384	1742	519	278	92	598	1148
Attrition (%)	24	28	21	27	20	18	24	11	13	28
Sample additions	1473	1076	866	614	2168	983	341	218	879	1609
Full samplec	8639	5565	3674	2527	12 752	4696	1877	1288	6339	6508

^aFull sample: number of women potentially eligible for next round. The full sample includes women who partially or fully completed the female survey, women who slept at the household the day before the survey and usual members of the household.

^bIneligible: women who aged out of the survey (>49 years) and those who did not consent to follow-up.

^cAll women who entered the survey at a given Phase, regardless of consent to follow-up.

COVID-19 subpanel

In addition to the main cohort survey, PMA conducted a phone follow-up survey 6 months after Phase 1 to collect COVID-19-related data to inform outbreak response interventions, policies and messages in Kenya, Burkina Faso, Nigeria and Kinshasa (DRC). All females from the PMA Phase 1 cohort who consented to and provided phone numbers for follow-up were eligible to participate and completed the survey (Supplementary Text S4, available as Supplementary data at IJE online).

What has been measured?

The household questionnaire contains a roster of household members and includes standard items on household assets, structure materials and water and sanitation that are used to create a wealth index. The female questionnaire contains a standard set of questions asked in all geographies, including sections on sociodemographic factors; reproductive history, including pregnancy, and fertility intentions; contraception; sexual activity; and women’s and girls’ economic and sexual and reproductive health (SRH) empowerment. A summary of the measures in the household and female surveys is listed in Table 4. Questionnaires were adapted at each phase based on prioritization of indicators or the performance of the survey items in the previous survey phases. Separate add-on modules, such as infertility, migration and gender-based violence were available at certain phases and included in the female survey. New questions on community-based gender norms and female genital mutilation (Nigeria and Burkina Faso only) were added at Phase 4. The Service Delivery Point survey measures items like contraceptive method availability and costs, provider training in FP services, FP client volume and FP service integration (Supplemental Text S5, available as Supplementary data at IJE online).

Table 4.

Performance Monitoring for Action household and female surveys content, 2019–23

Domain a	Measure/scale	Number of questions (range)	Survey cycle availability (C=Core; A = Add-on module/question)
			Phase 1	Phase 2	Phase 3
Household data (HQ)	Household roster information	10–12	C	C	C
	Household characteristics (assets and building materials)	7	C	C	C
	Water, sanitation and hygiene	2–5	C	C	C
Sociodemographic factors (FQ)	Age; overall health; education; marital/relationship status; current living situation	17	C	C	C
	Husband/partner’s level of education	1	na	C	C
	Relationship history	6	C	C	C
	Migration history	8	A	na	na
Reproduction, Pregnancy and Fertility Preferences (FQ)	Menstrual, pregnancy and birth history	11–12	C	C	C
	Fertility intentions	8–10	C	C	C
Contraception (FQ)	Knowledge of contraceptive methods	16–18	C	C	C
	Current contraceptive use	7–10	C	C	C
	DMPA-SCb self-injection	2–6	A	A	A
	Partner involvement in and support of contraceptive decisions	4–6	C	C	C
	Implant use	7	A	A	A
	Sterilization information	1	A	A	A
	Contraceptive source	5	C	C	C
	Perceived stigma related to obtaining contraception	2	C	C	C
	Non-users (future contraceptive intentions, reasons, decision-makers)	8	C	C	C
	Method information index plusc	4	C	C	C
	Provider pressure	2	na	C	C
	Contraceptive use history	4–6	C	C	C
	Family planning counselling, outreach, media and social norms	16–20	C	C	C
	Life goals among adolescents	7–14	C	C	C
	Health insurance	2	C	C	C
Sexual activity (FQ)	Sexual history	5–6	C	C	C
	Contraceptive use at first sex with current partner	2	C	C	C
	Contraceptive use at last sex	3	C	C	C
Women’s and girls’ empowerment (FQ)	Current employment and income	3	C	C	C
	Household decision-making related to household purchases and medical care	3	C	C	C
	Economic empowerment (household decision-making related to clothes purchases and earnings use; land ownership; earnings level compared with partner; income generating activities)	6	A	A	A
	Finance (savings accounts; level of financial literacy; financial goals)	2–5	A	A	A
	Internet access	4–5	na	na	C
	Personal beliefs on consequences family planning use	6	C	C	C
	Reproductive and sexual autonomy	10–20	A	A	A
Gender-based violence (FQ)	Sexual coercion	4	A	A	A
	Pregnancy coercion	5	na	C	C
	Intimate partner violence; household violence; help-seeking behaviour	15–19	na	A	A
Infertility (FQ)	Infertility (knowledge, experience, care)	9–14	na	na	A
COVID-19d (FQ)	Knowledge of; life impact; level of concern	15–18	C	na	na

FQ, female questionnaire; HQ, household questionnaire.

^aAdditional modules that were phase-specific and included in only certain country surveys are not included in this table. Add-on modules and questions included in this table have been offered to all Performance Monitoring for Action geographies during at least one phase of data collection.

^bDMPA-SC refers to self-injection of subcutaneous depo-medroxyprogesterone acetate.

^cMethod Information Index Plus is a score based on a set of four questions clients are asked at the end of a contraceptive visit. It gives insight into the type of information women receive during counselling and the extent to which women are informed about side effects and alternate methods.

^dAt Phase 1, COVID-19 questions were integrated into the FQ for Group B countries only (Uganda, India (Rajasthan), Niger, Côte d’Ivoire), as Group A countries (Kenya, Nigeria, Democratic Republic of Congo, Burkina Faso) had already finished collecting Phase 1 data by March 2020.

In addition to standard key indicators of FP and reproductive health, PMA survey has introduced innovative reproductive health measures, tracking knowledge of, attitudes towards and use of subcutaneous depot medroxyprogesterone acetate (DMPA-SC) (brand name Sayana Press), including self-injection. Implant use and removal questions have been added in countries with high implant use. Women and girls’ empowerment is studied across all PMA countries, with a focus on sexual, reproductive and economic empowerment.^7,8 PMA also developed an adolescent module in Rajasthan to explore girls’ SRH competencies, including autonomy in marriage and sexual decision-making; indicators were subsequently scaled to all PMA geographies in Phases 1–3. Given the potential for COVID-19 restrictions to increase gender-based violence, countries could opt into a gender-based violence mini-module beginning in 2020. The module includes questions on prevalence and intensity of intimate partner violence and household violence, as well as help-seeking behaviours.

What has it found?

With innovative measures and a longitudinal panel design, PMA has contributed to methodological advancements while providing new insights into key SRH topics like abortion, adolescent SRH, contraceptive dynamics, quality of FP care, women’s empowerment and reproductive coercion. Some of the more prominent findings are described below, and a full list of manuscripts published with PMA data can be found at the project’s Google Scholar site: https://scholar.google.com/citations?user=gYAacPsAAAAJ&hl=en&authuser=1

Methodological contributions

PMA has investigated the reliability of the contraceptive calendar, a widely used tool for key FP measures that has rarely been evaluated due to the lack of longitudinal panel data. The flexibility and frequency of PMA implementation allowed for a comprehensive assessment, revealing a ‘moderate to substantial’ level of reliability in nearly all PMA geographies, though less so for women using short-acting methods or women using contraception covertly from their partners.⁹

PMA also employed an anonymous third-party reporting technique (i.e. the ‘confidante’ method) in conjunction with innovative question framing to improve estimates of abortion incidence and safety in several countries and reporting of premarital sexual intercourse in Rajasthan.^10–15

Contributions to sexual and reproductive health and rights knowledge base

Using longitudinal data, PMA data have been used to examine the impact of the COVID-19 pandemic on FP and contraceptive use. Many predicted that COVID-19 would have a severe, detrimental impact on contraceptive use in sub-Saharan Africa. PMA provided the only population-based evidence in sub-Saharan Africa on the impact of COVID-19 on fertility intentions, contraceptive use and FP service provision early and after 1 year into the pandemic, looking at the impact on service delivery and on both population- and individual-level use of and demand for contraception. Contrary to the prevailing expectations, researchers using PMA data did not find a deleterious effect of COVID-19 in the earliest stages of the pandemic, extending to the first year into the pandemic, either at the individual- or population level; in fact, more women adopted than discontinued contraceptive use in Kenya and Burkina Faso.^16–19

PMA’s work on women and girls’ SRH empowerment, which involved mixed-method research in four geographies, underscores the dynamic nature of SRH empowerment, which works as a process starting with goal setting, followed by the ability to translate these goals into action and ultimately achieve SRH desired outcomes. The research also finds SRH empowerment is multidimensional as women may be empowered to make FP decisions but have no power in sexual decision-making. These results have informed the development of a cross-national SRH empowerment indicator, implemented in all PMA geographies, that can serve to track progress towards achieving the United Nations’ Sustainable Development Goal of gender equity and women’s empowerment, while also informing FP programs.^7,20,21

What are the main strengths and weaknesses?

PMA’s unique design fills several significant gaps in the field. Large-scale, nationally representative panel data in low-resource settings are rare in the FP/SRH field, but extremely valuable in permitting (1) more accurate measurement of change over time in reproductive and contraceptive outcomes for individual women and (2) sophisticated modelling of cause-and-effect relationships, contraceptive/reproductive trajectories and more. PMA also produces precise nationally and/or sub-nationally representative estimates of key FP/SRH indicators and do so more frequently than other large-scale population-based surveys. PMA has advanced measurement of FP/SRH and gender-related concepts by testing new approaches to measurement and assessing the reliability of standard measures. PMA includes both supply and demand side measures of the FP environment, by surveying both women and facilities that provide FP services in the same geographic context.²² Finally, PMA ensures that the data and results are tied to in-country policies and programs, which have been used extensively to develop and assess FP/SRH programs and policies in all geographies.

Some of PMA’s weaknesses relate to attrition between phases and implementation challenges during COVID-19 pandemic. First, although relatively low in all contexts, panel attrition is often selective, requiring post-stratification weights to reduce selection bias. Inverse probability weights partly address differential attrition, but do not account for differences in unmeasured characteristics. Second, COVID-19 delayed data collection by 6–9 months in several sites, including, Niger, Uganda, India, Côte d’Ivoire. Third, while collection of data from both demand (women) and supply (facilities) side is a rare feature for large-scale surveys, the PMA sampling of health facilities is not designed to be representative. Instead, the sampling is designed to capture both public and private facilities that serve the women in the PMA sample.

Can I get hold of the data? Where can I find out more?

Survey data are available for all countries at the PMA website (www.pmadata.org). Datasets are free to download and available to the public, but users are required to register and provide a description of the proposed research or analysis before data download (https://www.pmadata.org/data/request-access-datasets).

Ethics approval

PMA has been approved by the Institutional Review Board at the Johns Hopkins Bloomberg School of Public Health (IRB14702; MOD18860; MOD3903; MOD4748); the Comité d’Éthique pour la Recherche en Santé of the Ministère de l’Enseignement Superieur, de la Recherche Scientifique et de l’Innovation and the Ministère de la Santé, Burkina Faso (30-2019/CEIRES; 2020-11-258; 2021-11-253); the Comité National d’Ethique des Sciences de la Vie et de la Santé, Côte d’Ivoire (052-20/MHSP/CNESVS-km; 150-21/MHSP/CNESVS-km; 128-22/MSHPCMU/CNESVS-kp); the Comité d’Éthique de l’École de Santé Publique de l’Université de Kinshasa (ESP/CE/030B/2019; ESP/CE/160/2020; ESP/CE/159B/2021; ESP/CE/159C/2024); Indian Institute of Health Management Research University Institutional Committee for Ethics and Review of Research (IRB00008833); Kenyatta National Hospital—University of Nairobi Ethics and Research Committee (P801/09/2019); Comité national d’éthique pour la recherche en santé, Niger (034/2020/CNERS; 078/2021/CNERS); Health Research Ethics Committee of the Ministry of Health, Kano State (NHREC/1703/2018); Research Ethics Committee, Aminu Kano Teaching Hospital (NHREC/28/01/2020/AKTH/EC/3651), LASUTH Health Research and Ethics Committee, Lagos State University Teaching Hospital (LREC/06/10/1276; LREC/06/10/2215); Higher Degrees, Research and Ethics Committee, Makerere University, College of Health Sciences, School of Public Health (805).

Data availability

See ‘Can I get hold of the data?’ above.

Supplementary data

Supplementary data are available at IJE online.

Author contributions

S.R., Sa.A., Y.C., W.B., P.A., A.T., L.A.Z., E.G., C.M., C.K., S.W. and S.O.B. were responsible for the PMA study conception and design, or PMA sub-studies (e.g. abortion, adolescent, women’s and girls’ empowerment). G.G., R.M., F.R., P.A., A.K., P.G., M.T., S.A., S.O., E.O., F.M.O., M.S.Z., F.M. and S.P.S.K. supervised data collection in their respective geographies. A.S. and D.S. performed statistical analysis for this manuscript. A.S., M.E.B. and P.A. prepared the first draft of the manuscript. G.G., R.M., F.R., P.A., A.K., P.G., M.T., S.A., S.O., E.O., F.M.O., M.S.Z., F.M. and S.P.S.K. are Performance Monitoring and Action Principal Investigators and reviewed and assisted in finalizing the manuscript. All authors reviewed and approved the final manuscript.

Funding

This work was supported by the Bill & Melinda Gates Foundation (grant numbers 010481 and 130672). PMA has also received contributions from the Children’s Investment Fund Foundation (CIFF), the National Institutes of Health (NIH), the United Nations Population Fund (UNFPA), the Packard Foundation, the Hewlett Foundation, FHI 360, the Johns Hopkins Center for Communication Programs, a Large Anonymous Donor and Guttmacher. Overall, PMA has received more than $95 million in funding from various sources. A full list is provided in the Supplementary data file, available as Supplementary data at IJE online. Under the grant conditions of the Foundation, a Creative Commons Attribution 4.0 Generic License has already been assigned to the Author Accepted Manuscript version that might arise from this submission. The funders had no role in the study design, collection, analysis and interpretation of data, in writing of the report or in the decision to submit for publication.

Conflict of interest

None declared.

Use of artificial intelligence (AI) tools

No AI tools were used.