Emergency Contraception

The sample size for the study was calculated using Yamane’s formula [7 ] for proportions based on a population of 240 final year female nursing and midwifery students, under the assumption of 95% confidence interval, 5% margin of error, and 20% non-response rate. A sample size of 180 was estimated but was increased to 191 to broaden the scope of the study. The sample size was allocated to the nursing and midwifery programmes using proportion to size approach. Purposive sampling method was then used to sample 100 students from the nursing programme and 91 from the midwifery programme. The students were acquainted with the objective and purpose of the study and informed of their rights during the study. Data for the study were collected with a structured questionnaire on knowledge, perception, and use of emergency contraception. We also collected information on some socio-demographic characteristics (age, marital status, religion, and programme of study) of the students. The questionnaire was designed after considering variables that were included in similar studies. Two senior midwives and a public health professional reviewed the questionnaire for construct and content validity. Then the questionnaire was piloted on 30 first-year nursing and midwifery students to ascertain the clarity and practicability of the questions and to identify poorly constructed items and ambiguities that may be encountered during data collection. Suggested changes from the review and pilot study were made before the actual data collection. Three trained health tutors administered the questionnaire to the nursing and midwifery students in classrooms during school hours.

This study uses longitudinal data from women and facilities collected in 2011, prior to the start of ISSU program activities (baseline data), and 2015 (endline data) in all six cities. Longitudinal data for women allow us to use fixed effects methods that allow each woman to act as her own control [13] . At baseline, in each city, we collected data from a representative sample of women identified using a two-stage sampling design. In the first stage, we used the 2002 General Population and Housing Census's list of census districts (updated in 2009) as our primary sampling units (PSU) to select a random sample of PSU in each city. On average, census districts contain about 150–200 households. In total, we selected 268 PSU: 64 PSU from Dakar; 32 each from Guédiawaye, Pikine, and Mbao; and 54 each from Mbour and Kaolack. Prior to selection, we worked with municipal leaders to classify neighborhoods as poor and non-poor based on five characteristics: type of housing, residential security, neighborhood density, access to water, and access to flush toilets. Municipal leaders also gave an overall classification of the poverty level of the neighborhood. We weighted the five characteristics of neighborhoods (access to water had a weight of 2; toilets a weight of 1.5; and all others a weight of 1) and then summed. We considered those neighborhoods scoring in the lowest 40% as poor and all others as non-poor. Prior to selection, we stratified PSU by poor and non-poor and half of selected PSU were from the poor strata to increase inclusion of poor households and women. In the selected PSU, we listed/mapped all households and randomly selected 21 households for interview with equal probability [14] .
In total, we selected 5628 households. In each selected household, all women of reproductive age (15–49) were eligible for interview. A trained female interviewer approached all eligible women and asked for their signed consent to participate. At baseline, we interviewed 9614 women across the six cities to provide a representative sample for each city at baseline. At endline, we tracked all baseline women who were usual residents (not visitors) of the household (n=9421) and, if found, asked them for consent to be re-interviewed.
The Institutional Review Board at the University of North Carolina at Chapel Hill and the Comité National d'Ethique in Senegal approved all study procedures.
Women interviewed at both baseline and endline are the analysis sample for examining the impact of the program on modern contraceptive use over the four-year follow-up period. We also undertake a sub-sample analysis of “poor” women, identified from the matched sample as those women who were in the two lowest wealth quintiles (poorest and poor women) at both baseline and endline.
At baseline and endline, we also collected data from health facilities in the study cities. We approached all public and private facilities offering reproductive health services (n=269) for interview. At each study facility, an interviewer administered a facility audit, provider interviews (up to four per facility), and exit interviews [15] .
The key dependent variable for this analysis is use of modern contraception. At baseline and endline, trained interviewers asked women if they were using a contraceptive method to delay or avoid childbearing and if yes, interviewers asked what method the woman was using. Modern methods of contraception include male and female sterilization, daily pill, IUD, implants, injectables, male and female condoms, emergency contraception, Standard Days Method, and lactational amenorrhea. We coded women who reported traditional method use (e.g., rhythm method, withdrawal, or folkloric methods) as non-users. At each time point, we coded women who were abstinent or not sexually active as non-users.
This analysis examines the impact of exposure to various ISSU and non-ISSU demand generation activities and the ISSU and non-ISSU supply-side activities on the probability of modern method use. We classified the exposure variables as demand-side and supply-side. At endline, we included detailed questions on exposure to ISSU specific radio and television activities to assess the contribution of ISSU programming in target cities above and beyond the national-level radio and television programming (see Appendix Table A for details of questions used to measure exposure). We coded each of the exposure variables as dichotomous variables categorized as exposed versus unexposed. All exposure measures that were ISSU-specific and did not exist at baseline were coded zero (unexposed) at baseline.
We included four supply-side variables in this analysis. One of these is specific to the ISSU program: exposure to the ISSU Informed Push Model (IPM). To link the IPM variable to the women's data, we created a variable capturing whether the woman lived within 1 km of a facility with IPM at the time of the endline survey. This variable is coded zero at baseline since the IPM program activity did not exist at that time. We created three other supply variables based on this same 1 km buffer around where women live. These were: (a) woman lives within 1 km of a facility with any stock outs in the last 30 days; (b) woman lives within 1 km of a facility with a quality improvement committee; and c) woman lives within 1 km of a facility with observed FP guidelines.
We undertake descriptive analyses as well as multivariate analyses. We apply fixed effects regression to the pooled samples (baseline and endline) to reduce bias associated with self-selection, recall, and program targeting to underserved areas due to time invariant unobservables. Our estimation methods control for the possibility of endogenous attrition in the endline sample to the extent that attrition is due to unobserved fixed characteristics of the individuals. Fixed effect methods do not control for time varying unobservables. However, given the relatively close spacing of observations, time-varying unobservables are likely not a major source of residual bias in this setting. We estimated a classic fixed effects model whereby changes in modern contraceptive use depend on changes in time-varying individual characteristics and program exposure. All models control for marital status, age, education, city, religion, and wealth (contact corresponding author for full models). We perform two models, one including all women and the other for those in the two lowest wealth quintiles at baseline and endline. All analyses adjust for correlation at the community level using Huber-White type sandwich estimators for standard errors.

Momentum was an integrated FP and maternal and newborn health project. The project sought to increase contraceptive uptake and the adoption of health-seeking behaviors and household practices beneficial to mother and baby, and foster gender-equitable attitudes and behaviors among FTMs age 15–24 years and their male partners in Kinshasa, DRC. The study design was quasi-experimental, with three intervention and three comparison health zones. FTMs were recruited through convenience sampling at high-volume maternal health facilities and community sites and were followed up for 16 months. Enrollment criteria were: (a) six-months pregnant with the first child; (b) willing and mentally competent to provide informed consent; (c) ability to speak French or Lingala; and (d) residence in the intervention or comparison health zones. A total of 1,927 FTMs were completely interviewed in both surveys, of whom 761 were currently using a modern contraceptive.
In intervention health zones, trained nursing students conducted monthly group education sessions and home visits that included client-centered counseling on postpartum FP and birth spacing, offered a range of contraceptive methods (Implanon NXT, Sayana® Press, progestin-only pills, combined oral contraceptive pills, male condoms, emergency contraception, and Cycle Beads), and provided referrals. Group education sessions were based on Program M (21 ), a curriculum that focuses on equitable gender roles, empowerment in interpersonal relationships, and sexual and reproductive health and rights. Each FTM was assigned a dedicated pair of trained nursing students (one male and one female) who conducted monthly home visits in the prenatal and postnatal periods.
Data were collected through face-to-face interviews conducted by trained data collectors at baseline (September to November 2019) and endline (May to July 2020). Pretested questionnaires were used to gather data on the following FP-related topics: background characteristics; knowledge of, attitudes toward, and use of contraceptive methods, perceived norms, informed choice, decision making about contraceptive use, method satisfaction, and exposure to Momentum interventions. Questionnaires were administered using Open Data Kit software and Android smartphones.