Provision of family planning vouchers and early initiation of postpartum contraceptive use among women living with HIV in southwestern Uganda: A randomized controlled trial

Background Unwanted pregnancies remain a burden for women living with HIV (WLWH). Family planning prevents unplanned pregnancies while promoting longer birth intervals, key strategies to eliminate perinatal transmission of HIV and promote maternal and child health. We evaluated the effect of a family planning voucher, inclusive of immediate postpartum counseling, on uptake, early initiation, and continuation of modern contraceptive methods among recently postpartum WLWH delivering at a publicly funded regional referral hospital in rural, southwestern Uganda. Methods and findings We performed a randomized controlled trial between October, 2016 and June, 2018 at a referral hospital in southwestern Uganda. This interim analysis includes adult WLWH randomized and enrolled equally to receive a family planning voucher or standard of care (con-trol). Enrolled postpartum WLWH completed an interviewer-administered questionnaire at enrollment and 6 months postpartum. Our primary outcome of interest for this analysis is initiation of a modern family planning method within 8 weeks postpartum. Secondary outcomes included family planning initiation at 12, 14, 16, and 20 weeks postpartum, family planning discontinuation and/or change, pregnancy incidence, and mean time without contraception. The trial was registered with clinicaltrials.gov (NCT02964169). At enrollment, half of the women in both the voucher (N = 87, 55%) and control (N = 86, 54%) groups wanted to have a child in 2 years postpartum. Over 80% of referent pregnancies in the voucher (N = 136, 86%) and control (N = 128, 81%) groups were planned. All women were accessing ART. The mean CD4 count was 396 cells/mm3 (SD = 61) for those enrolled in thecontrol group versus 393 cells/mm3 (SD = 64) in the family planning voucher group. By 8 weeks postpartum, family planning was initiated in 144 (91%) participants in the voucher group and 83 (52%) participants in the control group (odds ratio [OR] 9.42; CI 4.67–13.97, P < 0.001). We also found high family planning uptake rates for both groups, with higher rates among the intervention group at 12 weeks (OR 5.66; CI 2.65–12.12, P < 0.001), 14 weeks (OR 2.51; CI 1.31–4.79, P < 0.001), 16 weeks (OR 4.02; CI 1.66–9.77, P = 0.001), and 20 weeks (OR 3.65; CI 1.40–9.47, P = 0.004) postpartum. The average time to family planning initiation was reduced to 5.9 weeks (SD = 2.4) for those in the voucher group compared to 9.3 weeks (SD = 5) in the control (P < 0.001). One pregnancy was recorded in the group receiving standard of care; none were reported in the voucher group. Method mix did not differ by group: injectables were selected by most women (N = 150, 50%), and 52% of this proportion were in the experimental arm, with <10% in each arm selecting condoms, oral contraception, or intrauterine devices (IUDs). Similar proportions of women changed contraceptive methods over the 6-month follow-up in the voucher and control groups (N = 8, 5% versus N = 5, 4%; P = 0.467). More women in the control group discontinued contraception for 1 to 2 weeks (N = 19, 13% versus N = 7, 5%; P = 0.008) or more than 4 weeks (N = 15, 10% versus N = 3, 2%; P = 0.002) compared to those given a family planning voucher. The main limitation of this study is that its findings may not be generalized to settings without improved availability of contraceptives in publicly funded facilities. Conclusion These findings indicate that a well-structured, time-bound family planning voucher program appeared to increase early postpartum contraceptive uptake and continuation in a setting in which users are faced with financial, knowledge, and structural barriers to contraceptive services. Further work should clarify the role of vouchers in empowering WLWH to avoid unintended pregnancies over time. This analysis includes preliminary data collected from WLWH enrolling in a randomized controlled trial at the maternity ward of a regional referral hospital in southwestern Uganda. Mbarara Regional Referral Hospital (MRRH) is a publicly funded teaching hospital serving 10 districts with a population of over 5 million people. The parent trial aims to evaluate the effect of family planning support versus standard of care on contraceptive use at 12 months postpartum ({"type":"clinical-trial","attrs":{"text":"NCT02964169","term_id":"NCT02964169"}}NCT02964169). Here, we present interim data up to 6 months postpartum. The hospital is equipped with trained staff, midwives, and obstetricians able to offer comprehensive family planning services. Women accessing care at this hospital represent varied social and demographic backgrounds. The hospital performs over 12,000 deliveries annually and reports a 13% HIV prevalence among women (hospital records). This study was initiated in October, 2016, and enrollment ended in May, 2017. Eligible participants were WLWH ≥ 18 years of age admitted to a postnatal ward within 5 days postpartum regardless of pregnancy outcome and qualified for any family planning methods available. The exclusion criteria included 1) HIV negative, 2) history of hypersensitivity to latex, 3) no male sexual partner and/or not anticipating one for the next 2 years, 4) only sexual partner has had vasectomy, 5) resides and works more than 20 km from the study site, and 6) inability to complete informed consent process as assessed by the study nurses. Trained research assistants (RAs) approached WLWH in the postnatal ward within 12 hours after delivery to capture all women delivering at this facility. RAs obtained voluntary written informed consent from all eligible participants in the local language in a private area of the hospital. All consenting participants gave written informed consent or, for those who could not write, a thumbprint was made on the consent form. A primary partner was defined either as a regular spouse who is also a regular sexual partner or the most recent sexual partner if no main partner was named. Because of structural and capacity challenges at the referent hospital site, routine discharge is often completed without family planning counseling. 1. Family planning voucher intervention: following delivery, the women randomized to the intervention group were counseled and given a family planning voucher by the study nurse. Our one-on-one educational counseling was semistructured, providing face-to-face standardized (a list of items to talk about was generated) information to the woman and available partner on the available contraceptive methods, family size, medical eligibility for the different contraceptive methods, dual contraception, when to start contraception, how to use the contraceptives, potential side effects and benefits/effectiveness, and where the different family planning methods can be accessed (Fig 1). Structured, immediate postpartum counseling was offered in a clinic setting in a private room by a well-trained study nurse and lasted up to 40 minutes (S1 Text). All women were given opportunity to ask questions to facilitate women’s informed choice of any of the 5 freely available family planning methods at MRRH (both long- and short-acting modern contraceptive methods, including condoms, injectables, contraceptive pills, copper IUDs, and contraceptive implants). WLWH were also counseled on the standard days method (SDM) and lactational amenorrhea method (LAM). The same voucher and counseling was also given to the male sexual partner, when available, because of its identified effect on family planning utilization [15]. The voucher was used as an incentive to motivate women to seek/demand and access family planning easily at family planning clinics. A well-trained nurse was available at the postnatal clinic to identify women with vouchers to access the relevant service provider within 1 hour of arrival. Waiting time to see a service provider at the clinic was reduced to 1 hour maximum. DOA, Date of Admission; DOD, Date of Discharge; FP, family planning; IP, In-patient; IUD, intrauterine device; PID, Pelvic inflammatory disease; UTI, Urinary Tract Infection. Although family planning is freely available in public health facilities, stockouts, especially of the long-acting contraceptive methods (implants and IUDs), attributed mainly to supply chain challenges are common [12]. The study promoted minimal stockouts of all methods at MRRH during the study period through regular involvement in forecasting and ordering. Private facilities rarely experience contraceptive method stockouts [12], and thus the family planning voucher also offered an opportunity for free administration (for example, injection, IUD placement, implant) of a contraceptive method purchased outside of the public healthcare facility. For this study, women who reside and/or work within 20 km from MRRH were enrolled, and thus all women were in close proximity to a facility with family planning services. The voucher was offered for free and had an expiration of 3 months from the date of delivery. The voucher included detailed information about side effects for the different contraceptive methods as well as a general overview on benefits of family planning (Fig 1). Within 3 months postpartum, the women were expected to have returned to a health facility for at least two of their scheduled routine postnatal visits and/or immunization appointments. 2. Control group: in order to align the control group with guidelines-based standard of care, women in the control group were offered routine family planning counseling at discharge as defined by the Uganda clinical guidelines [16] by a well-trained ward nurse and documented in the Ministry of Health discharge form (Fig 1). The control group was not given a voucher. Women from both groups were invited to start any available family planning method prior to discharge. The choice and place of family planning was entirely up to the participants regardless of group. All women accessing services at the hospital family planning clinic received care by a well-trained dedicated nurse to counsel and administer a chosen family planning method. Permission to contact spouses/sexual partners was obtained from all enrolled women. If permission to contact was given, a spouse/sexual partner was contacted, enrolled, and interviewed at baseline and 6 months for both groups. The spouses for controls were not given vouchers (S2 Text). A study biostatistician generated a randomization list with a block size of 20, totaling 160 participants equally in each of the two groups into which mothers could be randomly assigned and enrolled. The aim of the study and details of the procedures to be involved in the trial, were explained before randomization. Once mothers consented to participate in the study, a study number was allocated by the research assistant by taking the next in a series of similar opaque envelopes provided to conceal allocation of groups. These opaque envelopes were labeled with a computer-generated list of numbers with group allocation. The RAs were blinded to the group allocation until eligibility and study participation was confirmed. They were also blinded to the hypothesis of the study. WLWH were screened for eligibility and enrolled equally into the intervention arm (family planning support) and standard of care (control group) between October, 2016 and May, 2017. Women were followed for 1 year. All participants completed interviewer-administered interviews at baseline and 6 months postpartum. Interviews were conducted by two trained RAs fluent in English and the main local language in a private office space. Each interview took about 30–45 minutes. A different RA from the one enrolling participants was trained to specifically collect follow-up data to limit social desirability bias. Data were collected electronically. The data analyst was blinded to the group allocated to different study participants. A transport refund of US$3 was given on each visit. A blood sample was drawn at baseline to confirm HIV status and measure CD4 cell count. A structured face-to-face questionnaire was completed at enrollment to collect information on sociodemographics; depression and health[17]; reproductive history; partnership dynamics (for example, HIV serostatus disclosure, partner HIV serostatus); perception, use, and knowledge of contraception; decision-making [18–23]; food insecurity [24, 25]; alcohol use in the last year [26]; HIV stigma [27]; social support [28]; and pregnancy intentions or aspiration [29–31]. Our primary outcome of interest for this interim data was initiation of a modern family planning method within 8 weeks postpartum (the primary outcome for the full trial is effective contraceptive use at 12 months postpartum). Initiation of a modern family planning method at any facility of choice within 8 weeks postpartum was by both self-report and reviewing the participant’s postnatal chart by the study RA to identify and confirm initiation of family planning regardless of where the service was obtained. Outcomes from both reports were evaluated to confirm the internal validity and consistency of the two measures. Secondary outcomes included pregnancy incidence, mean time without contraception and family planning initiation at 12, 14, 16, and 20 weeks postpartum, and family planning discontinuation and/or change. Although postpartum counseling on contraceptive methods focused on the 5 methods—condoms, injectables, contraceptive pills (including progestin-only pills for breastfeeding mothers), copper IUDs, and contraceptive implants—as provided at MRRH, modern family planning was defined as use of these 5 and any other methods such as a diaphragm or cervical cap that participants could have obtained from other facilities. One-third of the pregnant WLWH in Uganda report an unmet need for contraception services and family planning support [32]. Provision of a family planning voucher has a significant impact on contraceptive uptake and long-term contraceptive use by an increase of 18 percentage points within 2 years of the reporting period among postpartum women [15]. The current contraceptive uptake among WLWH is 45% [1]. We therefore hypothesized that improved and focused family planning support through a voucher will increase contraceptive uptake among HIV-positive postpartum women to at least 63% within a follow-up period of 1 year postpartum. Allowing for a two-sided type I error of 5%, our target sample size was 320 postpartum women (with equal numbers of participants in the intervention and control groups) to enable 90% power to demonstrate a significant difference between groups. We described demographic and clinical data for the cohort using standard descriptive statistics. The Household Food Insecurity Access Scale (HFIAS) was calculated as recommended [33]. We compared dichotomous outcomes between study groups by estimating crude odds ratios (ORs) with 95% confidence intervals, and testing for differences between the two groups. We estimated P-values with chi-squared testing using a level of significance of 0.05. We compared continuous outcomes and estimated P-values using Student t tests. All primary and secondary outcomes were analyzed using intention-to-treat analyses (although no participants were misallocated a group [23]). Although our study was fully randomized, the differences in baseline characteristics noted between study groups was assessed for confounding by fitting multivariable logistic regression models. As per the revised CONSORT guidelines for reporting randomized trials [26], we assessed for subgroup effects for the following characteristics by testing the significance of interaction terms in a multivariable regression model: 1) children living in household below 18 years of age (dichotomized into <3 children and ≥3 children categories), 2) parity (dichotomized into 1–3 and >3), 3) prenatal visits (<3 and ≥3 visits), 4) household income (≤150,000 and >150,000 Ugandan shillings), 5) involvement in any domestic violence (involvement, no involvement), 6) religion (Catholic, Protestant, others), and 7) duration on ART (<4, 4–8, and >8 years). A Mantel–Haenszel test was also done to control for each of these variables. All statistical analyses were performed using STATA version 13.0 (Statacorp, College Station, TX, USA). All human subjects’ ethical approvals were obtained from Institutional Review Committees of Mbarara University of Science and Technology and Uganda National Council of Science and Technology and registered with clinicaltrials.gov ({“type”:”clinical-trial”,”attrs”:{“text”:”NCT02964169″,”term_id”:”NCT02964169″}}NCT02964169). A research assistant trained in human participant research conducted informed consent procedures with eligible participants in the local language in a private area. Written informed consent was obtained from all eligible participants.