Expanding contraceptive options for PMTCT clients: a mixed methods implementation study in Cape Town, South Africa

Working in consultation with the district health management team, the investigators designed a multi-faceted intervention that was informed by formative research previously completed by the investigators in South African PMTCT facilities. The earlier research revealed knowledge gaps on the part of health service providers and clients about use of long-acting and permanent contraceptive methods by women living with HIV[22]. The tested intervention focused on reinforcement of counseling on postpartum family planning delivered to PMTCT clients in antenatal, delivery, and child health services. Providers were trained to include the IUD and female sterilization among the methods promoted to clients, and clinical services for providing these methods were reinforced. Key intervention elements included the following (see Table 1, Column 1 for details):

While the study team provided technical support for intervention implementation and oversight, the aim was to introduce an intervention package with potential for broad implementation within public sector services. Accordingly, intervention elements were incorporated into routine service delivery and depended on the systems and providers already in place in the participating public sector health facilities.

Before intervention implementation and 6 months after it was fully in place, we interviewed a total of 1077 women exiting child health services (538 pre-intervention and 539 post-intervention), with approximately equal numbers of women at each time period who had and had not enrolled in PMTCT services during their most recent pregnancy. Eligible participants were adult women who reported having received antenatal care and were seeking child health services for their infants at the 5 participating facilities. Different samples of clients were interviewed at each time period, as opposed to following a cohort over the 9-month intervention period, given the additional eligibility criterion of participants being no more than 6 months post-partum. During the two data collection periods, each lasting about 2 months, study participants were recruited in the waiting areas of the facilities; all clients were screened for eligibility and administered informed consent until the sample size was attained. Research assistants administered a structured questionnaire in either English or Xhosa, according to the participant’s choice, to collect information on socio-demographic characteristics, contraceptive use, reproductive history, fertility intentions, exposure to family planning counseling, and knowledge about the IUD and sterilization. Interviews took approximately 40 minutes, and participants were not compensated for their time.

This analysis focused on the PMTCT clients. Using Stata 10.0, we computed summary statistics separately for the pre-intervention and post-intervention samples, and then analyzed the difference between samples. Differences between samples were analyzed using rank sum test for continuous variables and chi-squared and Fisher’s exact test for categorical variables. The study was powered to assess whether at least 30% of PMTCT clients interviewed post-intervention had correct knowledge about the IUD and female sterilization, respectively, as captured by the primary outcome variables. Power estimates were based on post-intervention estimates instead of a change over time since it was assumed that clients in the pre-intervention sample would have negligible knowledge about those methods that were largely unavailable prior to the study.

Approximately 8 months after intervention introduction, a research assistant conducted individual in-depth interviews with 16 providers who had participated in the study-sponsored trainings. Open-ended questions explored providers’ attitudes toward and experiences with promoting voluntary contraceptive services to HIV-positive clients, with particular focus on IUD and female sterilization services. Questions examined providers’ views on the safety and appropriateness of those methods for women living with HIV. Providers were further asked about any approaches they take in assessing clients’ fertility desires and tailoring family planning accordingly. Other questions explored factors facilitating and impeding providers’ ability to promote use of the IUD and female sterilization, considering health system resources, the supports introduced through the study intervention, and clients’ needs and priorities. Interviews, conducted in English, were digitally recorded and took approximately 60 minutes. Providers were not compensated for their time. One lead investigator and an analyst read the transcripts and developed a coding scheme shaped by the interview guide. An analyst coded and analyzed transcripts using QSR International’s NVivo 9. To complete thematic analysis, the analyst prepared memos to capture the divergence and convergence of ideas emerging from the text data. The lead investigator reviewed the memos and introduced modest refinements in consultation with the analyst.

The research team documented intervention implementation using a tracking tool. This document consisted of an Excel worksheet, similar to Table 1, with columns to record the intervention components, first as described in the research protocol and then as actually implemented. Investigators used a third column to record observations on challenges and considerations for future replication. Investigators filled in the tracking tool on an ongoing basis by abstracting information from trip reports, training reports, and team email correspondence.

The research was approved by the Health Science Faculty Human Research Ethics Committee of the University of Cape Town, the Western Cape Department of Health, and the Protection of Human Subjects Committee of FHI 360.

The demographic profiles of the participants sampled pre-intervention (herein “Pre-Int”) and post-intervention (“Post-Int”) were similar (Table 2). Participants, ranging in age from 18 to 53, were most commonly unmarried, in a steady relationship, and unemployed, with some secondary education. Significantly fewer Post-Int respondents reported being married or in a stable relationship compared to Pre-Int.

Most women at both time points reported their most recent pregnancy was unintended (Table 3). Of those interested in having another child, most wanted to wait at least three years or were unsure of desired timing. More than 80% of respondents in both samples reported current contraceptive use, primarily injectables. Fewer than 10% of women in both rounds had undergone sterilization, and just one woman, interviewed Post-Int, reported IUD use. Meanwhile, 77% of women Post-Int reported they wanted no future pregnancies. Asked if they would consider using the IUD, a third of the women Post-Int responded no; the most common reasons were that they lacked information about the method, were scared by it, or preferred sterilization. A quarter of the women Post-Int expressed no intention of undergoing female sterilization in the future, most commonly due to the permanency of the procedure or being afraid of it.

Table 4 presents the proportion of women demonstrating correct knowledge about the IUD and female sterilization as captured by the composite variables and their components. (Pre-Int participants were not asked about where they could access services since the research team understood those methods were largely inaccessible before the study.) Only 6% of women in the Post-Int survey answered correctly the four knowledge questions about the IUD, and 23% answered correctly the four knowledge questions about sterilization. The percentage of women correctly answering individual questions about the IUD declined between Pre-Int and Post-Int, while those correctly answering questions about female sterilization increased. Post-Int women were significantly more likely to know where HIV-positive women can undergo sterilization and that the procedure is safe for women using antiretrovirals. About one quarter of respondents Post-Int knew where they could access IUD services, compared to nearly two-thirds of women Post-Int who knew where they could access female sterilization services.

Roughly half the women interviewed at both time points reported having been counseled by a provider on family planning methods. Among those, the women reporting that they had been counseled on the IUD rose significantly from 8% Pre-Int to 24% Post-Int; women reporting counseling on female sterilization rose from 58% to 65%.

The 16 interviewed providers shared perspectives on serving the contraceptive needs of both PMTCT clients and postpartum women in general. Most providers clearly supported the right of all women to make informed and voluntary choices about planning or avoiding pregnancy. Some shared accounts of reassuring HIV-positive women that they could maximize the chance of a healthy pregnancy and baby with a well-timed pregnancy. By contrast, very few providers expressed disapproval of their HIV-positive clients becoming pregnant.

Several providers stated that prior to the intervention, they were not aware that methods other than pills, injectables, and condoms were available in health services. Nearly all stated that it would be useful to expand the range of contraceptive options for all women, not just HIV-positive women. Providers described difficulties women face in using short-term methods consistently, mentioning, for example, that women discontinue injectables due to missed clinic appointments for re-injections. Others recounted clients who had discontinued injectables due to concerns about irregular bleeding or amenorrhea. Providers reported that women discontinuing injectables for either reason then tended to abandon contraceptive use completely due to perceived lack of alternatives.

Providers shared perspectives on the importance of expanding the contraceptive options for HIV-positive clients in particular. Concerning hormonal contraceptive methods, about a third of providers mentioned the need to consider drug interactions and to modify re-injection schedules if the client is on specific anti-retroviral drugs or tuberculosis treatment. Most providers stated the importance of encouraging their HIV-positive clients to use condoms; only a minority reported that they promote simultaneous use of condoms and another contraceptive. Nearly all providers mentioned challenges women face in using condoms consistently, leading some to acknowledge that condoms alone provide inadequate contraceptive protection for most women. A few providers, however, indicated that condoms were the “best” contraceptive method for HIV-positive women.

Providers expressed favorable views toward promoting the IUD and female sterilization, with most reporting that both are appropriate for HIV-positive women. Some providers said alternatives to short-acting methods could ease the burden of HIV clients who already have frequent clinic visits. Half the providers specifically mentioned that an advantage of the IUD is the quick return to fertility upon its removal. Some explained that this characteristic is particularly helpful to women on antiretroviral therapy who are highly motivated to avoid pregnancy until the point when their health status has improved and they are ready to conceive. Providers similarly had favorable views toward sterilization but were clear in stating that it was appropriate only for women who were certain they had achieved their desired family size. One provider admitted that before attending study-sponsored training, she had not thought of educating HIV-positive clients about sterilization, reasoning it would not be necessary since they already use condoms.

Providers offered insights on the study intervention that helped to explain its failure. Several mentioned that the hands-on clinical component of the IUD insertion training should have been longer to ensure providers mastered the skills and gained the confidence to deliver the service. A few respondents referenced as a training model their experience learning to perform Pap smears, whereby clinicians stayed for an extended period in a high-volume facility to obtain extensive, supervised experience in performing the new procedure. Some providers suggested whole-facility trainings on the IUD and sterilization to ensure that all providers promote these methods. About half the interviewed providers mentioned that the intervention would have been strengthened with expanded client education about family planning. Several described a need for community education about the IUD in particular to counter concerns about the method and to publicize its advantages. A number of providers specifically mentioned clients’ fears about the IUD, including the pain of insertion, the risk of the device shifting position or being expelled, or the partner feeling the string. One provider recounted her successful experience stimulating interest in the IUD when she had an opportunity to discuss the method on a talk radio program.

Systematic documentation of intervention implementation captured information about the extent to which components were carried out as planned. Key intervention tracking findings are presented in Table 1, Column 2. The study team faced a host of challenges in implementing the intervention, some of which could be resolved with improved intervention design. For example, clearer criteria could be established to ensure that appropriate personnel are sent to trainings. Offering a condensed IUD service training to minimize providers’ absence from their work sites was the only option offered by district management, but it proved to be ineffective; instruction should have lasted longer and been conducted in a setting that provided trainees sufficient practical experience in insertion and removal techniques.

Other intervention shortcomings were rooted in health system constraints. Baseline provider capacity in basic family planning service delivery was weaker than anticipated; this fundamental deficiency needed to be addressed before training could be offered on IUD insertion and removal and the contraceptive considerations particular to HIV-positive women. Inadequacies of the health information system and limited time and resources for on-site supervision minimized district managers’ knowledge of the status of health service delivery within facilities and their influence over routine clinic operations. When the study-sponsored coach and research staff visited facilities, they sometimes noted high client loads for the staff available; at other times providers appeared to have a manageable number of clients to serve, but they demonstrated lack of motivation to take on additional responsibilities.