In June 2013, 56 international maternal health and HIV researchers and policy implementers were convened by the Maternal Health Task Force, the United States Agency for International Development, and the United States Centers for Disease Control and Prevention to share emerging evidence and identify priority knowledge gaps for improving programmatic responses to maternal health in contexts of high HIV prevalence. After this 2-day meeting, 30 researchers and policy implementers participated in a day-long meeting dedicated to reaching consensus on research and evaluation priorities for HIV and maternal health in sub-Saharan Africa. In addition to experience and technical expertise in HIV and maternal health, criteria for selecting the invited experts included representation of institutions from sub-Saharan Africa, North America, and Europe and a range of disciplinary backgrounds. Please see www.mhtf.org to consult the meeting agenda, presentations, and participant list. After a facilitated discussion of research and evaluation priorities, the experts divided into 3 groups tasked with reaching consensus on 4 priority areas for research and evaluation. Repetition of identified priorities allowed clustering in 3 broad areas for research and evaluation that were agreed on by consensus: (1) clinical management of pregnant and postpartum WLWH to address the leading causes of maternal death and evaluate the impact of expanded ART on maternal mortality and morbidity, (2) integrated service delivery models that simultaneously provide HIV screening and/or treatment and address malaria, tuberculosis, and/or the reproductive health of WLWH to optimize maternal health outcomes, and (3) an enabling social environment for women to begin and remain in care. The international working group that authored this article was formed to further develop the research and evaluation agenda and a conceptual framework (Fig. 1). Subsequently, narrative reviews of the peer-reviewed literature, relevant international guidelines and selected documents from the gray literature were conducted for each of the identified subtopics to summarize the existing evidence and knowledge gaps and generate additional specific research questions. The full working paper provides a comprehensive review of the literature and a detailed discussion of the proposed research and evaluation agenda.30 This article synthesizes broad research and evaluation questions that should be addressed to improve the maternal health of WLWH and provides illustrative examples. Table 1 outlines additional research questions.

The 3 thematic areas focus on different levels of the social and programmatic context relevant for HIV and maternal health (Fig. 1) as follows: (1) the “individual” level is focused on the clinical management and maternal health outcomes of WLWH; (2) the “health system” level considers how multiple priority health needs can best be addressed in resource-constrained health systems; and (3) at the “society” level, the intent is to assess how interventions that address gender and HIV-related discrimination and create a supportive social environment affect women's linkage to and retention in health services. Evidence-based interventions will resonate positively across all 3 levels. For instance, better information about the causes of maternal death among WLWH can inform clinical management and health system resource allocation. Subsequent reductions in maternal deaths can create more community demand for services, creating a positive cycle. Similarly, when the social context enables women to access and continue HIV care and adhere to ART, individual morbidity and mortality, sexual and vertical HIV transmission, and development of antiretroviral resistance will be reduced. Individual and community health outcomes and the sustainability of ART programs will be improved.

Tracking maternal health outcomes to assess the impact of ART on excess maternal morbidity and mortality is a priority as millions of WLWH begin and remain on treatment. Evidence from sub-Saharan Africa indicates that earlier initiation and longer duration of ART reduces the risk of maternal death among WLWH. When South Africa began initiating ART at a CD4⁺ cell count of ≤350 cells per microliter rather than at ≤200 cells per microliter, an 18% reduction in the institutional maternal mortality ratio for WLWH was observed during the first year.33 In Mozambique and Malawi, more than 30 days of ART before delivery has been associated with reduced risk of death during childbirth and postpartum.34 However, data from South Africa indicate higher maternal mortality rates among WLWH compared with HIV-negative women even when they are receiving ART,35 and results from a prospective cohort study of women initiating ART in Uganda found elevated risk of pregnancy-related mortality when pregnancy and the 1-year postpartum period were treated as a time-dependent risk factor.36 Data from the United States also indicate a persistently elevated risk of maternal morbidity and mortality among WLWH in the ART era.37 More studies, including cohort studies that allow valid comparisons by controlling for the health status of different groups of WLWH and that allow adjusted comparisons with HIV-negative women, are needed to tease out the relationships between HIV, pregnancy, time on ART, immune status, and maternal morbidity and mortality. As more countries offer lifelong ART for all pregnant and breastfeeding WLWH, it is critical to develop surveillance systems to track potential negative and positive effects of ART on women and infants.

Tuberculosis, malaria, pneumonia, and sepsis are the leading causes of maternal death among WLWH.38–40 Prevalence of anemia among pregnant WLWH is high and contributes to poor maternal and infant health outcomes.41–43 Data about greater incidence of hypertension or obstetric hemorrhage and outcomes for pregnant WLWH are inconclusive.44–49 Research to improve the prevention and clinical management of the leading causes of maternal morbidity and mortality among WLWH is needed. To exemplify, WLWH have 3 times the risk of puerperal sepsis as compared with HIV-negative women and a 6-fold greater risk when delivery is by caesarean.44 Maternal death audits in South Africa have identified an institutional maternal mortality ratio due to sepsis of 24.2 among WLWH as compared with 4.1 per 100,000 live births for HIV-negative women.50 Results from studies that examined provision of prophylactic antibiotics to WLWH before or during delivery have been mixed.51,52 Additional research is needed to inform sepsis prevention for WLWH during labor and delivery and as part of postpartum and postabortion care, as well as to elucidate the relationships between puerperal sepsis, immune status, and ART. Similar questions exist for other causes of maternal morbidity and mortality among WLWH (Table 1).

The promise of integrated service delivery is that meeting women's multiple health needs at a single point-of-care, possibly during a single visit and potentially by the same providers, can optimize women's contact with the health system and more effectively use scarce human and financial resources. Providing ART within maternal health services and eliminating barriers to ART initiation, for example by offering point-of-care CD4 testing53 or eliminating CD4 testing before initiation of ART,54 has been shown to dramatically increase uptake of ART among pregnant and breastfeeding women.55,56 Reviews of integration of interventions needed to prevent vertical HIV transmission with other maternal and child health (MCH) services have identified generally positive results in terms of coverage, but few studies provide information on health outcomes.57,58 Similarly, provision of contraceptives in HIV services and as a routine part of postnatal care has shown increased uptake by WLWH,59–61 but information about how these interventions affect unintended pregnancies or other health outcomes is lacking.62 Despite the overall positive impact, individual studies that identify negative effects in some settings, for instance, drops in child immunization rates in rural clinics when offer of HIV testing was added to immunization visits,63 deserve attention. There are also outstanding questions about the point on the MCH and HIV care continuum at which to provide services. For example, if WLWH begin ART in MCH services, when and how is the transition to chronic HIV care most successfully accomplished? Research and evaluation priorities include monitoring to ensure that service integration does not compromise the quality and coverage of core MCH and HIV services and medium-term assessment of health outcomes. Studies that evaluate the impact of adding other care components to MCH and HIV services, such as safer conception and contraceptive services, and malaria and tuberculosis prevention and treatment are also needed.

Increasing the quality of services and the number of interventions provided requires strengthening MCH and HIV services and ensuring adequate human resources. This strengthening is particularly important in the context of increased demands for ART provision to greater numbers of pregnant and breastfeeding WLWH and guidelines recommending earlier initiation of treatment for all people living with HIV. In addition, the skills required to deliver different services, for instance, HIV treatment adherence counseling or diagnosis and management of tuberculosis may differ from the skills required to deliver high-quality antenatal and obstetric care. As task shifting is implemented, evaluation is needed to assess health care staffing needs and skill level and optimal provider mix to deliver interventions safely and effectively while retaining health care providers on the job and pregnant and postpartum women in care.

Preconception counseling and safer conception strategies for women affected by HIV provide examples of the type of research and evaluation needed to guide the integration of service delivery. When and where on the continuum of MCH and HIV services should information about biomedical and behavioral interventions and services to support safer conception be offered to WLWH and to serodiscordant couples? How prepared are different cadres of HIV and MCH health care providers and community members to provide safer conception strategies? What facility- and community-based models are feasible and what resources are needed to ensure quality and coverage? Implementation research is needed to answer these questions and to assess the impact of information and service provision on reproductive behaviors, vertical and sexual HIV transmission, and maternal health outcomes. Additional questions about integrating interventions to support the reproductive choices of WLWH and to prevent, diagnose, and treat tuberculosis and malaria with MCH and HIV services are noted in Table 1.

Gender discrimination and HIV-related stigma violate human rights and contribute to poor health outcomes, in part by creating barriers to women's utilization of essential maternity and HIV services. Anticipation or experience of disrespect and abuse in maternity care and HIV-related stigma and discrimination, including internalized stigma, have been associated with avoidance of skilled delivery, refusal of antenatal HIV testing, and for WLWH, not enrolling in HIV services, and reduced adherence to ART for prevention of vertical HIV transmission.64–66 One of the most egregious forms of gender discrimination, Intimate Partner Violence (IPV), is associated with increased likelihood of acquiring and living with HIV66–68 and with poor reproductive health outcomes among all women.69–72 Qualitative research indicates that experiencing IPV can be a barrier to accessing HIV care and ART adherence for women in Africa.73 Studies from the United States and Canada associate IPV with reduced ART adherence, detectable viral load, and less engagement in HIV care.74–76 There is an urgent need to evaluate if facility- and community-based interventions that reduce disrespect and abuse during childbirth, HIV-related stigma, and IPV result in improved uptake and retention in care and better maternal health outcomes.64,77–81 Interventions, such as IMAGE and Stepping Stones, are particularly promising as they have been shown to simultaneously reduce IPV and HIV-related stigma and increase measures of women's empowerment,80,81 indicating the potential to address multiple social determinants simultaneously.

Research and evaluation of male involvement in HIV and MCH care is a related priority. When men attend prenatal care visits with their female partners and choose to test for HIV or women disclose their HIV status to male partners during pregnancy, male involvement has been associated with improved implementation of interventions to prevent vertical HIV transmission.82–84 Similarly, when women choose to disclose their HIV status, family-centered care has successfully engaged men and supported women to remain in care.85 However, how best to involve men systematically is a knowledge gap. For example, while some models of promoting couple counseling and HIV testing during pregnancy show promise for engaging men and assisting women with safe disclosure,86 others have resulted in half of the pregnant women asked to bring their male partner not returning for antenatal care.87 Research is needed to develop and evaluate programs that involve male partners while safeguarding women's access to and retention in care, safety, and autonomy.

In addition, policymakers and programmers should seriously consider the evidence associating participatory women's groups with reduced maternal mortality,88 and showing that community mobilization and peer support increase supply of, demand for, and implementation of interventions to prevent vertical HIV transmission, in addition to improving pregnant and postpartum WLWH's retention in care.15,89–91 Existing evidence indicates that further investment in and evaluation of strategies that actively involve and engage community members, including people living with HIV, is warranted.

Interventions designed to create an enabling environment for women to begin and remain in care and treatment should be sensitive to the diversity of women's experiences and needs, including those of women who may face additional barriers because of age, civil status, ethnicity, drug use, involvement in sex work, or other factors. Evaluation should consider proximate measures, such as uptake of services, retention in care, disclosure of HIV status, postpartum depression, and adherence to ART, as well as maternal health outcomes.