Interventions to address unequal gender and power relations and improve self-efficacy and empowerment for sexual and reproductive health decision-making for women living with HIV: A systematic review

Background Many women living with HIV experience gendered power inequalities, particularly in their intimate relationships, that prevent them from achieving optimal sexual and reproductive health (SRH) and exercising their rights. We assessed the effectiveness of interventions to improve self-efficacy and empowerment of women living with HIV to make SRH decisions through a systematic review. Methods and findings We included peer-reviewed articles indexed in PubMed, PsycINFO, CINAHL, Embase, and Scopus published through January 3, 2017, presenting multi-arm or pre-post intervention evaluations measuring one of the following outcomes: (1) self-efficacy, empowerment, or measures of SRH decision-making ability, (2) SRH behaviors (e.g., condom use, contraceptive use), or (3) SRH outcomes (e.g., sexually transmitted infections [STIs]). Twenty-one studies evaluating 11 intervention approaches met the inclusion criteria. All were conducted in the United States or sub-Saharan Africa. Two high-quality randomized controlled trials (RCTs) showed significant decreases in incident gonorrhea and chlamydia. Sixteen studies measuring condom use generally found moderate increases associated with the intervention, including in higher-quality RCTs. Findings on contraceptive use, condom self-efficacy, and other empowerment measures (e.g., sexual communication, equitable relationship power) were mixed. Studies were limited by small sample sizes, high loss to follow-up, and high reported baseline condom use. Conclusions While more research is needed, the limited existing evidence suggests that these interventions may help support the SRH and rights of women living with HIV. This review particularly highlights the importance of these interventions for preventing STIs, which present a significant health burden for women living with HIV that is rarely addressed holistically. Empowerment-based interventions should be considered as part of a comprehensive package of STI and other SRH services for women living with HIV.


Introduction
. Examined one or more interventions designed to address unequal gender power relations, increase self-efficacy, and/or increase empowerment around safer sex and reproductive decision-making for women living with HIV, 2. Compared women living with HIV who received the intervention to those who did not through a pre/post or multi-arm design, 3. Measured at least one of the following outcomes: (a) Self-efficacy, empowerment, or other measure of ability to make own decisions around condom use, pregnancy termination, birth spacing, childbearing, and other aspects of SRH, (b) SRH behaviors (such as condom use, contraceptive use, disclosure of HIV serostatus to partner) or (c) SRH outcomes (such as STIs, pregnancy). 4. Published in a peer-reviewed journal prior to the search date.
We included studies among all populations of women living with HIV, including adolescents (10-19 years), young people (20-24 years), adults (25+ years), and women of any age who were members of key populations (including female sex workers, women who use drugs, women in prisons or other closed settings, and transgender women) [13]. Given our focus on SRH decision-making, we excluded studies with children under ten years of age. If a study evaluated an intervention for both men and women, or for both women living with HIV and HIV-negative women, it was included only if outcome data were disaggregated for women living with HIV. We did not include self-efficacy for coping with HIV status; self-efficacy for adherence to medications; or general measures of self-efficacy, self-esteem, agency, or wellness not directly linked with SRH behaviors and outcomes. Articles from all countries and written in all languages were eligible for inclusion.

Data sources
The following electronic databases were searched for articles through January 3, 2017: PubMed, CINAHL, Embase, PsycINFO, and Scopus. We developed search terms for HIV, women, study design, and SRH to identify articles in PubMed (S1 Appendix), then adapted the search for other databases. Secondary reference searching was conducted on all included articles.

Data analysis
Initial screening of titles and abstracts was conducted by JR and SM. Potentially relevant citations were then independently screened in duplicate by JR and SM and resolved through discussion with CK. Full-text articles were reviewed for final eligibility decisions.
JR and SM independently extracted data in duplicate using standardized forms. Differences in data extraction were resolved through discussion and consensus. The following information was gathered from each included study: objectives, location, population characteristics, intervention description, study design, sample size, follow-up periods, loss-to-follow-up, analytic approach, outcome measures, comparison groups, effect sizes, confidence intervals, significance levels, conclusions, and limitations. JR and CK assessed study rigor using the Evidence Project's tool for evaluating multiple study designs in HIV behavioral intervention research [14], including assessment of comparison groups, random assignment and selection, followup rate, equivalency of comparison groups, and control for potential confounders.
Data were descriptively analyzed by coding categories and SRH outcomes. We did not meta-analyze due to differences in intervention design and outcome measurement across studies. However, we grouped similar measures (e.g., condom use self-efficacy) across studies and summarized findings by outcome.

Results
Database searches produced a total of 3,351 hits; 2,087 citations remained after removing duplicates (Fig 1). After initial screening, 151 citations were reviewed by two authors in duplicate, of which 73 were excluded for not meeting the inclusion criteria (e.g., qualitative studies, studies without relevant outcomes, or studies without findings for women living with HIV). Seventy-eight articles were pulled for full-text review, and 57 were excluded. Ultimately, 21 studies were included in the review covering 11 specific intervention approaches (Table 1).
Theoretical bases. All programs had an underlying theoretical basis. Theories used included social learning theory/social cognitive theory [15,21,22,25,30,34,35], the theory of gender and power [21,23,25,27,30,31,34], the empowerment and support model [28,29], the theory of reasoned action and theory of planned behavior [19,32,33], stages of change theory [16,23], and the AIDS risk reduction model [23]. Self-efficacy and empowerment review  (Table 1). Some interventions included multiple adaptations to different geographic context. In one instance, an in-person intervention was later adapted for multimedia [21]. Several were considered effective behavioral interventions by the U.S. Centers for Disease Control and Prevention.

Study outcomes
Sexually transmitted infections. Two studies measured STI incidence: the original WiL-LOW intervention in the southern USA and its South African adaptation. Both were highquality RCTs, although the South African adaptation had a shorter follow-up time (3 vs. 12 months) and smaller sample size (102 vs. 321 participants) [25,30]. Both studies showed significant decreases in STI incidence. The original WiLLOW intervention found a significant reduction in incidence of bacterial STIs (Chlamydia trachomatis and gonorrhea) over 12month follow-up in intervention versus control participants (OR = 0.20, 95% CI = 0.10-0.60). However, there was no significant change in Trichomonas vaginalis [25]. In the South African adaptation, the intervention group similarly showed a significant reduction in incidence of Chlamydia trachomatis (OR = 0.21, 95% CI = 0.07-0.59) and gonorrhea (OR = 0.10, 95% CI = 0.02-0.49) compared to the control group. The South African adaptation further showed a significant decrease in incidence of Trichomonas vaginalis (OR = 0.06, CI = 0.01-0.46), but no difference in incidence of bacterial vaginosis [30].

SISTA Adaptations (WiLLOW, PURSE)
Wingood Of the three SISTA adaptation RCTs that measured condom use, two showed significant increases [21,25], while the South African adaptation did not [30]. Most other high-quality RCTs also found significant positive impacts on condom use by various measurements [17,24,26,[31][32][33]35]. WDIP found progress (through stages of change) in condom use with main partner [16]. Three studies with high loss to follow-up rates (30-44% retention at followup) found mixed results on condom use [15,23,28]. Contraceptive use. Two studies measured contraceptive use. WDIP, an RCT, found that intervention participants were more likely to show progress (OR = 2.07, p = 0.08) and significantly less likely to relapse (OR = 0.43, p = 0.03) in contraceptive use compared to the comparison group [16]. PURSE, a non-randomized trial with 98 participants and high rates of followup, found that uptake of long-acting reversible contraception increased in both intervention and control groups three months after delivery, but there was no significant difference across groups (I: 87%, C: 81.8%, p = 0.34). The authors suggested this was due to both groups having access to nurses with training in enhanced family planning [34].
Self-efficacy and psychosocial measures. Four RCTs (the original WiLLOW, its South African adaptation, WDIP, and EVOLUTION) and one non-randomized trial (PURSE) measured condom use self-efficacy. The original WiLLOW program found that intervention participants had higher condom use self-efficacy over 12 months of follow-up (13.6 vs. 12.6; p = 0.001) [25]. PURSE also found significant increases in self-reported control over condom use (67.2% vs. 34.4%, p = 0.002) [34], and WDIP intervention participants showed higher selfefficacy for condom use with a main partner than control participants (OR = 2.01, p = 0.01) [16]. However, neither the small EVOLUTION pilot study nor the South African WiLLOW adaptation found a significant difference between intervention and control groups in condom use self-efficacy [30].
Other psychosocial outcomes also showed mixed results. The multimedia WiLLOW adaptation reported improvement in sexual communication self-efficacy (mean difference = 3.40, p = 0.004) [21], while EVOLUTION found no significant impacts on self-efficacy for sexual discussion or self-efficacy for limiting HIV risk behavior [27]. PURSE found significant increases in relationship power (2.5 vs. 2.1, p = 0.01) [34], whereas the South African WiL-LOW adaptation found no significant results for relationship control or condom negotiation [30]. Finally, PURSE intervention participants were more likely to report disclosing their HIV status to a partner (98.4% vs. 87.5%, p = 0.04) and vice versa (75.8% vs. 55.2%, p = 0.04) [34].

Discussion
All women living with HIV must be supported in their voluntary choices around sexual relationships and be given information and resources to engage in safe, enjoyable sexual experiences, or to not engage in sex based on their personal preference, with counselling and support tailored to their decision-making, desires and needs. Supporting women living with HIV in all their diversity to achieve their sexual and reproductive health and rights in all epidemic contexts requires overcoming major barriers to service uptake such as social exclusion and marginalization, criminalization, stigma, and gender inequality [37]. Addressing unequal gender and power relations and empowering women living with HIV may be one part of a comprehensive approach to achieve these goals.
This systematic review highlights the potential for increasing condom use and reducing incident STIs through empowerment interventions for women living with HIV. STIs continue to be an important public health issue that can facilitate sexual transmission of HIV and trigger some cancers. As stated in the WHO Global Health Sector Strategy on Sexually Transmitted Infections, 2016-2021, "the burden of morbidity and mortality worldwide resulting from sexually transmitted pathogens compromises quality of life, as well as sexual and reproductive health" [37]. Women living with HIV have high rates of STI co-infection, with a mean STI prevalence of 15.8% (standard deviation: 9.9) across studies in a recent global systematic review [38]. Although STI screening and treatment are a recommended part of the package of care for people living with HIV by the WHO [39,40] and PEPFAR [41], a comprehensive, rights-based approach to addressing STIs and other SRH issues is needed to facilitate STI prevention as well as treatment for women living with HIV.
Findings from our review were more mixed, however, for other outcomes, including contraceptive use, self-efficacy, and psychosocial measures. While these interventions hold promise, further work is needed to determine which components of interventions make them successful, for which populations, and on which outcomes.
Conclusions from this review are limited by the nature of the evidence base. The range of outcomes measured by the included studies was narrow, with the majority measuring condom use. Only a few studies measured other SRH outcomes, or more proximal outcomes such as empowerment and self-efficacy. Consequently, it is difficult to assess the impact of the interventions on women's self-efficacy or empowerment, and to understand the association between empowerment and SRH outcomes. Not measuring other outcomes limits the evidence for pathways to improved health for women living with HIV and their partners. Additionally, studies used a wide range of measures for condom use that affected our ability to compare across interventions and precluded us from conducting meta-analysis. Condom use reported in these studies was affected by high rates of initial reported use, creating a ceiling for measuring intervention impact. Many measures were also self-reported, introducing the possibility of recall and social desirability bias. Finally, the included studies were of mixed quality, with many limited by small sample size and low follow-up rates. The evidence base is further limited in geographic and population scope. Many important populations of women living with HIV, such as transgender women, were not included in any studies. Most included studies were conducted in the USA or were adaptations of interventions originally implemented there. Nevertheless, some interventions were determined to be effective when adapted to multiple contexts and feasible across settings. Finally, we did not include unpublished ("grey") literature or qualitative studies in our inclusion criteria; these studies may have provided additional insights into the effectiveness and outcomes of interventions.
Although this review focused on interventions with women, interventions with men that seek to address unequal gender and power relations are also essential to empower women in their SRH decisions. Recent evidence suggests that gender-transformative interventions to engaging men in HIV [42] and gender-based violence [43] hold promise; such programs seek to directly discuss and reconfigure gender roles in the direction of more gender equitable relationships [44]. Additionally, many gender inequalities exist at a structural level through cultural norms, laws, and institutions. Future research should also seek to implement structural-level interventions so that women may live in environments that better facilitate their control over their own sexual and reproductive health. Though structural-level interventions can be challenging both to implement and evaluate, they can have significant impact [45]. This is the first systematic review of interventions to improve self-efficacy and empowerment around safer sex and reproductive health decision-making for women living with HIV. The limitations of the existing evidence indicate a need for further research to determine the impact of empowerment and self-efficacy interventions. Future studies should include measurement of a wider range of sexual and reproductive health and rights outcomes, including both proximal empowerment and more distal health outcome measures. Studies should ensure the meaningful participation of the community of women living with HIV in study design. Interventions should also be explicit about how their content addresses unequal gender power relations. Such studies would allow for clear conclusions on how these types of interventions may improve the SRH of women living with HIV.