Non-specialist psychosocial support interventions for women living with HIV: A systematic review

ABSTRACT

Many women living with HIV experience a range of physical, social, and psychological challenges linked to their HIV status. Psychosocial support interventions may help women cope with these challenges and may allow women to make better decisions around their sexual and reproductive health (SRH), yet no reviews have summarized the evidence for the impact of such interventions on well-being and SRH decision-making among women living with HIV. We systematically reviewed the evidence for non-specialist delivered psychosocial support interventions for women living with HIV, which are particularly relevant in low-resource settings. Outcomes of interest included mental, emotional, social well-being and/or quality of life, common mental health disorders, and SRH decision-making. Searching was conducted through four electronic databases and secondary reference screening. Systematic methods were used for screening and data abstraction. Nine articles met the inclusion criteria, showing positive or mixed results for well-being and depressive symptoms indicators. No studies reported on SRH decision-making outcomes. The available evidence suggests that psychosocial support interventions may improve self-esteem, coping and social support, and reduce depression, stress, and perceived stigma. However, evidence is mixed. Most studies placed greater emphasis on instrumental health outcomes to prevent HIV transmission than on the intrinsic well-being and SRH of women living with HIV. Many interventions included women living with HIV in their design and implementation. More research is required to understand the most effective interventions, and their effect on sexual and reproductive health and rights.

KEYWORDS:

• HIV
• psychosocial
• women
• review
• systematic

Introduction

Living with HIV impacts physical, social and psychological aspects of a person’s health and life (Lingen-Stallard, Furber, & Lavender, 2016; WHO, 2016b). A higher proportion of people living with HIV experience depression and anxiety than the general population: some studies reported higher prevalence in females than males, and a negative impact of psychological distress on quality of life (Lowther et al., 2016; Niu, Luo, Liu, Silenzio, & Xiao, 2016; Vo et al., 2016). A large global survey led by, and for, women living with HIV on their sexual and reproductive health and rights found that 82% of women reported depressive symptoms and 78% reported experiencing rejection (Orza et al., 2015). These and other experiences, including social exclusion, stigma, insomnia, and rights violations, challenged their overall well-being, their sexual and reproductive health (SRH), and their ability to exercise their rights (Orza et al., 2015).

Psychosocial support addresses the ongoing psychological and social problems of people living with HIV, their partners, families and caregivers (WHO, 2016b). Despite high prevalence of psychosocial challenges among women living with HIV, identified associations between psychosocial concerns and SRH outcomes such as child-bearing and sexual desire (EngenderHealth, 2009; Malary, Khani, Pourasghar, Moosazadeh, & Hamzehgardeshi, 2015), and recommendations to include psychosocial support in HIV care (Paudel & Baral, 2015; WHO, 2016b), no reviews have investigated the impact of psychosocial support interventions on SRH health decision-making among women living with HIV. Existing reviews concerning psychosocial support have focused on: (1) clinical outcomes or care engagement outcomes (Bateganya, Amanyeiwe, Roxo, & Dong, 2015; Genberg et al., 2016), (2) specific intervention types including support groups (Bateganya et al., 2015; Mundell et al., 2011; Paudel & Baral, 2015), or (3) outcomes for mixed-gender groups (Crepaz et al., 2008; Scott-Sheldon, Kalichman, Carey, & Fielder, 2008b). More research has been recommended to understand psychosocial support interventions for people living with HIV, including gender considerations (Bateganya et al., 2015; Crepaz et al., 2008; Paudel & Baral, 2015). Many psychosocial support interventions can be delivered by non-specialist healthcare workers or lay providers (Kaaya et al., 2013; Mundell et al., 2011; Richter et al., 2014). This is particularly important in resource-limited settings (WHO, 2016b).

To inform the World Health Organization (WHO) Consolidated guideline on the sexual and reproductive health and rights of women living with HIV, we conducted a systematic literature review on psychosocial support for women living with HIV to identify effective forms of non-specialist psychosocial support for improving women’s intrinsic well-being and ability to make SRH decisions (for example, around contraceptive use, abortion, or prevention of sexually transmitted infections). We focused on intrinsic well-being – health outcomes that directly benefit the woman, herself – rather than secondary or instrumental outcomes that benefit others, such as onward HIV transmission to partners or children. Outcomes were guided by the holistic, interconnected priorities identified through a global survey on SRH priorities of women living with HIV (Salamander Trust, 2014). We focused on health service providers without specialized training in mental health, as interventions with trained mental health practitioners providing formal mental health services are covered in other reviews and existing WHO guidelines (Genberg et al., 2016; Kennedy, Fonner, Armstrong, O’Reilly, & Sweat, 2015; Kennedy, Haberlen, Amin, Baggaley, & Narasimhan, 2015; Perazzo, Reyes, & Webel, 2016; WHO, 2013).

Methods

Inclusion criteria

Articles meeting the following criteria were included in the review:

1. Published in a peer-reviewed journal through the search date of 25 February 2016.

2. Measured one or more of the following primary or secondary outcomes: Primary: (1) Measures of mental/emotional/social well-being and/or quality of life (e.g., decision-making capacity, coping skills, stigma reduction). (2) Measures of common mental health disorders (e.g., depression, anxiety). Secondary: Decision-making around SRH services (e.g., contraception, abortion, cervical cancer screening, sexually transmitted infection services, etc.).

3. Controlled trials (randomized and non-randomized), comparing provision of psychosocial support by providers without specialized training in mental health to people who received alternative care or treatment. Providers could include a trained, certified health professional without specialized training in psychiatric or psychological care (e.g., doctor, nurse without specialized mental health training) or a lay provider with or without brief training in provision of psychosocial support services (WHO, 2012).

4. Presented data for women living with HIV who received the intervention, either because the study included only women living with HIV or because data were disaggregated by gender and HIV status.

Search strategy, screening, and data extraction and analysis

We searched four electronic databases: PubMed, PsycINFO, CINAHL, and EMBASE. The following terms comprised the PubMed search, which was adapted for the other databases: (“HIV positive” [tiab] OR “living with HIV” [tiab] OR “HIV infected” [tiab]) AND (randomized controlled trial[pt] OR controlled clinical trial[pt] OR randomized[tiab] OR random*[tiab] OR clinical trials as topic[mesh:noexp] OR “non-randomized trial” [tiab]) AND (“psychosocial support” [tiab] OR “social support” [tiab] OR “support groups” [tiab] OR “counselling” [tiab] OR counseling [tiab]) AND (depression or stress OR coping OR cope OR self-esteem OR self-confidence OR wellness OR stigma OR “mental health” OR “well-being” OR “quality of life”). We also conducted secondary reference searching on nine systematic reviews identified in our search or known by the researchers to be related to the topic (Biswas, 2007; Carrico & Antoni, 2008; Feyissa, Lockwood, & Munn, 2015; Kennedy, Fonner, et al., 2015; Kennedy, Medley, Sweat, & O’Reilly, 2010; Mwai et al., 2013; Paudel & Baral, 2015; Scott-Sheldon, Kalichman, Carey, & Fielder, 2008a; Zajac et al., 2015).

Citations identified through the search strategy underwent initial screening by one reviewer, based on title and abstract. Two reviewers then conducted second level screening of all potentially relevant citations, comprising those related to HIV and well-being, quality of life, mental health, or sexual or reproductive health decision-making, based on title, abstract, citation information, and descriptor terms, conferring as needed. One reviewer then screened potentially relevant full text articles for outcomes measured, study design, type of healthcare provider, and data exclusively disaggregated for women living with HIV to determine final study selection. We contacted, and received clarification from, two authors about articles for which we could not determine study eligibility based on published information. Queries about article eligibility were discussed among the research team and resolved through consensus. For the included studies, study identification, intervention descriptions, provider type, theoretical grounding, measurement approaches, study design, and outcomes were abstracted into standardized tables by the reviewer who conducted the full text screening. Quality assessment of included articles was conducted using an adapted version of the Evidence Project’s tool (Fonner, Armstrong, Kennedy, O’Reilly, & Sweat, 2014; Kennedy, Fonner, et al., 2015; Zajac et al., 2015).

Meta-analysis was not conducted due to the limited number of articles measuring the same outcomes using a consistent approach. We present a descriptive analysis of the study findings based on the data extracted. We conducted sub-group analyses of outcomes that were measured by three or more studies.

Results

Studies identified

The search string identified 451 articles with duplicates removed (Figure 1). An additional 68 articles were identified through secondary reference searching. After initial abstract screening, 44 articles were pulled for full-text review. Nine studies met the inclusion criteria.

Figure 1. PRISMA flow diagram of identified and included citations.

Study descriptions

A descriptive summary of included studies is presented in supplementary material.

Settings

Included studies were conducted in three countries: South Africa (Eloff et al., 2014; Futterman et al., 2010; Mundell et al., 2011; Richter et al., 2014), Tanzania (Kaaya et al., 2013), and the southern United States (US) (Echenique et al., 2013; Klein, Lomonaco, Pavlescak, & Card, 2013; Miles et al., 2003; Wingood et al., 2004). All nine studies recruited participants from health facilities, care agencies, or health departments.

Populations

Four of the five studies in Africa included only pregnant women (Futterman et al., 2010; Kaaya et al., 2013; Mundell et al., 2011; Richter et al., 2014), with two specifically studying women with recent HIV diagnosis (Mundell et al., 2011; Richter et al., 2014). Participant mean age was 26–27 years; the youngest specified age was 16 years (Futterman et al., 2010). The remaining African study included women with mean age 33 years who had children aged 6–10 years (Eloff et al., 2014). Many women in these studies had low income. Three of the four US-based studies included only, or majority, African American women (Klein et al., 2013; Miles et al., 2003; Wingood et al., 2004). One of those included only mothers (Miles et al., 2003). Mean age in these studies ranged from 35–40 years. Most women had low or very low income. The remaining US-based study focused on older women (over age 45) including a mix of Hispanic, White, and Black women (Echenique et al., 2013). No literature included data specific to women living with HIV from key populations, such as women who inject drugs, are in prison, transgender, lesbian or bisexual, adolescents, or sex workers (WHO, 2016a).

Study designs, quality, interventions, and outcomes

Study designs included six individually-randomized controlled trials (RCT) (Echenique et al., 2013; Eloff et al., 2014; Kaaya et al., 2013; Klein et al., 2013; Miles et al., 2003; Wingood et al., 2004), one facility-randomized controlled trial (Richter et al., 2014) and two non-randomized trial designs (Futterman et al., 2010; Mundell et al., 2011). Total sample size per study ranged from 71–1200 participants. Follow-up time extended from eight weeks to eighteen months. Length of follow-up time was not consistently associated with improved outcomes. Study quality varied. Most studies were relatively methodologically rigorous, with randomization to study arms; however, not all were comparable between arms on baseline sociodemographic characteristics and outcome measures (see Table 1).

Table 1. Study Rigor assessment.

Study	Random assignment (group or individual) to the intervention	Comparison groups equivalent at baseline on socio-demographics	Comparison groups equivalent at baseline on outcome measures	Time period and follow-up rate
Mundell et al. (2011)	No	No	No: Depression higher in intervention group	2 months & 8 months 69% comparison; 90% intervention
Kaaya et al. (2013)	Yes	Yes	Yes	8–10 weeks 84.3% standard of care; 81.5% intervention
Eloff et al. (2014)	Yes	No	Yes	18 months 88% standard of care; 81% intervention; 72% completed all 3 interviews
Futterman et al. (2010)	No	No	No: Depression higher in intervention group	6 months 40% control; 48% intervention
Richter et al. (2014)	Yes	Yes	Yes	1.5 months 71% standard of care; 69% intervention
Wingood et al. (2004)	Yes	No	Yes	12 months 90.3% comparison; 85.2% intervention
Klein et al. (2013)	Yes	Yes	No: Having someone to confide in about HIV lower in intervention group^a	3 months 98.9% control; 93.1% intervention
Echenique et al. (2013)	Yes	Yes	Not reported^t	6 months 41% control; 32.5% intervention
Miles et al. (2003)	Yes	No	Not reported	1 month & 6 months 1 month: 74.0% control; 62.7% intervention 6 months: 58.0% control; 50.8% intervention

^aAlso not equivalent on other outcome measures not of direct interest to this review; ^tEquivalent on other outcome measure not of direct interest to this review.

Psychosocial support interventions encompassed a range of intervention modalities and intended outcomes. Eight of the nine studies used an explicit theoretical framework that guided the intervention design and implementation (Eloff et al., 2014; Futterman et al., 2010; Kaaya et al., 2013; Klein et al., 2013; Miles et al., 2003; Mundell et al., 2011; Richter et al., 2014; Wingood et al., 2004). Four studies worked with women living with HIV to design and/or refine the intervention (Eloff et al., 2014; Kaaya et al., 2013; Mundell et al., 2011; Wingood et al., 2004).

The most common mode of intervention delivery was a support group or facilitated group (Echenique et al., 2013; Eloff et al., 2014; Futterman et al., 2010; Kaaya et al., 2013; Mundell et al., 2011; Wingood et al., 2004). Two studies used an individual mentorship approach (Miles et al., 2003; Richter et al., 2014), and one used interactive computer modules (Klein et al., 2013). Interventions were delivered in sessions, ranging from 2–24 exposures. Intervention providers included nurses, midwives, masters-level psychology students, community health workers and peer educators. Five studies utilized interventions delivered or co-delivered by women living openly with HIV (Echenique et al., 2013; Futterman et al., 2010; Mundell et al., 2011; Richter et al., 2014; Wingood et al., 2004). The women’s well-being was the primary objective in two studies (Miles et al., 2003; Richter et al., 2014). The other studies’ primary aims were risk reduction, disclosure or child health. The included studies used different metrics to measure a range of psychosocial support outcomes (Table 2). None of the included studies measured our secondary outcome of interest, SRH decision-making.

Table 2. Psychosocial support outcome measures in included studies.

Psychosocial support outcome measured	Studies utilizing measure	Measurement tool
Self-esteem	Mundell et al. (2011), Klein et al. (2013)	Rosenberg Self-Esteem Scale 10-item likert; 10-item scale (unspecified)
Coping (Active, Positive, Self-coping	Eloff et al. (2014), Futterman et al. (2010), Mundell et al. (2011), Klein et al. (2013)	Brief COPE (3 studies including active, self, positive), 18-item coping scale (unspecified)
Avoidant coping	Mundell et al. (2011), Eloff et al. (2014)	Brief COPE (2 studies)
Coping: help from others	Eloff et al. (2014)	Brief COPE
Positives Social support	Mundell et al. (2011), Klein et al. (2013)	Multidimensional Social Support Inventory, 23-question scale (unspecified)
Social support availability	Wingood et al. (2004), Futterman et al. (2010)	Number of network ties; Medical Outcomes Study Social Support Survey
Social support satisfaction	Futterman et al. (2010)	Medical Outcomes Study Social Support Survey
Negative social support	Mundell et al. (2011)	Multidimensional Social Support Inventory
Depression	Eloff et al. (2014), Futterman et al. (2010), Miles et al. (2003), Mundell et al. (2011), Kaaya et al. (2013), Richter et al. (2014)	Center for Epidemiologic Studies Depression scale (CES-D) (4 studies); 15-item depression subscale of the Hopkins Symptom Checklist-25 (HSCL-25); General Health Questionnaire (GHQ)
HIV discomfort	Futterman et al. (2010)	Read scenarios, replied with discomfort level on Likert scale
HIV stigma (perceived)	Echenique et al. (2013), Miles et al. (2003)	HIV Stigma Scale; Demi HIV Stigma Scale
Mood	Miles et al. (2003)	Profile of Mood States
HIV worry	Miles et al. (2003)	HIV Worry Scale designed for study
Stress	Klein et al. (2013)	14-item stress scale (unspecified)

Study results

Results for each outcome measured are presented in Supplementary material.

Mental, emotional, social well-being and/or quality of life

Coping

Included studies measured three different types of coping: (1) active, positive self-coping (Eloff et al., 2014; Futterman et al., 2010; Klein et al., 2013; Mundell et al., 2011), (2) avoidant (Eloff et al., 2014; Mundell et al., 2011), and (3) coping: help from others (Eloff et al., 2014). Metrics included the brief COPE (Eloff et al., 2014; Futterman et al., 2010; Mundell et al., 2011) and an 18-item coping scale (Klein et al., 2013). Results were mixed. Only one study, using a non-randomized trial design, found significant improvements in active (t = 2.68, p-value = 0.01) and avoidant coping (t = −2.02, p-value = 0.04) (Mundell et al., 2011); all other results were null. Active coping was the only coping outcome measured by three or more studies. There were no clear patterns in active coping outcomes based on types of interventions (group versus individual), target populations (pregnant women versus others), the amount or frequency of interactions (2–24 sessions), follow-up time (3–18 months), or primary intervention objective (well-beingversus instrumental).

Social support

Included studies measured four types of social support: positive support (Klein et al., 2013; Mundell et al., 2011), support availability (Futterman et al., 2010; Wingood et al., 2004), support satisfaction (Futterman et al., 2010), and negative support (Mundell et al., 2011). Metrics included the “multidimensional social support inventory” for positive and negative support (Mundell et al., 2011), the “medical outcomes study social support survey” for availability and satisfaction (Futterman et al., 2010), a 23-question scale for positive support (Klein et al., 2013), and the number of network ties for support availability (Wingood et al., 2004). No studies showed improvement in positive social support or social support satisfaction. However, both studies measuring social support availability, a facility-randomized controlled trial (regression-b = 9.32, SE = 3.53, p-value = < 0.05) (Futterman et al., 2010) and an individually-randomized controlled trial (% relative change = 12.0, 95% CI: 0.5, 23.4, p-value = 0.02) (Wingood et al., 2004) saw significant improvements among intervention compared to control participants. The only study measuring negative social support, a non-randomized trial, saw a significant decline among a sub-analysis of participants with high attendance at 8 months of follow-up (t = −2.41, p-value = < 0.05) (Mundell et al., 2011).

Self-Esteem

Self-esteem was measured by two studies with mixed results. One showed significant improvement in self-esteem among a sub-group of participants with high attendance at 2 months of follow-up using the Rosenberg Self-Esteem Scale (t = 2.11, p-value = < 0.05) (Mundell et al., 2011). Results from the second study, an RCT using an unspecified 10-item scale, were null (Klein et al., 2013).

Perceived HIV stigma

Both studies that measured perceived HIV stigma, one using the HIV Stigma Scale (Echenique et al., 2013), and the other using the Demi HIV Stigma Scale (Miles et al., 2003), saw significant improvements using individual RCTs. The intervention group in the first study had a significant decline in mean stigma score (baseline = 104.6, endline = 96.5, p-value = 0.05) (Echenique et al., 2013). The second study found a significant decrease in perceived stigma (glm effect = −0.25, p-value = 0.03) (Miles et al., 2003).

Other measures

The following outcomes were each measured by one study: HIV discomfort, using scenarios and discomfort levels on a Likert scale (Futterman et al., 2010); mood, using Profile of Mood States (Miles et al., 2003); HIV worry, using an HIV worry scale (Miles et al., 2003); and stress, using a 14-item stress scale (Klein et al., 2013). Findings were null for all outcomes other than stress, which declined significantly in the intervention group compared to the control group (% relative change = −12.08, 95% CI:−3.76, −13.72) (Klein et al., 2013).

Common mental health disorders

Depressive symptoms

Six of the nine studies measured depressive symptoms, with mixed results. Two studies, one facility RCT and one non-randomized, facility-based trial design, showed significant differences in depression: one measured odds of not having depressive symptoms (OR = 2.55, 95%CI: 1.37, 4.76, p-value = 0.002) (Richter et al., 2014), the other measured presence of depressive symptoms (Regression-b = −4.43, SE = 1.62, p-value = <0.01) but had unequal prevalence of depressive symptoms between comparison groups at baseline (Futterman et al., 2010). All other study results for this outcome were null (Eloff et al., 2014; Kaaya et al., 2013; Miles et al., 2003; Mundell et al., 2011). Measurement tools included the CES-D (Eloff et al., 2014; Futterman et al., 2010; Miles et al., 2003; Mundell et al., 2011), the 15-item depression sub-scale of the Hopkins Symptom Checklist-25 (Kaaya et al., 2013), and the General Health Questionnaire (Richter et al., 2014). As with coping, there were no clear patterns in depressive symptom outcomes based on types of interventions (group versus individual), target populations (pregnant women versus others), amount or frequency of interactions (6–24 sessions), follow-up time (6 weeks–18 months), or the primary intervention objective (well-being versus instrumental).

Intervention characteristics

All studies that included a woman living with HIV as at least one of the intervention providers showed significant results (Echenique et al., 2013; Futterman et al., 2010; Mundell et al., 2011; Richter et al., 2014; Wingood et al., 2004). Five of the six studies that showed significant results where people, not a computer, delivered the intervention had women living with HIV as providers (Echenique et al., 2013; Futterman et al., 2010; Mundell et al., 2011; Richter et al., 2014; Wingood et al., 2004). All but one of the studies that showed significant results reported extensive grounding in theory and formative work to tailor the intervention to the target population (Futterman et al., 2010; Klein et al., 2013; Miles et al., 2003; Mundell et al., 2011; Richter et al., 2014; Wingood et al., 2004). All but four of the studies (Eloff et al., 2014; Kaaya et al., 2013; Klein et al., 2013; Wingood et al., 2004) experienced poor participant retention of less than 80% retained, with three studies experiencing retention of 51% or lower (Echenique et al., 2013; Futterman et al., 2010; Miles et al., 2003).

Discussion

Our systematic review identified 9 articles measuring the impact of psychosocial support interventions delivered by non-specialist providers to women living with HIV. Study designs were of variable quality and generally showed mixed or positive results for well-being and depressive symptoms indicators. In the literature, psychosocial support interventions were primarily designed and evaluated with regard to instrumental outcomes, such as care engagement (Wanyenze et al., 2011), medication adherence (Cook, Carrington, Schmiege, Starr, & Reeder, 2015; van Loggerenberg et al., 2015), sexual risk reduction (Echenique et al., 2013; Klein et al., 2013; Morin et al., 2007; Wingood et al., 2004), and disclosure (Mundell et al., 2011). No studies measured SRH decision-making. Women are more often supported in their role as expecting mothers or mothers (Eloff et al., 2014; Futterman et al., 2010; Richter et al., 2014), and less frequently supported as individuals with multiple, interactive identities and priorities. These gaps reflect several priority issues raised in the global survey of women living with HIV, including the desire for affordable, accessible, continuous, holistic, and integrated psychological support in HIV and SRH care from well-trained, compassionate providers covering topics including self-confidence, self-stigma, anxiety, isolation, acceptance and positive body image, and knowing and believing in rights (Salamander Trust, 2014). To improve well-being, SRH, and uphold women’s rights, both physical and psychological health must be prioritized. Priorities and SRH preferences of women living with HIV as individuals should remain at the center of healthcare provision, which should consequently not focus solely on reducing vertical HIV transmission. More research is needed on the role of psychosocial support in SRH and rights.

Most articles included in this review reported the meaningful involvement of women living with HIV in the design and delivery of interventions. Meaningful inclusion of women living with HIV was noted as a critical element of HIV interventions in the global survey (Salamander Trust, 2014). This, and formative research conducted with women living with HIV, will help to ensure that interventions are contextually appropriate (Kako, Wendorf, Stevens, Ngui, & Otto-Salaj, 2016) and uphold the rights of women living with HIV (Orza et al., 2015). Theoretically-grounded research helps identify anticipated mechanisms and pathways of change, which is particularly important given the multi-faceted nature of psychosocial support (Glanz, Rimer, & Viswanath, 2008). Understanding the explicit theoretical grounding guiding behavior change interventions may help to support replication. However, further research is needed. Until more studies using similar outcome metrics are available, it is difficult to evaluate the effect of theoretical grounding on outcomes for women living with HIV.

Women living with HIV from key populations may have multiple additional identities which can render them even more socially disadvantaged (Orza et al., 2015). This disadvantage is associated with an increased experience of mental health issues, and these women are thereby more vulnerable to human rights violations and intersectional violence (Orza et al., 2015; WHO, 2016a). It is critical that future studies actively include them and conduct analyses to understand sub-group intervention effectiveness.

Implementation of psychosocial support interventions should work to increase participation retention, which may be related to study follow-up procedures, intervention timing (evening or weekend v. weekday), participant engagement with the subject matter and delivery, or other factors. The instrumental versus intrinsic focus of the intervention may also play a role. Also, research on psychosocial support interventions should also consider long-term follow-up post-intervention to determine the length of intervention effect, its perceived relevance to the women concerned, and possible follow-on support required.

Evidence base and study limitations

The narrow population and geographical scope of the existing evidence base may limit the generalizability of the results. Four of the nine studies included only pregnant women with two more focused on mothers or primary caregivers. Only three countries were represented. The included studies all recruited women from healthcare organizations, so women living with HIV not currently engaged in formal care were not included.

Seventeen articles, including two from China and several with alternative delivery modalities (e.g., telephone), were not included because they did not disaggregate results by gender. While the intervention modalities included in this review, support groups and individual mentorship, reflect common intervention approaches, several approaches hypothesized to improve psychosocial wellbeing, such as physical exercise, were not used in any of the nine included articles.

This systematic review was limited to peer-reviewed articles. This helps ensure the inclusion of high quality research, but may miss potentially valuable, unpublished interventions, which may be more likely to be voluntary, locally-appropriate or sustainable without external support. Also, a single reviewer screened articles and conducted data abstraction, though a second coauthor did review coding results.

Conclusions

While outcomes were mixed, the available evidence suggests that support group and individual mentorship psychosocial support interventions may improve coping, self-esteem, and social support, and reduce depression, stress, and perceived stigma among women living with HIV. Interventions are optimally delivered by women living openly with HIV or by them together with others. Interventions should be relevant, acceptable and tailored to the health of women living with HIV. This may be best facilitated through theoretically-grounded interventions designed using formative research. Future studies using consistent outcome metrics are needed to understand the most effective intervention types and delivery mechanisms, as well as the relationship between psychosocial support and sexual and reproductive health and rights.

Acknowledgments

We wish to acknowledge the contributions of Dr. Sabina Haberlen.

Disclosure statement

The authors have no competing interests to declare.

Additional information

Funding

This work was supported by the World Health Organization, Department of Reproductive Health and Research.