Food insecurity, sexual risk behavior, and adherence to antiretroviral therapy among women living with HIV: A systematic review

ABSTRACT

Gender inequalities shape the experience of food insecurity among women living with HIV (WLHIV). We systematically reviewed the impact of food insecurity on sexual risk behaviors and antiretroviral therapy (ART) adherence among WLHIV. We included qualitative or quantitative peer-reviewed articles, extracted data in duplicate, and assessed rigor. Seven studies, from sub-Saharan Africa, North America, and Europe, met inclusion criteria. Food insecurity was associated with increased sexual risk through transactional sex and inability to negotiate safer sex. Hunger and food insecurity were barriers to ART initiation/adherence. Multidimensional programming and policies should simultaneously address poverty, gender inequality, food insecurity, and HIV.

Background

Food insecurity, defined as existing when “the availability of nutritionally adequate and safe foods or the ability to acquire acceptable foods in socially acceptable ways is limited or uncertain” (Anderson, 1990), is likely to disproportionately affect women because of global gender inequalities (Amin, 2015; Ivers & Cullen, 2011). Women are more vulnerable given their lack of access to and control over resources including land, employment, and money (United Nations Department of Economic and Social Affairs, 2015). Food insecurity has important implications for nutrition and health, particularly in women. For women, food insecurity is related to poor mental health such as depression and anxiety (Whitaker, Phillips, & Orzol, 2006), increased risk of high-risk sexual behavior (Weiser et al., 2007), and staying in abusive relationships (Miller et al., 2011). In addition, nutrition demands are higher during pregnancy, and food insecurity has been linked to poor pregnancy outcomes such as low birth weights and gestational diabetes (Borders, Grobman, Amsden, & Holl, 2007; Laraia, Siega-Riz, & Gundersen, 2010). Special consideration is warranted in understanding the impact of food insecurity on women because of the primary role women often play within families in preparing food and feeding families (Ivers & Cullen, 2011).

Women living with HIV face specific, and possibly greater, challenges related to food insecurity compared to other women. Food insecurity has been associated with poor clinical outcomes for people living with HIV, including lower CD4 counts (Aibibula et al., 2016; Weiser et al., 2014; Weiser et al., 2013), increased opportunistic infections (Weiser et al., 2012), increased hospitalizations (Weiser et al., 2013; Weiser et al., 2012), decreased bioavailability and efficacy of antiretroviral therapy (ART) (Bartelink et al., 2014), incomplete viral load suppression (Feldman, Alexy, Thomas, Gambone, & Irvine, 2015; Koss et al., 2015; Weiser et al., 2014) and increased risks of mortality even on ART (Anema et al., 2013; Feldman et al., 2015; Weiser et al., 2009), and suboptimal ART adherence (Singer, Weiser, & McCoy, 2015). Pregnant and lactating women living with HIV have additional nutritional requirements that may make them particularly vulnerable (Montgomery, 2003), and food insecurity has been associated with an increased risk of mother-to-child transmission of HIV among pregnant women living with HIV (McCoy et al., 2015; Villamor et al., 2005).

Food insecurity can have other consequences related to sexual and reproductive health for women. Researchers have identified food insecurity as a motivation for HIV-negative women to engage in transactional sex to obtain food and/or resources, or to engage in sex work (Dunkle et al., 2004; Oyefara, 2007; Tsai, Hung, & Weiser, 2012; Weiser et al., 2007), thereby increasing their risk for HIV. In a large study from Botswana and Swaziland, food insufficiency was correlated with both inconsistent condom use with a non-primary partner and intergenerational sexual relationships (Weiser et al., 2007). The correlation between food insecurity and HIV risk behaviors was also observed in a study of sexually active women in Brazil: food insecurity was associated with reduced odds of consistent condom use in the past year and condom use at last sex (Tsai et al., 2012).

To inform the World Health Organization Guideline on the sexual and reproductive health and rights of women living with HIV, we conducted a systematic review of the impact of food insecurity on two key outcomes: (1) sexual risk behaviors and (2) ART adherence among women living with HIV. Previous reviews have examined the association between food insecurity and ART adherence, but these have not disaggregated findings by gender and have only included quantitative studies (Singer et al., 2015; Young, Wheeler, McCoy, & Weiser, 2014). To our knowledge, no previous studies have reviewed the relationship between food insecurity and sexual risk behaviors among women living with HIV. Our goal was to review the literature on these relationships and inform strategies for creating an enabling environment that will address the sexual and reproductive health needs of women living with HIV and ensure that HIV and food security are jointly considered in the implementation of policies and programs.

Methods

We used systematic search and screening methods to identify articles meeting our inclusion criteria, and followed PRISMA guidelines for reporting systematic reviews (Moher et al., 2015).

Inclusion criteria

To be included in the review, an article had to (a) present primary data, either qualitative or quantitative, collected among women living with HIV; (b) describe the associations between food insecurity and ART adherence, safer sex, or sexual risk behaviors; and (c) be published in a peer-reviewed journal prior to the search date of February 18, 2016.

No restrictions were placed based on the type of population (specific sub-populations of women living with HIV) or study location. When studies included both men and women living with HIV, we included them if they stratified results by gender.

Search strategy

We searched the following electronic databases to identify articles for inclusion: PubMed, CINAHL (Cumulative Index to Nursing and Allied Health Literature), PsycINFO, and Sociological Abstracts. We did not include search terms around the concepts of adherence and sexual risk behavior in order to broaden the search and reduce likelihood of missing potentially relevant studies. The following terms were used in PubMed and adapted for each database:

(“HIV positive” [tiab] OR “living with HIV” [tiab] OR “HIV infected” [tiab]) AND (Women's Health [mesh] OR women [tiab] OR woman [tiab] OR female* [tiab] OR gender [tiab]) AND (“food security” [tiab] OR “food insufficiency” [tiab] OR “food insecurity” [tiab] OR hunger [tiab] OR “food scarcity” OR (“food security” AND “risk”)).

We also performed secondary reference searching on all studies included in the review, as well as articles included in two previous reviews of food insecurity and ART adherence among people living with HIV (Singer et al., 2015; Young et al., 2014).

Screening abstracts

Titles, abstracts, citation information, and descriptor terms of citations identified through the search strategy were screened by a study staff member. Full text articles were obtained of all selected abstracts, and final decisions on inclusion were made by two members of the study team. All articles underwent data abstraction in duplicate with differences in interpretation between reviewers resolved through discussion. Rigor for quantitative studies was assessed using an approach designed for multiple study designs in the field of HIV research by the Evidence Project (Denison, O'Reilly, Schmid, Kennedy, & Sweat, 2008; Medley, Kennedy, O'Reilly, & Sweat, 2009). Qualitative studies were excluded from study rigor assessments, given the limitations of using checklists to assess rigor of qualitative studies, but we did consider aspects of rigor where appropriate (Higgins & Green, 2011).

Results

Of 66 unique articles identified through the database search and secondary searches, seven articles ultimately met the inclusion criteria for the review (Figure 1). Included studies came from Canada, the Democratic Republic of Congo, France, Swaziland, the United States, Uganda, and Zambia (Table 1). Five qualitative studies (Fielding-Miller, Mnisi, Adams, Baral, & Kennedy, 2014; Miller et al., 2011; Murray et al., 2009; Musumari et al., 2013; Whittle et al., 2015), one prospective cohort study (Shannon et al., 2011), and one cross-sectional study (Peretti-Watel et al., 2006) were included. While the majority of the qualitative studies used in-depth interviews to highlight main themes among study populations, one study employed free listing techniques and key informant interviews to elicit results (Murray et al., 2009). In the prospective cohort study, loss to follow-up was not reported. People living with HIV who inject drugs were surveyed at baseline and every 6 months for 2 years; data were drawn from 470 participants who completed at least one follow-up survey (1310 total observations), and the analysis was adjusted for potential confounders (Shannon et al., 2011). A random stratified sample of people living with HIV at hospitals and outpatient clinics were recruited for the cross-sectional study (Peretti-Watel et al., 2006). Findings are summarized narratively by major themes below.

Figure 1. PRISMA flow diagram of the search and screening processes of a systematic review.

Table 1. Descriptions of included studies.

Author, year	Location/setting and population	Study objective	Study design and methods	Sample size and food security status	Outcomes
Fielding-Miller et al., 2014	Location: Four sites in Swaziland: a rural border town, a peri-urban community, and two side in the urban Manzini-Mbabane corridor Population: HIV-positive female sex workers	Self-identified female sex workers were recruited for participation in this study through purposive and snowball sampling in order to examine the relationship between food insecurity and risky sexual practices.	Study design: Qualitative Participant selection: Purposive Sampling Methods: Two semi-structured in-depth interviews were conducted with each participant.	Sample size: Total: 20 Food security status: Not specified	Relevant themes: Hunger and poverty: Hunger was often the reason participants first engage in sex work. Insufficient resources for themselves and their families, particularly in relation to adequate food remained a concern for participants. The use of food to manage HIV status: Participants emphasized the role and importance of food in managing HIV leading some to continually engage in sex work. Others cited the necessity of adequate food in order to take ART medications. Social stigma: Stigma associated with both sex work and HIV resulted in a continued dependence on sex work to address hunger.
Miller et al., 2011	Location: Mbarara, Uganda Population: HIV-positive adults both on and off ARV treatment	Interviews were conducted with people living with HIV attending an ARV distribution clinic in Mbarara, Uganda, in order to examine the relationship between food insecurity and risky sexual behaviors.	Study design: Qualitative Participant selection: Purposive Sampling Methods: Semi-structured in-depth interviews	Sample Size: Total: 41 Women: 25 Men: 16 Food security status: Not specified	Relevant themes: Food insecurity contributes to women's decision to engage in transactional sex The vulnerability of women whose husbands died of HIV: Widows are more likely to be food insecure and exchange sex for food, and men frequently approach these women for sex Food insecurity and control over sexual relationships: Dependence on male sexual partners for food and resources made condom negotiation extremely difficult for many women Staying in high-risk relationships for food: Lack of resources and dependence on male partners forced some women to remain in abusive relationships in order to access food.
Murray et al., 2009	Location: Urban Zambia Population: HIV-positive pregnant and breastfeeding women attending an HIV clinic	Conducted among women involved in the Zambia Exclusive Breastfeeding Study (ZEBS), this cohort study examined mother to child transmission of HIV through breast milk to uncover barriers to ART adherence.	Study design: Qualitative Participant selection: Purposive Sampling Methods: Free listing and key informant interviews	Sample size: Total: 80 Free listing: 47 Key informant interviews: 33 Food security status: Not specified	Relevant themes: Food insecurity acts as a barrier to starting ART Lack of consistent access to food was identified as a barrier to continued ART adherence
Musumari et al., 2014	Location: Kinshasa, Democratic Republic of Congo Population: HIV-positive adults on ART, those on ART after an interruption, and those who had been off ART for 3+ consecutive months.	Interviews were conducted with people living with HIV of different treatment profiles to identify barriers and facilitators of ART adherence in Kinshasa, DRC.	Study design: Qualitative Participant selection: Purposive Sampling Methods: Semi-structured in-depth interviews	Sample Size: Total: 38 Women: 24 Men: 14 Food security status: 42.1 % reported having irregular meals (≤1 meal/day)	Relevant themes: Food insecurity as a barrier to ART adherence: Food insecurity was the most common theme as a barrier to daily ART adherence and was driving force behind treatment interruption across the different treatment profiles. Participants described an increased incidence of negative side-effects of ART when taken without food. Adherence was also affected by the belief of some participants that the medication is not effective or can be harmful when taken on an empty stomach.
Peretti-Watel et al., 2006	Location: France Population: HIV-infected patients at hospital departments delivering HIV care	A national cross-sectional study was conducted among a random stratified sample of HIV patients to investigate the correlates of non-adherence to ART.	Study design: Cross-sectional Participant selection: Random stratified sampling	Sample size: Total: 1809 Women: 408 Men: 1401 Food security status: 17.2% of heterosexual women living with HIV (70/408) experienced food privation in the household in the prior 4 weeks	ART adherence: Among heterosexual women living with HIV who reported food privation in the household in the prior 4 weeks, 56% were non-adherent to ART and 44% were adherent (p = 0.04); among those who did not report food privation, 43% were non-adherent to ART and 57% were adherent. (OR = 1.30, 95%CI: 0.79–2.14)
Shannon et al., 2011	Location: Vancouver, Canada Population: HIV-positive people who inject drugs both on and not on HAART	HIV-positive people who inject drugs in inner city Vancouver were recruited through snowball sampling and street outreach to participate in a prospective observational cohort study, also known as the AIDS Care Cohort to evaluate Exposure to Survival Services, 2005–2008	Study design: Prospective Cohort Participant selection: Non-random Follow-up time: Observations taken every 6 months between Dec 2005 and May 2009	Sample size: Total: 470 Women: 183 Men: 287 Food security status: 39% of those severely food insecure were women and 38% of those not severely food insecure were women (No difference between men and women p=0.789)	Behavioral outcomes: Unprotected sex (vaginal and/or anal) and severe food insecurity Unadjusted OR: 2.98 (1.56–5.05) p < 0.10 Adjusted OR: 2.68 (1.49–4.82) p < 0.10 *Adjusted for age, sex/gender, married/cohabitating partner, binge drug use, homelessness, and HAART use Unprotected sex (vaginal and/or anal) among females Unadjusted OR: 1.54 (1.00–2.37) p < 0.10 Adjusted OR: 1.5 (1.02–2.22) p < .10 *Adjusted for age, sex/gender, married/cohabitating partner, binge drug use, homelessness, and HAART use
Whittle et al., 2015	Location: San Francisco and Alameda County, CA, United States Population: Low-income people living with HIV	Semi-structured in-depth interviews were conducted with participants recruited from a program providing 3 meals a day with the aim of documenting experiences with food insecurity and risky sexual practices.	Study design: Qualitative Participant selection: Purposive Sampling Methods: Semi-structured in-depth interviews	Sample Size: Total: 34 Women: 6 Men: 28 Food security status: Not specified but all participants were receiving daily food assistance	Relevant themes: Transactional sex as coping strategy when facing food insecurity: Absolute lack of food and hunger motivated engagement in transactional sex Food insecurity leading to unprotected sex: Obtaining food and resources were given precedence over using condoms when experiencing food insecurity

Food insecurity and sex as a means of economic survival through transactional sex and sex work

In three studies, food insecurity directly contributed to women's decisions to have sex in exchange for resources and food, either through informal transactional sex or through formal/commercial sex work (Fielding-Miller et al., 2014; Miller et al., 2011; Whittle et al., 2015). In a qualitative study from Uganda among women living with HIV, many women relied on transactional sex as a survival tool to buy food, but this was not considered “prostitution” (Miller et al., 2011). Participants in the study (6 men and 26 women living with HIV) distinguished women who had sex for food from “prostitutes,” with the former viewed as women engaging in longer term, informal transactional relationships and the latter generally viewed as women in bars and urban settings who sought out one-time customers. Women reported receiving multiple solicitations from men for sex in exchange for food or resources. Women, particularly those with children, whose husbands or primary partners had died were especially targeted by men for transactional sex. Although knowledge of a woman's HIV status did not always deter men from offering food in exchange for sex, other women reported feeling hesitant to disclose their HIV-positive status to a new sexual partner for fear they would be rejected and therefore unable to provide for themselves or their children.

In a qualitative study with 20 female sex workers living with HIV in Swaziland, hunger was frequently cited as both the main reason for initiating sex work and as the primary motivation to continue to sell sex (Fielding-Miller et al., 2014). Similar to the findings from Uganda, single women with children in Swaziland were especially vulnerable to food shortages and, as a result, engaged in transactional sex to feed and support their families. Many women described their introduction into sex work as precipitated by the death or departure of a partner. Though women originally entered sex work to cope with hunger, few women in the study identified themselves as currently food secure. Both HIV and sex work are stigmatized in Swaziland, and many women worried they would lose support, including food, from family and friends if they disclosed either their HIV status or their occupation. The authors described the women's situations as a “cycle of food insecurity, HIV risk, and social marginalization.”

In a third qualitative study conducted in the San Francisco Bay area of the United States among low-income people living with HIV, participants described how women and men who have sex with men resorted to transactional sex in times of food insecurity (Whittle et al., 2015). Women in this population explained how food insecurity and severe hunger served as motivators to sell sex. High rent in urban areas and homelessness was also found to be a facilitator of transactional sex for food and shelter. A notable difference between these transactional relationships and those found in resource-poor countries included in this review is that they were generally not long-term arrangements.

Food insecurity and lack of control in sexual relationships

Food insecurity was related to a lack of control in sexual relationships in four studies. In the study from Uganda described above, women living with HIV with primary partners often reported a lack of control in their sexual relationships due to their reliance on their male partners for food (Miller et al., 2011). Some women reported that their male sexual partners would deny food sources in order to control the circumstances around sex, including the timing. Many women stayed in high-risk violent or abusive relationships because of their dependence on partners for food. This was particularly true for women with children and women living far from family who could provide assistance. This lack of control prevented women from successfully negotiating condom use; many women reported being threatened with loss of access to food or even violence if they insisted on using condoms.

In the qualitative study from Swaziland, female sex workers living with HIV also reported food insecurity as a barrier to condom use (Fielding-Miller et al., 2014). Clients were often willing to pay more for sex without a condom, which was hard to refuse in the face of hunger. Many of the women believed that they had contracted HIV through lack of condom use during sex. Similarly, in the qualitative study conducted in San Francisco, United States, participants described how food insecurity led to a decreased ability to insist on protection during sexual encounters (Whittle et al., 2015). Women also related how their experiences with transactional sex sometimes compromised their safety, as they engaged with people with whom they did not feel comfortable.

In a study among both men and women living with HIV who inject drugs in Vancouver, Canada, Shannon et al. (2011) found no difference in food insecurity by sex, but women had 1.5 times the odds of reporting unprotected sex compared to men. Controlling for sex, people who were severely food insecure had nearly three times the odds of condom-less sex compared to those who were not severely food insecure (adjusted odds ratio: 2.68, 95% confidence interval: 1.49–4.82).

Food insecurity as a barrier to ART initiation and adherence

In a national cross-sectional study in France, women living with HIV who experienced food privation in the household in the prior 4 weeks were more likely to be non-adherent to ART, but the correlation was not statistically significant (odds ratio: 1.30, 95% confidence interval: 0.79–2.14) (Peretti-Watel et al., 2006).

Researchers investigated factors related to decision-making around initiation and adherence to ART among urban women with access to free comprehensive HIV care in Zambia (Murray et al., 2009). In this qualitative study involving a sample of 47 women living with HIV, hunger emerged as a key barrier to consistently taking medication. Participants discussed the difficulty of eating sufficient quantities of food and the associated negative side effects of treatment. Many women believed that it was better not to start ART due to fears that the drugs increased appetite, and many had limited access to food. Relatedly, several women described lack of financial independence, unemployment, and dependence on a husband or partner for resources as barriers to accessing food. Hunger and lack of food also emerged as major themes for why women stopped taking ART. Women taking ART complained of hunger, articulating a need to eat more, while also vomiting the little food they could consume after taking their medication.

Similar results were reported in a qualitative study in the Democratic Republic of Congo exploring factors influencing ART adherence among HIV-positive adults with different treatment profiles (currently on ART, on ART re-initiation following an interruption, and lost to follow-up) (Musumari et al., 2013). Women in this sample cited lack of food as the most significant barrier to daily ART adherence and the main reason for treatment interruptions. Participants described increased side effects experienced when taking ART without food. Many believed that treatment was ineffective or harmful when taken without sufficient quantities of food. These intensified side effects and beliefs impeded adherence.

Though the female sex workers living with HIV in the Swaziland study did not directly address food insecurity and ART medication adherence, many discussed difficulties related to taking ART on an empty stomach (Fielding-Miller et al., 2014). Some women feared that inability to obtain food could result in deteriorating health, motivating some to continue engaging in sex work. Local clinics advised them to eat “healthy foods” to manage their HIV infection, and women often talked about “healthy food” when asked how they coped with their HIV status but had different conceptions of what that entailed.

Discussion

There is some evidence that people living with HIV who begin antiretroviral therapy without adequate nutrition have lower survival rates (de Pee & Semba, 2010; Gupta et al., 2011). For women living with HIV, particularly in poorer communities, treatment adherence is even more important since economic or social dependency may lead them to engage in high-risk activities such as transactional or commercial sex or stay in high-risk or abusive sexual relationships. Despite the extensive amount of literature published on HIV, we identified only seven studies evaluating the relationship between food insecurity and sexual risk behaviors and ART adherence among women living with HIV. Studies included populations of pregnant and breastfeeding women, sex workers, people who inject drugs, and women both on and off ART. Study design and rigor varied across the five qualitative studies, one prospective cohort study, and one cross-sectional study. Results suggest that food insecurity remains a challenge for many women living with HIV across diverse settings, resulting in risky coping strategies like transactional sex and sub-optimal ART adherence.

Findings from this review indicate that food-insecure women living with HIV often engage in transactional sex and sex work in order to obtain resources and food for themselves and their families. Our findings among women are consistent with previous research on food insecurity and sexual risk conducted among both men and women living with HIV in high-resource settings (Vogenthaler et al., 2011; Vogenthaler et al., 2013; Whittle et al., 2015). The correlation between food insecurity and HIV-risk behaviors has also been extensively studied among HIV-negative populations across various settings (Davidoff-Gore, Luke, & Wawire, 2011; Dunkle et al., 2004; Eaton et al., 2014; Greif, 2012; McCoy, Ralph, Njau, Msolla, & Padian, 2014; Oyefara, 2007; Raiford et al., 2014; Tsai et al., 2012; Wang et al., 2013; Weiser et al., 2007). In a study conducted in South Africa, economic survival strategies, which included transactional sex with non-primary partners, were associated with a 1.5-fold increased risk of HIV infection (Dunkle et al., 2004). Food insecurity has been cited as a significant motivator for women to enter sex work and as a barrier to condom use, despite high rates of HIV awareness (Fielding-Miller et al., 2014; Oyefara, 2007).

This review also found that food insecurity serves as a barrier to ART initiation and adherence. Again, our findings among women align with a larger evidence base linking food insecurity and sub-optimal adherence among both men and women living with HIV across a variety of settings (Singer et al., 2015). Several plausible mechanisms for this relationship have been suggested, including fears or experiences of increased hunger on ART, fears of intensified side effects of ART if taken without adequate food, and having to choose between buying food and accessing ART (often causing people to sell or trade ART for food or resources) and other medical care, resulting in poor treatment retention (Young et al., 2014). People living with HIV are often advised to “eat healthy foods” by health care workers and health promotion campaigns (Makoae, 2011), yet some people must choose between paying for ART or paying for food (Jones, 2011). Conversely, researchers conducting a prospective cohort study among 602 treatment-naïve people in Uganda found that initiating ART was linked with reduced odds of food insecurity, perhaps due to improved mental and physical health and increased ability to work (Palar, Wagner, Ghosh-Dastidar, & Mugyenyi, 2012). These findings, along with substantial evidence linking food insecurity with HIV risk behaviors and ART adherence, point to the bidirectionality and syndemicity of food security and risk behaviors like non-adherence and condomless sex or transactional sex (de Waal & Whiteside, 2003; Himmelgreen et al., 2009; Palar et al., 2012; Weiser et al., 2011).

The particular risks of food insecurity and transactional sex as well as links to adherence and initiation of ART must be understood within the context of gender inequalities. Globally, women often depend on men for food and financial support. Women are less likely to have access to and control over land (for food production and income generation), employment, money, and other economic resources (e.g. inheritance) (Knox et al., 2007; Swaminathan et al., 2007). In many settings, women also have primary roles in procuring/buying/preparing food and feeding children, other family members, and community members in ways that might shape the impacts of food insecurity on their ability to adhere to ART or their reasons for engaging in sexual risk behaviors. Similarly, in many settings where clean and safe drinking water is not available, it is the primary responsibility of women to travel long distances in search of clean water and to transport clean water, requiring them to carry heavy loads and walk long distances (Buor, 2004; Ray, 2007; United Nations Department of Economic and Social Affairs, 2015; West, Hirsch, & El-Sadr, 2013). The role of women in securing food, water, preparing meals, and feeding families also places a heavy burden on their time as shown by time use surveys (United Nations Department of Economic and Social Affairs, 2015). This “time poverty” has implications for their ability to take care of their own health and well-being. For women living with HIV, the stigma and discrimination related to HIV status pose additional burdens as they face the risk of being abandoned, deprived of family property, land, inheritance, and housing in ways that increase their food insecurity (Knox et al., 2007; Swaminathan et al., 2007).

There are limitations to the conclusions we can draw from this review. The small number of studies identified that met our criteria limits our understanding of the resulting themes. Due to the limited number of studies stratifying results by gender, we broadened our inclusion criteria to include any study that presented at least some data separately for women, presenting challenges in separating results by gender. Some of the included studies relied on observational data involving self-reporting of sexual risk behaviors and food security, potentially allowing for social desirability bias or recall bias. Additionally, the included studies were conducted in only seven countries, with four in sub-Saharan Africa, two in North America, and one in Europe. With this limited geographic variation, results may not be generalizable across all settings. We focused only on outcomes related to sexual behavior and ART adherence, which did not capture the full range of effects that food insecurity may have on the lives of women living with HIV. We also did not include lack of access to clean water in our definition of food insecurity, though women often bear a disproportionate burden in obtaining water when it is scarce and lack of clean water has severe implications for the health of women living with HIV in particular due to their risk of infections (Amin, 2015; Levy, Webb, & Sellen, 2010; West et al., 2013). There may be other important health consequences of food insecurity for women living with HIV that merit further research, such as maternal and infant health, stress, depression, or violence.

Conclusion

In summary, current literature points to an association between food insecurity, lack of control within sexual relationships, and various behavioral outcomes related to HIV and STI risk, as well as HIV medication adherence, among women living with HIV. Our findings have the following implications for programs and policies that seek to address the needs of women living with HIV. The HIV response must create an enabling environment for women living with HIV by implementing programs that reduce hunger and food insecurity, especially for women. This includes efforts to provide information and education about growing, preparing, and eating healthy foods. Programs should address the political, social, and economic contexts in which women experience food insecurity (Aberman, Rawat, Drimie, Claros, & Kadiyala, 2014) to identify which interventions and programs are effective in order to improve women's legal and social rights and empower them economically such as cash transfers, microfinance, vocational and skills training (Himmelgreen et al., 2009), particularly those that promote personal property and inheritance rights in conjunction with food assistance and income generation (Hardee, Gay, Croce-Galis, & Peltz, 2014). The paucity of studies included in our review suggests that more research is needed that specifically addresses food security, sexual risk behavior, and ART adherence among women living with HIV. Furthermore, disaggregation of results by sex would allow future researchers to better understand gender dynamics of HIV and food insecurity.

Acknowledgments

The authors would like to acknowledge the members of the WHO guideline development group who provided input to this review.

Funding

This review was funded by the World Health Organization, Department of Reproductive Health and Research.