Childhood Sexual Trauma and Antiretroviral Therapy Adherence: A Mixed-Methods Systematic Review

Childhood sexual abuse (CSA) has been shown to be more prevalent among populations living with HIV. Antiretroviral therapy (ART) adherence is crucial for populations living with HIV as it significantly increases the likelihood of attaining and maintaining viral suppression. Previous findings on the association between CSA and ART adherence have been mixed. The current mixed-methods systematic review aimed to identify quantitative and qualitative studies from CINAHL, PsycInfo, PubMed, and Web of Science examining the relationship between CSA and ART adherence. Authors were also contacted if relevant data were unpublished. Studies had to be published from January 1, 2000 to April 1, 2019, written in English, and examined CSA as an exposure and ART adherence as an outcome. Four domains were combined: 1) childhood sexual abuse; 2) child; 3) antiretroviral; and 4) adherence. Eight quantitative and two qualitative studies were retained. The results showed that four quantitative studies found no association while the other four found factors such as timing of victimization, mental health and gender influenced the association between CSA and ART adherence. Themes emerging from the qualitative studies included use of ART evoking memories of CSA; CSA impacting mental health; and mental health treatment improving ART adherence. Mixed insights included the intricate links between CSA and ART adherence and the role of external factors on the relationship. ART adherence intervention programs may be needed for people who have experienced CSA. However, future studies are needed that will examine the association between CSA and ART adherence and include subgroup analyses.

ART adherence. Authors were also contacted if relevant data were unpublished. Studies had to be published from January 1, 2000 to April 1, 2019, written in English, and examined CSA as an exposure and ART adherence as an outcome. Four domains were combined: 1) childhood sexual abuse; 2) child; 3) antiretroviral; and 4) adherence. Eight quantitative and two qualitative studies were retained. The results showed that four quantitative studies found no association while the other four found factors such as timing of victimization, mental health and gender influenced the association between CSA and ART adherence. Themes emerging from the qualitative studies included use of ART evoking memories of CSA; CSA impacting mental health; and mental health treatment improving ART adherence. Mixed insights included the intricate links between CSA and ART adherence and the role of external factors on the relationship. ART adherence intervention programs may be needed for people who have experienced CSA. However, future studies are needed that will examine the association between CSA and ART adherence and include subgroup analyses.

The syndemic of childhood sexual abuse and other forms of trauma
Populations who experience CSA often undergo multiple forms of trauma in their lifetime, including physical abuse, emotional abuse, and intimate partner violence (Dong, Anda, Dube, Giles, & Felitti, 2003). Increased severity and frequency of experience with CSA and other forms of trauma can influence greater likelihood of poor health outcomes among those affected (Adams, Mrug, & Knight, 2018;Boudewyn & Liem, 1995;Zanarini et al., 2002). Particularly, previous research has found that exposure to adverse events such as CSA is linked to substance abuse, future chronic disease development, and the development of mental health problems such as PTSD and depression (Batchelder et al., 2017;Brown et al., 2017;Danese et al., 2009;Markowitz et al., 2011;Pretty, O'Leary, Cairney, & Wade, 2013;Spies et al., 2012). Those who have experienced CSA are also more likely to have increased hospitalizations, disease progression and mortality, especially among PLWH .

Experiencing childhood sexual abuse and risky behaviors
High rates of CSA among PLWH should be given marked consideration given that past research has shown that this population is likely to engage in risky sexual behaviors (Anaya, Swendeman, & Rotheram-Borus, 2005;Pence et al., 2012;Whetten et al., 2012). Cohen et al.'s (2000) study using 1,288 women living with HIV and 357 HIV-negative women aged 18 and older found that CSA was associated with several risky sexual behaviors, including having greater than 10 lifetime male sexual partners, having had sexual intercourse with a male or female partner at risk for HIV acquisition, and having had sex for money, drugs, or shelter. (Cohen et al., 2000). Black and White populations living with HIV, and PLWH who abuse substances are known to most frequently report CSA (Anaya et al., 2005;Kidman & Violari, 2018;Markowitz et al., 2011;Paxton, Myers, Hall, & Javanbakht, 2004). Several studies have also found that CSA is linked with increased risk for engaging in unprotected sex, adult sexual revictimization, and having greater number of sex partners among heterosexual men and men who have sex with men (MSM) living with HIV (Markowitz et al., 2011;O'Leary, Purcell, Remien, & Gomez, 2003;Pantalone et al., 2015;Whetten et al., 2012).

The importance of antiretroviral therapy adherence
Higher likelihood of engagement in risky sexual behaviors, substance abuse, and mental health problems among PLWH are of great concern given their association with increased likelihood of onward HIV transmission (Colfax et al., 2002;Esser, Krotzek, Dirks, Scherbaum, & Schadendorf, 2017;Fox et al., 2009;Sikkema et al., 2010). Achieving and maintaining HIV viral suppression is, thus, particularly imperative for high risk groups such as PLWH who have experienced CSA. Using antiretroviral therapy (ART) and adhering to prescribed ART regimens is crucial in attaining and maintaining viral suppression, to prevent HIV transmission, and to prevent the advancement of HIV infection (Bartlett, 2002;Bunnell et al., 2006). PLWH taking ART are recommended to adhere to prescribed ARTs regimens at least 95% of the time in order to fully benefit from treatment (Bartlett, 2002).

Childhood sexual abuse and antiretroviral therapy adherence
Past research suggests that gender, experiencing more CSA events or lifetime traumas, age at CSA, and substance abuse are key influences for poor HIV medication adherence among people living with HIV who have experienced CSA (Markowitz et al., 2011;M. Mugavero et al., 2006;Wilson, Sikkema, & Ranby, 2014). CSA may be intrinsically linked to poor HIV medication adherence due to its association with low social support and trauma symptoms like stress and negative coping (Johnson, Heckman, Hansen, Kochman, & Sikkema, 2009;McCall et al., 2018;Meade, Hansen, Kochman, & Sikkema, 2009;Wilson et al., 2014). Anxiety, depression, and PTSD are also common trauma-related consequences of CSA that have been found to be individually associated with poor ART adherence among PLWH (Sauceda, Wiebe, & Simoni, 2016;Willie et al., 2016).

Study rationale and aims
Previous research has had mixed results on how CSA may be indirectly or directly linked to poor ART adherence across various populations of PLWH (Kidman & Violari, 2018;Liu et al., 2006;Mugavero et al., 2006;Wilson et al., 2014). Kidman and Violari found that childhood adversity was associated with partner violence victimization, which was linked to poor medication adherence to antiretroviral therapy. CSA severity was associated with substance use, which was related to lower levels of adherence. However, a direct link between severity of CSA and poorer ART adherence was not found (Liu et al., 2006). Lifetime traumatic events, including CSA, were linked to poorer ART adherence (Mugavero et al., 2006) and within the context of CSA, gender may play a moderating role in the influence of psychosocial factors on ART adherence (Wilson et al., 2014). Past reviews have largely focused on assessing the relationship between CSA or childhood trauma in relation to outcomes such as HIV risk (Lloyd & Operario, 2012;McGeough & Sterzing, 2018;Thornton & Veenema, 2015). Therefore, collectively examining research findings on the relationship between CSA and HIV medication adherence is warranted. This paper outlines results from a systematic review on the association between CSA and HIV medication adherence.

Literature search
The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Oxford., 2015) were used to guide this systematic review. Quantitative and qualitative studies that provided information on the relationship between childhood sexual trauma and ART adherence were retrieved. The published peer-reviewed literature from January 1, 2000 to April 1, 2019 was searched in four databases: CINAHL, PsycInfo, PubMed, and Web of Science. Search algorithms were developed with assistance from a public health librarian (AE) and combined four domains: 1) childhood sexual abuse; 2) child; 3) antiretroviral; and 4) adherence. The search was expanded to include specific text words. The search algorithm included only English publications. Study selection and extraction of data were completed by at least two team members (MJB, ADB, MRH, AE).

Review Process
Four reviewers participated in the first and second phases of review (MJB, ADB, MRH, AE). If there was a disagreement between two reviewers in each phase, a third reviewer made the final decision.
First phase of review: Four reviewers reviewed the titles and abstracts of articles independently. All studies were retained for further review if the title and/or abstract included concepts related to "HIV", "trauma", "abuse" or "adherence".

Second phase of review:
The full texts for all review studies and primary studies that were considered relevant were retrieved. If additional data were needed that were not found in articles, the investigators were contacted to obtain these data. This method of obtaining data had a 57.1% success rate (4/7).
As shown in Figure 1, a total of 1,793 records were identified and after duplicates were removed (n=952), the titles and abstracts of 841 articles were screened. A total of 690 were excluded after title/abstract review. Therefore, the full text of 151 articles were assessed. There were 141 studies excluded, which resulted in eight (8) articles to be included in a quantitative synthesis, and two (2) for qualitative synthesis. The research team captured relevant information from the studies to be included in final tables.

Descriptive synthesis
The research team developed shell tables to include pertinent information about the studies. The PRISMA guidelines as well as the research questions were used as a guide to determine what information to include in the tables. For the quantitative studies (Table 1), information garnered included: definition of CSA, study purpose/questions, geographic location, source of sample, number of participants in the sample, methods, and main findings. For the qualitative studies (Table 2), information obtained included: research theoretical framework, study purpose/questions, geographic location, source of sample, methods and main findings. Shared or mixed insights were also obtained from quantitative and qualitative studies. Qualitative data were analyzed with a framework thematic analysis (Kimera et al., 2019).

Quality assurance of systematic review
At least two reviewers were responsible for evaluating the titles, abstracts and full texts of relevant articles. Research team meetings and group discussions were held throughout the extraction process to raise and discuss any related issues, which came up.

Types of participants
Participants included people living with HIV who were on ART. Study participants were obtained from a variety of settings: a sample of PLWH enrolled in a primary care-based alcohol intervention study in San Francisco; women from the Canadian HIV Women's Sexual Health and Reproductive Health Cohort Study and the Women's Interagency HIV Study (WIHS) in the US; PLWH from the Coping with HIV/AIDS In Tanzania (CHAT) longitudinal cohort study; PLWH from the Centro de Salud Familiar La Fe Care Center, a federally qualified health center in El Paso Texas; urban university-affiliated outpatient clinics in the Northwestern US; participants from substance use treatment centers in Massachusetts and Rhode Island; and from street outreach in East Harlem, New York.

Quantitative findings
A total of eight quantitative studies were found to meet inclusion criteria for this review. Although this review assessed studies from 2000 to 2019, the final eight quantitative studies that were eligible for inclusion were published between 2008 to 2018. Results for the eight quantitative studies can be seen in Table 1. The samples varied greatly.
Across study samples, reported CSA among participants ranged from 6.8% to 55%. No two studies used the same method to measure CSA. For HIV medication adherence, six studies used self-reported measures of adherence either over the past week, month, three months, or six months.
Findings on associations between CSA and HIV medication adherence varied across studies. Particularly, Whetten's 2013 study among 451 men and women living with HIV in Tanzania found that compared to participants who had not experienced sexual abuse before puberty, participants who had experienced sexual abuse before puberty were 78% more likely to report incomplete adherence or non-adherence to ART (RR = 1.78, 95% CI: 1.13, 2.79, p = .0120). The Markowitz et al. (2011) study among 119 men and women living with HIV with history of drug abuse found that experience of CSA both before the age of 13 and between age 13 to 16 years old was associated with poor HIV medication adherence after adjustment for confounders (β = −0.20, p = 0.04). However, experiences of CSA either before the age of 13 or between age 13 to 16 years old was not linked to poor HIV medication adherence. Kang's 2008 study using 220 men and women living with HIV who were drug users found that CSA was associated with decreased self-reported use of HIV medications after adjustment for confounders (OR=0.51; 95%CI: 0.27,0.96, p=0.04). Yet after stratifying by gender, only men were found to have a significant association between CSA and use of HIV medications (OR=0.49, p = 0.04). Preliminary analysis from Carter's large cohort study found that women who reported CSA had slightly lower unadjusted odds of optimal HIV medication adherence, but this result was not statistically significant (OR=0.933; 95% CI: 0.705-1.236, p = 0.630). Likewise, Young-Wolff's 2019 study among 566 men and women living with HIV who were at risk for alcohol abuse found no statistically significant association between CSA and HIV medication adherence (adjusted OR=0.95 ;95% CI: 0.52, 1.71, p = 0.855). Pantalone's 2015 study also found that among 166 MSM living with HIV, there was no statistically significant association between CSA and HIV medication adherence (adjusted OR= 1.11; 95% CI: 0.48, 2.56, p = 0.814).
Dale's 2014 study among a sample of 139 women living with HIV used a composite measure of any history of abuse, which included CSA, adult sexual abuse, physical abuse, or domestic violence. Overall, this study found that there was no statistically significant association between any history of abuse and HIV medication adherence. However, when considering interaction between resilience and history of sexual abuse in childhood and/or adulthood, increase in resilience among women with history of sexual abuse in childhood and/or adulthood was associated with an increase in ≥95% HIV medication adherence (OR=1.15; 95% CI 1.02-1.29). Similarly, Sauceda's 2016 study using Latino MSM living with HIV found that depression significantly mediated the relationship between CSA and poor HIV medication adherence, yet worse adherence was found in CSA participants with the lowest resilience percentile indexes. Mean difference on HIV medication adherence between CSA groups through the indirect effect of depression equaled to −6.77 (SE=3.44), −4.48 (SE=2.36) and −2.96 (SE=1.71) at the 10th, 25th, and 50th percentile of the resilience index, respectively.

Qualitative findings
Overall, only two qualitative studies met inclusion criteria for this review. One study included eight women living in Cape Town, South Africa. This study aimed to examine the lives of women who had experienced any history of sexual abuse. A key theme discovered through this study were that women felt that using HIV ART evoked memories of their experience with CSA. The other study included three participants from psychiatric care at an HIV clinic. This case study aimed to investigate the relationship between childhood trauma, PTSD, and medication adherence. Study results showed that all participants who experienced sexual abuse additionally suffered from depression and PTSD. Psychodynamic therapy sessions for the male participant led to HIV medication adherence. One of the two female participants did not adhere to HIV medications.

Integration
Data from both quantitative and qualitative studies were used to obtain mixed insights. Based on the themes from the qualitative studies and analyses in quantitative studies, mixed insights included: 1) the intricate link of the effect of CSA on taking ART as prescribed; and 2) the role that external factors such as mental health, mental health treatment, substance use, and repeat victimization may play in the relationship between CSA and adherence ( Figure 2).

Discussion
To our knowledge, this is the first study to perform a systematic review on the association between CSA and ART adherence. Previous reviews examined the association between CSA or childhood trauma in relation to outcomes such as HIV risk. Overall, we found that data allowing the estimation of the association between CSA and ART adherence were sparse. This lack of research could be due to the sensitive nature of CSA and challenges in accurately measuring ART adherence. We propose that more studies address this relationship acknowledging the sensitive nature of CSA and relying less on self-report of ART adherence with the use of biomarkers.
The majority of quantitative studies included were cross-sectional studies and only three were cohort studies. Only two qualitative studies were found with very small sample sizes (N=8 and N=3). Therefore, this systematic review was most akin to the mixed-methods explanatory sequential design with an emphasis on the quantitative data (Figure 2). For a summary of the critical findings and implications, see Table 3 and Table 4, respectively.
Although majority of the cross-sectional studies found no statistically significant association between CSA and ART adherence, one cross-sectional study examining sex as a moderator of the relationship between CSA and ART adherence found that the relationship was statistically significant among men but not among women (Kang et al., 2008). This study suggests that sex differences may be present when examining the association between CSA and ART adherence and should be considered as an important factor in future studies.
Additional important factors that became evident, especially from the integrated analysis, was the potential role that external factors played in the relationship between CSA and ART adherence: mental health, mental health treatment, substance use history, repeat victimization, and CSA occurring before puberty. For example, the qualitative studies found that depression and PTSD were important factors to consider in this relationship. Furthermore, women found that taking their ART medication resulted in memories of their experience with CSA. Therefore, future ART adherence interventions may need to consider improving coping mechanisms, especially for women who were exposed to CSA. Finally, an additional interesting finding was that repeat sexual abuse before age 13 and between ages 13 and 16 was found to be predictive of poorer medication adherence among women and men living with HIV. Another study, however, found that CSA before puberty was also associated with lower ART adherence. These findings suggest that repeat sexual abuse revictimization and timing of CSA may play an important role in lowering ART adherence and should also be a focal point in ART adherence interventions.

Limitations
There are some limitations to consider when interpreting the study's findings. Only eight quantitative and two qualitative studies were included in the study. The research team also endeavored to contact investigators with a 57% success rate with 43% not responding to requests or indicating that there was no time available to conduct additional analyses. Nevertheless, this success rate was better than previous research with a response rate of 25% (Kimmes et al., 2019).

Methodological Limitations.
We were unable to conduct a meta-analysis due to the diversity and stark differences in the studies, especially in the operationalization of key variables such as CSA and ART adherence. For example, CSA operationalizations ranged from using sexual-abuse related items on an ACE questionnaire of CSA experiences before age 18 to reports of experiencing sexual abuse at age 12 or younger. Based on our findings, we suggest operationalizing CSA based on related experiences before age 18: someone at least 5 years older touching the respondent sexually or forcing the respondent to touch them sexually; and/or being forced to have sex (including oral sex, anal or vaginal sexual intercourse). Diversity was also seen in the operationalization of ART adherence with operationalizations ranging from self-report of ART adherence in the past month using the visual analog scale to self-report of ART adherence in the past week. The percentage of optimal adherence also varied from >90% to >95%. These findings suggest that standardized measures of CSA and ART adherence are needed and when available should be used in order to facilitate better comparisons across studies. Indeed, Mathews and Collin-Vézina (2019) proposed a conceptual model of CSA to distinguish CSA from other concepts. One quantitative study was also conducted in Canada. Comparisons across countries due to cultural differences, if done, should be done with caution (Winzer, Krahé, & Guest, 2019). Diversity was also present in the samples of the studies with regards to age, gender, sexual orientation, and race/ethnicity.

Conclusions
The small number of relevant studies indicate that more research examining the impact of sexual trauma on ART adherence is needed. For example, there are no studies examining the impact of CSA on HIV treatment outcomes such as ART adherence or viral suppression among older adults. Filling this gap of the literature is crucial as people are living longer with HIV and CSA is a common traumatic event among people living with HIV. Kimera et al. suggests that additional studies examining the impact of CSA on health behaviors are needed so as to develop the necessary interventions to improve healthy behaviors and medication adherence among youths living with HIV (Kimera et al., 2019). Whetten et al. states that additional research into trauma and its related risk factors and consequences is warranted, especially with respect to HIV risk behaviors and ART adherence (Whetten, Whetten, Ostermann, & Itemba, 2008). The Living in the Face of Trauma (LIFT) intervention by Sikkema and colleagues (2013), which was found to reduce unprotected sexual intercourse among participants living with HIV with CSA histories, may provide some intervention components for future consideration to improve ART adherence. These may include: using a group-based approach, helping participants to identify and address stressors related to living with HIV and their trauma history; and taking part in related discussions and skill building exercises (Conduent, 2020). We also recommend future studies on CSA and ART adherence conduct subgroup analyses, for example, by gender and mental health status. In addition, due to the lack of studies among older adults and that this population continues to grow, we argue that these studies are also warranted for adults aging with HIV/AIDS.  Mixed-methods systematic review, analysis and integration. NSA: No statistically significant association

Critical Findings
Having a CSA history may be associated with lower ART adherence.
The relationship between CSA and ART adherence may be influenced by factors such as gender, mental health, substance use, and mental health treatment.
Taking ART may be linked to memories of CSA, which may lower adherence to ART.
More studies (quantitative and qualitative) examining the relationship between CSA and ART adherence, including subgroup analyses, are needed.

Implications
ART adherence intervention programs may be needed for people who have experienced CSA.
Targeted ART adherence interventions addressing CSA should consider gender differences, and presence of mental health and substance use disorders, and treatment.
ART adherence for people who have experienced CSA may need to address coping skills and strategies specific to CSA experience. Quantitative and qualitative studies and related subgroup analyses will help to determine the intricate links of the relationship between CSA and ART adherence and how this relationship may differ by specific groups. This approach will help to determine specific target groups for interventions.