A scoping review of the associations between mental health and factors related to HIV acquisition and disease progression in conflict-affected populations

A scoping review of the associations between mental health and factors related to HIV acquisition... The association between poor mental health and factors related to HIV acquisition and disease progression (also referred to as HIV-related factors) may be stronger among conflict-affected populations given elevated rates of mental health disorders. We conducted a scoping review of the literature to identify evidence-based associations between mental health (depression, anxiety, and post-traumatic stress disorder [PTSD]) and factors related to HIV acquisition and progression in conflict-affected populations. Five electronic databases were searched on October 10, 2014 and updated on March 7, 2017 to identify peer-reviewed publications presenting primary data from January 1, 1994 to March 7, 2017. Articles were included if: 1) depression, anxiety, and/or PTSD was assessed using a validated scale, 2) HIV or HIV-related factors were a primary focus, 3) quantitative associations between depression/anxiety/PTSD and HIV or HIV-related factors were assessed, and 4) the study population was conflict-affected and from a conflict-affected setting. Of 714 citations identified, 33 articles covering 110,818 participants were included. Most were from sub-Saharan Africa (n = 25), five were from the USA, and one each was from the Middle East, Europe, and Latin America. There were 23 cross-sectional, 3 time-series, and 7 cohort studies. The search identified that mental health has been quantitatively associated with the following categories of HIV-related factors in conflict-affected populations: markers of HIV risk, HIV-related health status, sexual risk behaviors, and HIV risk exposures (i.e. sexual violence). Further, findings suggest that symptoms of poor mental health are associated with sexual risk behaviors and HIV markers, while HIV risk exposures and health status are associated with symptoms of poor mental health. Results suggest a role for greater integration and referrals across HIV and mental health programs for conflict-affected populations. Keywords: Mental health, HIV, Conflict settings, Depression, Anxiety, PTSD Background and socially increases risk for neuropsychiatric conditions The relationship between mental health and HIV acqui- [7]. In this paper we consider a broad range of factors as- sition and disease progression (also referred to as HIV- sociated with HIV acquisition and disease progression, related factors) is bi-directional. Having symptoms of such as markers of HIV risk, HIV-related health status, post-traumatic stress disorder (PTSD), depression, and/ sexual risk behaviors, and other potential HIV risk expo- or anxiety has been linked to HIV risk factors in various sures not under individual control (i.e. sexual violence). populations both prospectively [1] and cross-sectionally As conflict-affected populations often have elevated [2–6]. Being HIV-positive physiologically, psychologically, rates of PTSD, depression, and anxiety [8–12], the associ- ation between poor mental health and risk for HIV acqui- sition and disease progression may be stronger among * Correspondence: koeglere@health.missouri.edu these populations. Conflict can shape population move- Department of Health Sciences, University of Missouri, 512 Clark Hall, ments, opportunities for sexual partnering, and mortality Columbia, MO 65211, USA Department of International Health, Johns Hopkins Bloomberg School of patterns in ways that might increase or decrease HIV Public Health, 615 N. Wolfe Street Room E5547 Baltimore, Baltimore, MD prevalence [13, 14]. Epidemiological evidence suggests 21205, USA © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Koegler and Kennedy Conflict and Health (2018) 12:20 Page 2 of 22 elevated HIV prevalence in the five years prior to conflict, measures. This review was conducted following PRISMA but an overall decrease in HIV prevalence during and just guidelines [20]. after conflict [15, 16]. However, vulnerable populations may remain at elevated risk for HIV acquisition during Definition of terms political and socioeconomic instability [14, 17]. A seminal We sought to illuminate the ways a broad range of factors paper discussing population vulnerability to HIV trans- related to HIV acquisition and disease progression have mission in conflict-affected settings discusses health fac- been examined in relationship to mental health. There- tors but does not detail the ways poor mental health can fore, we defined factors related to HIV acquisition and impact population vulnerability to HIV [13]. Poor mental disease progression to include factors such as: markers health, conflict, and being HIV positive are independently of HIV risk (i.e. sexually transmitted infections (STIs)); related to morbidity and mortality. Co-occurrence of these HIV-related health status (i.e. HIV seropositive status, factors can contribute to increased vulnerability to mor- CD4 count); sexual risk behaviors (e.g. unprotected sex, bidity and mortality. multiple sexual partners, exchange sex, etc.); and other Other reviews that have examined associations be- potential HIV exposures not under individual control (i.e. tween mental health and HIV risk behaviors or care and sexual assault). A range of factors related to HIV were in- treatment programs have focused on migrant popula- cluded in order to provide a comprehensive understand- tions [18] and populations from developing countries ing of how researchers have quantitatively examined the [18, 19]. It is yet unknown how the vulnerabilities of relationships between specific mental health disorders and poor mental health and factors related to HIV acquisi- HIV-related factors. tion and disease progression operate in conflict-affected Conflict-affected settings were defined according to populations. Understanding how mental health is associ- UNESCO as areas with ‘explosive’ (over 200 battle-related ated with HIV acquisition and disease progression in deaths in a year) or ‘protracted’ (over 1000 battle-related conflict-affected populations can inform program and deaths over ten years) events [21]. Both active conflict and policy work in these settings. The aim of our study was post-conflict settings were included. Since not all popula- to conduct a scoping review of the literature to identify tions in conflict-affected countries are directly affected by evidence-based associations between common mental conflict, the study population had to be affected by con- health conditions (depression, anxiety, and PTSD) and flict and described as such by the article’sauthors.All factors related to HIV acquisition and disease progression combat-affected populations were considered for inclu- in conflict-affected populations. We sought to understand sion, both combatants and civilians. Conflict-affected the bi-directional associations between these mental populations across the economic spectrum were consid- health conditions and various measures of HIV-related ered for inclusion to examine the relationship between factors, and to examine the strength and directionality of mental health and HIV-related factors in a variety of associations to offer suggested directions for future re- economic situations. search, policy, and interventions. Search strategy Methods Five electronic databases (PubMed, PsycINFO, SCOPUS, Peer-reviewed publications that presented primary data CINAHL, and EMBASE) were searched first on October from January 1, 1994 to March 7, 2017 were included in 10, 2014 and updated on March 7, 2017. Search terms this review if they met the following inclusion criteria: 1) included combinations of terms for mental health, HIV one or more of three common mental health conditions risk, and conflict-affected settings (Additional file 1). We (depression, anxiety, PTSD) was a primary or substantive also searched reference lists of included articles and focus of the article and was assessed using a validated hand searched the table of contents of Conflict and scale, 2) HIV serostatus or factors related to HIV acquisi- Health and Medicine Conflict and Survival. Only articles tion and disease progression (defined below) was a pri- with an abstract in English were screened. mary or substantive focus of the article, 3) the quantitative relationship between HIV or HIV-related factors and the Data extraction and management mental health condition(s) was discussed, and 4) the study Articles were screened and data extracted by one reviewer reported that participants were conflict-affected and from (EK), with uncertainty resolved through discussion with a a conflict-affected setting. All age groups were included in second reviewer (CK). A third reviewer verified all data this review. All study designs were considered as long as presented in Tables 1 and 2. First, titles and abstracts iden- the four inclusion criteria were met. Articles were ex- tified through the search strategy were screened. Full text cluded if they measured HIV-related factors on a war articles were obtained for all selected abstracts. An eligibil- events scale but did not present data for the relationship ity form was completed to determine final study selection. between the HIV-related factor alone and mental health Data were extracted using a standardized data extraction Koegler and Kennedy Conflict and Health (2018) 12:20 Page 3 of 22 Table 1 Description of included studies Primary author Country Study design Sampling strategy Sample size and Participant characteristics and Year participation rate Mental health and HIV serostatus/HIV-related outcomes Adedimeji et Rwanda Cross-sectional Non random selection N = 928 Women over age 15 who experienced the al., 2015 Baseline data from 2005 RWISA HIV+ and HIV- women approximately 50% 99% of N = 936 included 1994 genocide, 76% HIV+ prospective cohort of whom experienced rape during the 20.5% < 30 years genocide 48.4% 30–40 years 31.1% 40+ years 100% female Adler et al., USA Time-series Non random selection N = 647 Active duty USA soldiers in a brigade combat 2011 Time 1: 4 months after return from Part of a larger study on post deployment 39% of N = 1651 included team who had returned from a 15-month deployment transition who completed both deployment in Iraq Time 2: 4 months later assessments Age not reported 96% male 4% female Kinyanda et Uganda Cross-sectional Random selection N = 1560 Vulnerable (widows, orphans, single mothers) al., 2012 Nested in study on HIV-related psychiatric Multistage sampling to include vulnerable 98.5% of N = 1584 included and non-vulnerable individuals in a war- & psychosocial vulnerabilities in war- and non-vulnerable individuals who completed the affected community affected community interview Aged 15 years and older 56% were aged between 18 and 44 years 43% male 57% female Kinyanda et Uganda Cross-sectional Random sampling N = 1110 Vulnerable (widowed, divorced, orphan, al., 2016 Nested in study addressing HIV-related Multistage sampling for representative 71.2% with complete data suffered torture, mental illness, etc.) and non- psychiatric and psychosocial vulnerabilities sample of vulnerable and non-vulnerable included of N = 1560 vulnerable individuals in a war-affected in the war-affected community individuals community Aged 15 years and older 56% were aged between 18 and 44 years 43% male 57% female Malamba et Uganda Cross-sectional Random selection N = 2388 Conflict affected individuals aged 13–49 al., 2016 Baseline data from a longitudinal cohort Two-stage stratified sampling for 97.5% who had HIV results 29.1% 13–19 years study to determine HIV prevalence and risk representative sample included of N = 2449 20.2% 20–24 years factors to inform program development consenting individuals 19.6% 25–29 years 12.4% 30–34 years 18.6% 35+ years 40% male 60% female Svetlicky et al., Lebanon Cross-sectional Non random selection N = 180 Mean age = 29.95 years (SD = 5.82; range = 20 2010 Collected 6 months post-conflict, collected Combat reserve soldiers who sought 65.7% of N = 274 included to 54 years). for 4 months. treatment in the Combat Reaction Unit in who completed 100% male the wake of the Second Lebanon War questionnaires Most were Israeli-born (82.8%) Talbot et al., Rwanda Time-series Random selection N = 120 94% were orphaned from the genocide 2013 Collected at baseline, 5, 9, and 12 months Orphans selected via random number 95% of N = 120 completed Mean age = 18 years (range 15–25) generation from a list of all eligible orphans all 4 assessments; all Male 47% enrolled in program participants were included Female 53% in analysis Koegler and Kennedy Conflict and Health (2018) 12:20 Page 4 of 22 Table 1 Description of included studies (Continued) Primary author Country Study design Sampling strategy Sample size and Participant characteristics and Year participation rate B.E. Cohen et USA Retrospective cohort Non random selection N = 71,504 Veterans of Operations Enduring and Iraqi al., 2012 From a roster of all USA veterans from 2 Separated USA veterans who were new Freedom operations users of Department of Veterans Affairs Mean age = 28.5 to 29.5 healthcare 100% female Sexual violence and mental health outcomes Amone P’Olak Uganda Cross-sectional Random selection N = 539 Aged between 18 and 25 years et al., 2013 Baseline data nested in a before and after War-affected youth who had been 83% of N = 650 who were 61% male study abducted and lived in rebel captivity for at invited to the study 39% female least 6 months 86% Acholi ethnic group Roberts et al., Uganda Cross-sectional Random selection N = 1210 Adults living in camps for internally displaced 2008 Multi-stage cluster sampling of camps, persons administrative zones, and individuals Mean age = 35.3 years 40% male 60% female 91% Acholi ethnic group Nakimuli- Uganda Time series Non random selection N = 375 Demographic data reported all patients N = 2868, Mpungu et al., Collected at baseline, 3, and 6 months Analysis included only adults with a history 59% of N = 631 included many of whom were not included in the main 2013 of war-related traumatic experiences who were present for at analysis least 2 visits Mean age adult men = 34.5 Mean age adult women = 37.3 47% male 53% female Okello et al., Uganda Case control Random selection N = 153 War affected adolescents 2007 Cross- sectional, unmatched Systematic recruitment, every 3rd name at Formerly abducted Boys mean age = 15.5 years Cases were formerly abducted youth 2 sites: a children’s support organization N =82 Girls mean age = 15.2 years Controls were non abducted youth (case) and a mixed boarding school Non-abducted N =71 Cases: 64% male; 36% female (control) Controls: 61% male; 39% female 100% of controls in secondary school, 12.2% of cases in secondary school Betancourt, Sierra Leone Prospective cohort Non random selection N = 152 Former child soldiers Agnew-Blais et Collected at baseline and time 2 Two stage method: 1) master list of youth in 60% of N = 260 interviewed Mean age = 17.4 years al., 2010 care 2) Invited youth between ages 10–18 at both times 89% male with contact information 11% female Betancourt et Sierra Leone Cross-sectional Non random selection N = 273 Former child soldiers al., 2011 Partially nested in a longitudinal study Longitudinal participants from those who N = 146 from longitudinal Mean age = 16.55 (SD 2.61) participated in one follow up visit, new study and 71% male participants recruited with NGO outreach N = 127 newly recruited for 29% female lists study (50% male, 50% female) Betancourt, Sierra Leone Prospective cohort Non random selection N = 156 Former child soldiers Borisova et al., Collected at baseline and time 2, Two stage method: 1) master list of youth 60% of N = 260 interviewed Mean age = 15.13 years 2010 approximately 2 years later in care 2) Youth aged 10–18 who did not at both times 88% male have a severe disability participated 12% female Sierra Leone Prospective cohort Non random selection N = 260 Former child soldiers Koegler and Kennedy Conflict and Health (2018) 12:20 Page 5 of 22 Table 1 Description of included studies (Continued) Primary author Country Study design Sampling strategy Sample size and Participant characteristics and Year participation rate Betancourt, Collected at baseline (2002), time 2 (2004), Sample from a master list of youth assisted 56.5% (N = 147) assessed at Mean age at time 1 = 15.13 (SD = 2.22) Brennan et al., and time 3 (2008) by program. Youth aged 10–17 with time 2 89% male 2010 contact information invited to participate. 68.8% (N = 179) assessed at 11% female time 3 Johnson et al., Liberia Cross-sectional Random selection N = 1666 Adults in Liberia; 1/3 were former combatants 2008 Population based multi stage random 98.2% of N =1696 Mean age = 41 years cluster of households attempted interviews 47.2% male 52.8% female Johnson et al., Democratic Cross-sectional Non random selection N = 998 Adults in conflict-affected provinces and 2010 Republic of Accessible population based cluster (some 98.9% of N =1005 districts Congo originally selected villages were inaccessible households surveyed Mean age = 40.1 years due to weather and security concerns) 40.6% male 59.4% female Johnson et al., Kenya Cross-sectional Random sampling N = 916 Adults in Kenya 2014 Systematic sampling of 90 villages and 10 95.8% of N = 956 households Mean age = 37.3 years households to assess election-related samples 40% male violence 60% female Cardozo et al., Kosovo Cross-sectional Random selection N = 1358 Kosovar ethnic Albanians aged 15+ years 2000 Two-stage cluster sampling Only women included in 45.3% 15–34 relevant analysis, N = 825 34.1% 35–54 10.9% 55–64 9.7% 65+ 37.7% male 62.3% female Sabin et al., Guatemalan Cross-sectional Non random selection N = 170 Adults and children in Mayan refugee camps 2003 refugees Convenience sample of 5 camps; all 93% of N = 183 households Mean age = 37.9 years living in households sampled in 4 camps, every 3rd 42% male Mexico house in 1 camp 58% female Wolfe et al., USA Retrospective cohort Non random selection N = 160 Returned veterans of the Persian Gulf War 1998 Nested in longitudinal study. Baseline Included women who completed the 66.7% of N = 240 women Mean age = 28.2 years (SD = 6.8) within 5 days of return from deployment, mailed sexual harassment questionnaire assessed at baseline 100% female time 2 18–24 months later. Washington et USA Cross-sectional Random selection N = 3598 Veterans who had been called to duty al., 2013 Pertinent result presented as case (PTSD) Population-based stratified sample 99.6% of N =3611 Mean age = 46.8 (SD = 17.3) for PTSD positive control (no PTSD) Included those who completed the PTSD and 57.4 (SD = 17.0) for PTSD negative screener women 100% female Kang et al., USA Case control Random selection N = 11,441 Gulf War veterans 2005 Nested data from a population based Stratified sample to include each subgroup 76.3% of N = 15,000 sampled Mean age: survey of military personnel Females: with PTSD = 39.1; without PTSD = 38.1 Cases: PTSD Males: with PTSD 40.4; without PTSD 39.6 Controls: did not meet criteria for PTSD 81.4% male 18.6% female HIV acquisition/disease progression and mental health outcomes Koegler and Kennedy Conflict and Health (2018) 12:20 Page 6 of 22 Table 1 Description of included studies (Continued) Primary author Country Study design Sampling strategy Sample size and Participant characteristics and Year participation rate Epino et al., Rwanda Cross-sectional Non random selection N = 610 HIV-positive adults who initiated lifelong ART 2012 From a prospective cohort Patients from clinics Mean age = 38 (SD = 10) 38% male 62% female Mean CD4 count =214 (SD = 92) Mugisha, Uganda Cross-sectional Random selection N = 2361 Adult residents of 3 of the most war affected Muyinda, Baseline data from a project delivering a Two-stage cluster sample stratified at the 98% with complete data of districts Wandiembe et kinship intervention for post-conflict sub-county N = 2406 23.5% 18–24 years al., 2015 mental health 27.3% 25–34 years 20.8% 35–44 years 28.5% 45–54 years 37.5% male 62.5% female Mugisha, Uganda Cross-sectional Random selection N = 2361 Adult residents of 3 of the most war affected Muyinda, Nested in project delivering a kinship Multistage sampling for a representative 98% who had complete districts Malamba et intervention for post-conflict mental health sample from 3 districts data included of N = 2406 23.8% 18–24 years al., 2015 27.1% 25–34 years 20.7% 35–44 years 28.4% 45+ years 37.5% male 62.5% female Muldoon et Uganda Cross-sectional Non random selection N = 129 Formerly abducted by the Lords Resistance al., 2014 From a larger community-based study of Recruited through peer/sex worker led Army sex workers outreach in bars and hotels, and Median age = 22 years (IQR:20–26) community-led outreach to former IDP 100% female camps 96.1% from Acholi tribe M.H. Cohen et Rwanda Cross-sectional Non random selection N = 850 HIV-positive and HIV-negative women al., 2009 Baseline data from a prospective cohort Mainly recruited by Rwandan women’s 91% of N = 936 with About half of each group experienced study associations available mental health data genocidal rape Mean age = 36.4 100% female M.H. Cohen et Rwanda Prospective cohort Non random selection N = 698 HIV-positive and HIV-negative women al., 2011 Baseline, 6, 12, and 18 months later Recruited from Rwandan women’s 74.6% of N = 936 who 50% of each group experienced genocidal associations and HIV clinics in Kigali completed baseline HTQ rape and at least 1 post-baseline Mean age = 36.7 (SD = 8.3) HTQ 100% female Other associations between mental health and HIV acquisition and disease progression Gard et al., Rwanda Cross-sectional Non random selection N = 922 HIV-positive and HIV-negative women 2013 Baseline data nested in a prospective Recruited Rwandan women’s associations 98.5% of N = 936 women 50% of each group experienced genocidal cohort study and clinical sites for HIV patients who completed the Health- rape Related Quality of Life 20.8% under 30 years measure 48.4% aged 30–40 years 30.8% over 40 years 100% female Cross-sectional Non random selection N = 315 Conflict-affected adult women Koegler and Kennedy Conflict and Health (2018) 12:20 Page 7 of 22 Table 1 Description of included studies (Continued) Primary author Country Study design Sampling strategy Sample size and Participant characteristics and Year participation rate Kohli et al., Democratic Baseline data from a randomized Included if provided family rejection 1.9% 16–19 years 2014 Republic of community trial information and had experienced at least 1 14.6% 20–24 years Congo traumatic event in the past 10 years 28.25% 25–34 years 22.54% 35–44 years 29.52% 45–60 years 3.17% over 60 years 100% female Sinayobye et Rwanda Cross-sectional Non random selection N = 710 HIV+ women over age 15, ART naïve al., 2015 Baseline data from 2005 RWISA HIV+ women, approximately 50% of whom Mean age = 34.9 ± 7.0 prospective cohort experienced rape during the genocide 100% female ART Antiretroviral therapy, HIV Human Immunodeficiency Virus, HTQ Harvard trauma questionnaire, IDP Internally displaced person, PTSD Post traumatic stress disorder, USA United States of America Koegler and Kennedy Conflict and Health (2018) 12:20 Page 8 of 22 Table 2 Study outcomes for association between mental health and HIV risk First author & Mental Mental health scales HIV risk measures Results Year health disorders Mental health and HIV serostatus/HIV-related outcomes Adedimeji et al., Depression Center for Epidemiologic Studies Depression HIV serostatus Depression (p < 0.001) but not PTSD (p = 0.06) was related 2015 Rwanda PTSD Scale (CES-D) Had sex last 6 months to HIV serostatus Harvard Trauma Questionnaire (HTQ) Condom use at least 50% of time last 6 months Depression (p = 0.002) but not PTSD (p = 0.09) was History of ever exchanging sex for cash or help related to sex in the last 6 months; women who had History of a non-HIV STI sex did not have different odds of depression scores between 16 and 26 (OR = 0.88, CI 0.64, 1.22) but had decreased odds of depression scores 27+ (OR = 0.57, CI 0.04, 0.81) and no different odds of symptomatic PTSD (OR = 0.78, CI 0.60, 1.03) Depression (p = 0.04) and PTSD (p = 0.006) were related to 50% condom use in the last 6 months; women who used condoms had greater odds of depression scores between 16 and 26 (OR = 1.84, CI 1.20, 2.82) but not scores 27+ (OR = 1.36, CI 0.87, 2.54) and decreased odds of symptomatic PTSD (OR = 0.60, CI 0.42, 0.86) Depression (p = 0.02) and PTSD (p = 0.003) were related to exchange sex; women who had exchanged sex had greater odds of depression scores between 16 and 26 (OR = 1.82, CI 1.19, 2.77) and 27+ (OR = 1.74, CI 1.10, 2.76) and greater odds of being symptomatic for PTSD (OR = 1.68, CI 1.19, 2.36) Depression (p = 0.04) but not PTSD (p = 0.74) related to history of a non-HIV STI; women with a non-HIV STI had greater odds of depression scores between 16 and 26 (OR = 2.02, CI 1.39, 3.09; AOR = 1.64, CI 1.01, 2.65) but not depression scores of 27+ (OR = 1.50, CI 0.94, 2.41; AOR = 1.11, CI 0.65, 1.89) nor symptomatic PTSD (OR = 1.07, CI 0.77, 1.50) Adler et al., PTSD PTSD Checklist (PCL) Risked STD by having unprotected sex PTSD at time 1 predicted sex without a condom four 2011 USA months later (OR = 1.57, CI 1.20, 2.04) Kinyanda et al., Depression Hopkins Symptom Checklist (HSCL-15) High risk sexual behaviors: High-risk sexual behavior was marginally related to MDD 2012 Uganda sex outside marriage; amongst males in univariate analysis (OR = 1.61, 95% CI sex in exchange for gifts; 0.99–2.62, p = 0.06) but not females (OR = 1.17, 95% CI sex in exchange for money; 0.68–2.01, p = 0.57). sex in exchange for protection; High-risk sexual behavior was related to MDD amongst sex with an older person; males (OR = 1.70, 95% CI 1.01–2.86, p = 0.05) in sex with someone known for less than a day; multivariable analysis but not females (OR = 1.03, 95% CI sex with uniformed personnel; 0.59–1.80, p = 0.91). sex with more than one partner Kinyanda et al., Depression HSCL-25 Sexual intimate partner violence (IPV) (‘force you to have Females who experienced sexual IPV had greater odds of 2016 Uganda sex when you don’t want to’) probable MDD (AOR = 4.20, CI 1.54, 11.46) Malamba et al., Depression HSCL-25 HIV serostatus Those with MDD symptoms had greater odd of testing 2016 Uganda PTSD HTQ positive for HIV (UOR = 2.70, CI 1.95, 3.75; AOR = 1.89, CI 1.28, 2.80) Koegler and Kennedy Conflict and Health (2018) 12:20 Page 9 of 22 Table 2 Study outcomes for association between mental health and HIV risk (Continued) First author & Mental Mental health scales HIV risk measures Results Year health disorders Those with PTSD symptoms had greater odds of testing positive for HIV (UOR = 1.90, CI 1.30, 2.78; AOR = 1.44, CI 1.06, 1.96) Svetlicky et al., PTSD PTSD Inventory Risky sexual activities (3 items including sex without No relationship was found between PTSD and risky sexual 2010 Lebanon protection against sexually transmitted diseases) activities Talbot et al., PTSD PCL Laboratory STI testing Rates of STI were too low to evaluate associations with 2013 Rwanda HIV risk taking behavior: PTSD make any conclusions. Exchanging sex for drugs, money, or favors; Higher PTSD symptoms correlated with increased HIV risk- Having sex with an HIV-infected or status unknown taking behavior (r = 0.24, p = 0.006) at baseline. partner; PTSD symptoms were related to baseline HIV risk (0.01, Having two or more sexual partners within the past p = 0.002) in a growth model; for each 1 point increase 3 months of trauma symptoms there was a 0.01 unit increase in baseline HIV risk B.E. Cohen et al., Depression ICD-9-CM diagnostic codes Sexually transmitted infections: cervical dysplasia; All STIs except chlamydia were associated with PTSD. 2012 USA PTSD genital herpes; Cervical dysplasia AOR = 1.86 (CI 1.61–2.16), Comorbid genital warts; Genital herpes AOR = 1.69 (CI 1.36–2.08), depression chlamydia; Genital warts AOR = 1.83 (CI 1.45–2.31), and PTSD gonorrhea; Chlamydia AOR = 1.66 (CI 0.93–2.96), trichomonas; Gonorrhea AOR = 3.12 (CI 1.51–6.44), and other STIs Trichomonas AOR = 1.60 (CI 1.08–2.39), Other STIs AOR = 1.83 (CI 1.52–2.21) All STIs were associated with depression. Cervical dysplasia AOR = 2.35 (CI 2.12–2.59), Genital herpes AOR = 2.51 (CI 2.20–2.87), Genital warts AOR = 2.44 (CI 2.09–2.86), Chlamydia AOR = 2.21 (CI 1.49–3.27), Gonorrhea AOR = 3.99 (CI 2.38–6.71), Trichomonas AOR = 2.38 (CI 1.85–3.06), Other STIs AOR = 2.21 (CI 1.95–2.53) All STIs were most strongly associated with comorbid PTSD and depression. Cervical dysplasia AOR = 2.65 (CI 2.41–2.91), Genital herpes AOR = 2.55 (CI 2.24–2.91), Genital warts AOR = 2.97 (CI 2.56–3.43), Chlamydia AOR = 2.58 (CI 1.80–3.70), Gonorrhea AOR = 4.74 (CI 2.91–7.71), Trichomonas AOR = 3.75 (CI 3.01–4.66), Other STIs AOR = 2.92 (CI 2.59–3.28) Sexual violence and mental health outcomes Amone-P’olak et Depression Acholi Psychosocial Assessment Instrument Sexual abuse measured by one item in the War Trauma Sexual abuse (β = 0.32, SE = 0.16, p < 0.001) predicted al., 2013 and anxiety (APAI) Screening scale symptoms of depression and anxiety for female but not Uganda male youths in multivariate analysis. Roberts et al., PTSD HTQ Rape or sexual abuse Those who reported rape or sexual abuse had greater 2008 Uganda odds of PTSD symptoms (AOR = 1.76, CI 1.01, 2.75) but not depression symptoms (NR) Koegler and Kennedy Conflict and Health (2018) 12:20 Page 10 of 22 Table 2 Study outcomes for association between mental health and HIV risk (Continued) First author & Mental Mental health scales HIV risk measures Results Year health disorders Nakimuli- Depression Self- reporting questionnaire (SRQ-20) Experienced sexual violence Experiencing sexual violence was significantly related to Mpungu et al., PTSD HTQ HIV serostatus PTSD symptom scores (β = 3.75, SE = 1.01, p < 0.05) but 2013 Uganda not depression symptom scores (β = 0.54, SE = 0.45). Being HIV-positive was not significantly related to depression (β = 0.51, SE = 0.43) or PTSD (β = −1.41, SE = 0.94) scores. Okello et al., Depression MINI-KID Sexual torture (undefined) Quantitative results not presented in a table, but the 2007 Uganda Anxiety Being forced to marry stated that no trauma event (including sexual torture and PTSD being forced to marry) showed any significant relationship with any diagnosis of PTSD, major depression and generalized anxiety disorder. Betancourt, Depression A measure developed by the Oxford Refugee Rape as part of Child War Trauma Questionnaire Surviving rape predicted an increase in depression over time Agnew-Blais, et and anxiety Studies Program for use among former child (b =2.58, p = 0.01) after controlling for demographic and al., 2010 Sierra soldiers includes a subscale for anxiety, war-related experiences. When perceived discrimination was Leone depression, and hostility included, the strength of the relationship between rape and depression is reduced, (b = 1.65, p = 0.08). When protective factors were added, there was no longer a relationship between rape and depression. Surviving rape was significantly associated with higher levels of anxiety (b = 5.35, p < 0.001) even after perceived discrimination and protective factors were controlled for. Betancourt et Depression HSCL-25 Rape as part of Child War Trauma Questionnaire No significant relationship between rape and al., 2011 Sierra and anxiety depression after controlling for multiple variables b =2.42 Leone (CI -0.99, 5.84). Rape was significantly related to anxiety b = 2.85 (CI 0.45, 5.26, p = 0.05). A smaller percentage of boys experienced rape (5%) compared to girls (44%), but the effect of rape on anxiety was significant among male child soldiers and not for females (b = −6.42, p = 0.05). Betancourt, Depression Oxford Refugee Studies Program measure for Rape as part of Child War Trauma Questionnaire Rape was correlated to depression symptoms (r = 0.24, Borisova, et al., and anxiety use among former child soldiers p ≤ 0.01) and anxiety symptoms (r = 0.38, p ≤ 0.001). 2010 Sierra Rape was not predictive of depression at T2, adjusting for Leone all covariates (b = 1.74, CI -0.53, 4.00). Rape was the strongest predictor of anxiety at T2 controlling for anxiety levels at T1 (b = 4.06, CI 1.49, 6.62, p < 0.05) and adjusting for all other covariates. Betancourt, Depression Oxford Refugee Studies Program measure for Rape as part of Child War Trauma Questionnaire Rape was associated with higher baseline levels of Brennan et al., and anxiety use among former child soldiers internalizing problems (depression/anxiety) (b =4.60, p < 0.05). 2010 Sierra After adjusting for all hardship and protective factors, among Leone time-invariant predictors, only being raped remained significantly related to depression/ anxiety (b =4.34, p = 0.039). Johnson et al., Depression Patient Health Questionnaire 9 Sexual violence defined as any violence, physical or Adults who experienced sexual violence were more likely 2008 Liberia PTSD PTSD Symptom Scale Interview (1 month psychological, carried out through sexual means or by to meet criteria for PTSD (69% vs. 38%, p < 0.001) and recall) targeting sexuality and included rape and attempted rape, MDD (57% vs. 37%, p = 0.002) compared to adults who molestation, sexual slavery, being forced to undress or did not experience sexual violence. Koegler and Kennedy Conflict and Health (2018) 12:20 Page 11 of 22 Table 2 Study outcomes for association between mental health and HIV risk (Continued) First author & Mental Mental health scales HIV risk measures Results Year health disorders being stripped of clothing, forced marriage, and insertion The weighted prevalence of PTSD (81% vs. 46%, p < 0.001) of foreign objects into the genital opening or anus, forcing and MDD (64% vs.42%, p = 0.003) was higher among male 2 individuals to perform sexual acts on one another or former combatants who had experienced sexual violence harm one another in a sexual manner, or mutilating a compared to those who had not. person’s genitals. The weighted prevalence of PTSD (74% vs. 44%, p = 0.005) was higher but not MDD (63% vs.55%, p = 0.51) among female former combatants who experienced sexual violence compared to those who had not. Noncombatant sexual violence was not related to MDD (32% vs. 29%, p = 0.73) nor PTSD (39% vs. 36%, p = 0.74) for men nor MDD (48% vs. 36%, p = 0.15) nor PTSD (56% vs.36%, p = 0.09) for women. Those who experienced lifetime sexual violence had 1.39 (p = 0.04) the odds of MDD and 2.67 (p < 0.001) the odds of PTSD compared to those who did not experience sexual violence. Johnson et al., Depression Patient Health Questionnaire–9 Sexual violence – defined above The prevalence of MDD was significantly higher for those 2010 PTSD PTSD Symptom Scale Interview (PSS-I) who experienced sexual violence (60.4%) compared to those Democratic who did not experience sexual violence (30.7%, p < 0.001); Republic of The prevalence of PTSD was significantly higher for those Congo who experienced sexual violence (70.2%) than those who did not experience sexual violence (40.3%, p < 0.001) . The prevalence of MDD for females who experienced conflict-related sexual violence was significantly higher (67.7%) than for those who did not experience conflict- related sexual violence (30.3, p <0.001). The prevalence of PTSD for females who experienced conflict-related sexual violence was significantly higher (75.9%) than for those who did not experience conflict- related sexual violence (44.4%, p < 0.001). There were no differences in the prevalence of MDD (47.5% vs. 36.3%, p = 0.18) or PTSD (56% vs. 41.7%, p =0.17) for men who did and did not experience conflict-related sexual violence. There were no differences in the prevalence of MDD (50.7% vs. 38.4%, p = 0.38) or PTSD (61.5% vs. 44.1%, p = 0.34) for men nor of MDD (72.9% vs. 40.1%, p = 0.07) or PTSD (83.6% vs. 52.4%, p = 0.06) for women who experienced community based sexual violence. Johnson et al., Depression Patient Health Questionnaire–9 Sexual violence – defined above 31% of those who experienced sexual violence had 2014 Kenya PTSD PSS-I anxiety and depression before the 2007 election, 45% who experienced sexual violence had anxiety and depression during the election, and 33.7% who experienced sexual violence had anxiety and depression after the 2007 election. The weighted prevalence of MDD (41.0%, CI 27, 55 vs. 35.0%, CI 29.2, 40.8) and PTSD (40.1%, CI 28.6, 51.6 vs. Koegler and Kennedy Conflict and Health (2018) 12:20 Page 12 of 22 Table 2 Study outcomes for association between mental health and HIV risk (Continued) First author & Mental Mental health scales HIV risk measures Results Year health disorders 30.9%, CI 25, 36.8) were not significantly different between those who reported sexual violence and those who did not report sexual violence. Cardozo et al., PTSD HTQ Rape Rape was not related to PTSD symptoms: 2000 Kosovo 21.6% or women who reported rape had symptoms of PTSD vs. 16.92% of women who did not report rape, p =0.49; AOR = 1.68, CI 0.69, 4.08 Sabin et al., Depression HSCL-25 Sexual abuse or rape reported as traumatic event Sexual abuse or rape was independently associated with 2003 Anxiety HTQ anxiety (p = 0.02) but sexual abuse did not remain Guatemalan PTSD significant in the full model. refugees living All rape survivors (N = 6, 100%) experienced anxiety. in Mexico Sexual abuse or rape was not related to PTSD or depression. Wolfe et al., PTSD Mississippi Scale for Combat-related PTSD Sexual assault defined as a sexual experience that was Women who were sexually assaulted experienced a 1998 USA unwanted and involved the use or threat of force significant 18.9 point increase in PTSD scores (M = 91.83, (attempted or completed rape) either by strangers or SD = 22.69) compared to women with no sexual people you knew harassment (M = 71.36, SD = 17.53). Women who were sexually assaulted had increased risk for PTSD compared to women who were only physically (12.5 point difference) or verbally (15.9 point difference) harassed. Washington et PTSD 7-item screen for DSM IV PTSD History of military sexual assault Women with PTSD were significantly more likely to have al., 2013 USA had experienced sexual assault in military (43% vs. 5.1%, p <0.001). Kang et al., 2005 PTSD PCL Sexual assault Among female (AOR = 5.41; 95% CI 3.19, 9.17) and male USA (AOR = 6.21 CI 2.26, 17.04) veterans, sexual assault was significantly associated with PTSD even while controlling for other covariates. HIV acquisition/disease progression and mental health outcomes Epino et al., Depression HSCL-15 CD4 count There was not a significant difference in depression for 2012 Rwanda those with <=200 CD4 cell count (25.5) and > 200 CD4 count (26) (p = 0.58). Mugisha, PTSD Mini-International Neuropsychiatric Interview HIV status Those reporting HIV+ status had greater odds of having Muyinda, (MINI) Sexual trauma events PTSD (UOR = 2.09, CI 1.48, 2.95) Wandiembe et Those who experienced 1–2 sexual trauma events had al., 2015 greater odds of having PTSD in the unadjusted (UOR = 2.6, Uganda CI = 1.63, 4.15) but not the adjusted (AOR = 1.23, CI 0.73, 2.07) model Those who experienced 3+ sexual trauma events had greater odds of having PTSD (UOR = 5.65, CI 3.33, 9.61; AOR = 2.02, CI 1.08, 3.76) Mugisha, Depression MINI HIV status Muyinda, High risk sexual behaviors Koegler and Kennedy Conflict and Health (2018) 12:20 Page 13 of 22 Table 2 Study outcomes for association between mental health and HIV risk (Continued) First author & Mental Mental health scales HIV risk measures Results Year health disorders Malamba et al., Receiving HIV treatment HIV+ status was related to MDD (UOR = 2.85, CI 2.04, 3.96), 2015 Uganda after adjusting for sex and age (AOR = 2.63, CI 1.87, 3.70), and in the multivariate model (OR = 1.83, CI 1.22, 2.74) High risk sexual behavior was not related to MDD in the unadjusted (UOR = 1.13, CI 0.77, 1.67) or adjusted model (AOR = 1.37, CI 0.91, 2.09) Receiving HIV treatment was related to MDD in the adjusted model (AOR = 3.22, CI 1.08, 9.57) but not the unadjusted model (UOR = 2.03, CI 0.85, 4.85) Muldoon et al., Depression APAI All participants had exchanged sex for money or resources For all participants the mean score for the depression 2014 Uganda and anxiety in the previous 30 days sub-scale was 12.84 (SD = 4.79) and the mean score for the anxiety sub-scale was 8.76 (SD = 5.14). No cut off score is defined for symptomatic for either subscale. M.H. Cohen et Depression CES-D About 50% of participants in each group of HIV-positive Women with HIV infection were more likely than HIV- al., 2009 PTSD HTQ and HIV-negative experienced genocidal rape negative women to have clinically significant Rwanda CD4 cell counts depression (81% vs. 65%, p < 0.0001) and MDD (31% vs. 23%, p <0.047). Women with more advanced HIV, indicated by CD4 cell counts < 200 = mL (OR 4.97, CI 2.93, 8.45), were the most likely to have depressive symptoms. Women who had experienced genocidal rape were more likely to have PTSD in unadjusted analyses (OR = 1.63, CI 1.23, 2.15). Depressive symptoms were higher in women who had a history of genocidal rape (OR = 1.56, CI 1.12, 2.16). M.H. Cohen et Depression CES-D About 50% of participants in each group of HIV-positive HIV-positive status was related to increased symptoms of al., 2011 PTSD HTQ and HIV-negative experienced genocidal rape depression (81.5% vs. 63.8%, p < 0.0001), marginally related Rwanda to symptoms of PTSD (59.6 vs. 67.5%, p = 0.081), and not related to MDD (29.2% vs. 22.7%, p = 0.11) compared to HIV-negative status at baseline. There was a continued reduction in PTSD at each follow-up visit for both HIV-positive and HIV-negative groups (6 month change = −0.78, p < 0.0001; 12 month change = − 0.9, p < 0.0001; 18 month change = − 0.84, p < 0.0001). HIV-positive status (b = 0.03, p = 0.38) was not related to PTSD improvement from baseline to 18-month follow up. All participants had fewer depressive symptoms at 18 months follow up compared to baseline (77% vs. 57%). In changes from baseline to visit 4, experiencing genocidal rape was significantly associated with reduced PTSD. Other associations between mental health and HIV acquisition and disease progression Gard et al., 2013 Depression CES-D About 50% of participants in each group of HIV-positive Rwanda PTSD HTQ and HIV-negative experienced genocidal rape Koegler and Kennedy Conflict and Health (2018) 12:20 Page 14 of 22 Table 2 Study outcomes for association between mental health and HIV risk (Continued) First author & Mental Mental health scales HIV risk measures Results Year health disorders HIV-positive women had higher depression scores than HIV- negative participants (23.67, SD = 9.19 vs. 20.79, SD = 9.60, p <0.001). More HIV-positive women met criteria for depression than HIV-negative women (81.46% vs.64.58%, p < 0.001). There was no difference in PTSD scores between HIV- positive and HIV-negative women (2.31, SD = 0.69 vs. 2.4 SD = 0.67, p = 0.09). A greater percentage of HIV-negative compared to HIV- positive women experienced elevated PTSD scores (65.63% vs. 57.8%, p = 0.05). Kohli et al., 2014 Depression HSCL-15 Rape Rape or sexual assault in the past 10 years was related to Democratic PTSD HTQ increased symptoms of PTSD (β = 0.35, p < 0.001) and Republic of depression (β = 0.29, p < 0.001) in multivariate regression. Congo Sinayobye et al., Depression CES-D CD4 count Depression scores were associated with CD4 count (p < 0.001) 2015 Rwanda PTSD HTQ with: CD4 counts > 350 having a mean depression score of 22.4 ± 9.3; CD4 count 200–350 having a mean depression score of 23.0 ± 8.2; CD4 count < 200 having a mean depression score of 25.8 ± 9.1 PTSD scores were not associated with CD4 count (p =0.60) with: CD4 counts > 350 having a median (IQR) PTSD score of 2.1 (1.7–2.7) CD4 count 200–350 having a median (IQR) PTSD score of 2.1 (1.8–2.8) CD4 count < 200 having a median (IQR) PTSD score of 2.2 (1.8–2.8) AOR adjusted odds ratio, ART Antiretroviral therapy, CI confidence interval, MDD major depressive disorder, NR not reported, OR odds ratio, SE Standard Error, SD standard deviation Effect size data are reported where available; textual descriptions of results are reported when that was all that the authors present Koegler and Kennedy Conflict and Health (2018) 12:20 Page 15 of 22 spreadsheet. For each included study the following infor- considered the outcome. Results are organized by the mation was extracted where applicable: citation; location, outcome variable reported by the authors (using the au- setting and target group; study design; sample size; age thors’ expected directionality of the association) as fol- range; gender; random or non-random selection of partici- lows: mental health and HIV serostatus/HIV-related pants; length of follow up; outcome measures; comparison outcomes; sexual violence and mental health outcomes; groups; effect sizes; confidence intervals; significance HIV acquisition/disease progression and mental health levels; measures of HIV risk, mental health conditions and outcomes; and other associations. Under other associa- measures; funding source; and study limitations. To assess tions, we include studies that met the inclusion criteria study quality, we extracted data on study design, sampling but that did not specify the expected directionality of strategy, sample size, and participant characteristics, the associations. Twenty-five studies were from sub- presented in Table 1. These factors were then consid- Saharan Africa, five were from the United States of ered in relation to study quality as presented in the America (USA), and one each was from the Middle East, results and discussion. We did not conduct meta- Europe, and Latin America. Most studies were cross- analysis due to the diversity of populations, study de- sectional, though three time-series studies and seven co- signs, and measured outcomes. hort studies reported longitudinal data. Most studies had non-random sampling of participants, though 14 studies Results utilized a form of random sampling. Overall, 110,818 Of 714 citations identified through the search strategy, 33 participants were included across articles. However, publications were included in this review (Tables 1 and 2). there was overlap in participants across four papers dis- Figure 1 presents a flowchart of the search and screening cussing child soldiers in Sierra Leone [22–25], in five pa- process. Eighteen articles were identified via database pers discussing HIV-positive and negative women, half searching, thirteen through reference searching, and two of whom survived rape during the genocide in Rwanda through searching journal table of contents. [26–30], in two papers from Uganda [31, 32], and in two Table 1 provides information on location, study design, papers from the Wayo-Nero Study in Uganda [33, 34]. sampling strategy, study size, and participant characteris- The smallest study included 120 participants and the lar- tics for each included article. Although most studies gest study included 71,504 participants. were cross-sectional, many articles presented the ex- Table 2 presents each of the included studies by author pected directionality of the relationship based on which and year, country, mental health disorders and measure- variable was considered the exposure and which was ment scales, HIV risk measures, and the relationship Fig. 1 Flow Diagram of review. This flow diagram depicts the search and screening process for the review. Of the 714 records identified through database searching and other sources, 33 articles were ultimately included in this review Koegler and Kennedy Conflict and Health (2018) 12:20 Page 16 of 22 found between mental health and HIV-related outcomes. conflict in Iraq and Afghanistan who had PTSD were Most studies reported mental health outcomes either in significantly more likely to have six of seven different relation to sexual violence or other factors related to STIs compared to veterans without any mental health HIV acquisition and disease progression. Eight studies diagnosis; female veterans with depression were more reported HIV serostatus or HIV risk outcomes. Three likely to have all seven STIs compared to those without studies reported other outcomes but included analyses any mental health diagnosis [39]. Those with comorbid of the association between mental health and HIV ac- depression and PTSD were even more likely to have all quisition and disease progression. In several studies, seven STIs compared to those without any mental health the relationship between HIV-related factors with mental diagnosis with adjusted OR between 2.55 and 4.74. health was not the main objective, but rather one of many This study had the largest population amongst the results reported. included articles (N = 71,504) and represented the entire population of female veterans who met inclu- Mental health and HIV serostatus/HIV-related outcomes sion criteria. Of the eight studies that reported HIV serostatus or Finally, one study from Uganda reported intimate part- HIV-related outcomes, five included HIV sexual risk ner sexual violence as an outcome. In this a cross-sectional behaviors, two reported HIV serostatus, two reported study, females who experienced sexual violence from an in- other STIs, and one reported intimate partner sexual timate partner had greater odds of probable depression violence. (AOR = 4.20, CI 1.54, 11.46) [31]. Four HIV sexual risk behavior studies found some re- Overall, studies reported strong associations between lationship between mental health and HIV risk behav- depression and PTSD and HIV serostatus/HIV-related iors. A study from Uganda found a relationship between outcomes in African and American conflict-affected depression and at least one of eight sexual risk behaviors populations. amongst males in multivariate analysis in a war-affected community (odds ratio (OR) = 1.70, p = 0.05) [32]. In one Rwandan study, increased symptoms of PTSD were cor- Sexual violence and mental health outcomes related with at least one of three HIV risk-taking behav- Twenty-two studies reported mental health outcomes: iors (r = 0.24, p = 0.0006) [35]. In another Rwandan two reported depression only; six PTSD only; six depres- study, women who had exchanged sex had greater odds sion and anxiety; six depression and PTSD; and two de- of depression (OR = 1.74, confidence interval (CI) 1.10, pression, anxiety, and PTSD. A large number of studies 2.76) and PTSD (OR = 1.68, CI 1.19, 2.36) [30]. Women in this review (n = 16) were conducted in sub-Saharan who used condoms at least 50% of the time had de- Africa and reported the relationship between mental creased odds of PTSD (OR = 0.60, CI 0.42, 0.86), but in- health and sexual violence, rape, or sexual abuse (as an creased odds of elevated depression scores (OR = 1.84, HIV-related measure out of the victims’ control). The CI 1.20, 2.82) [30]. Women who had sex in the last six remaining studies in this category were conducted in the months had decreased odds of depression (OR = 0.57, CI USA (n = 3), with Kosovar ethnic Albanians (n = 1), and 0.04, 0.81) [30]. A longitudinal study from the USA with Guatemalan refugees in Mexico (n = 1). showed that PTSD predicted unprotected sex four Most studies found a positive association between sex- months later (OR = 1.57, CI 1.20, 2.04) [36]. In Lebanon, ual violence and poor mental health; three studies found no relationship was found between PTSD and risky sex- no association. Sexual abuse predicted depression and ual activities [37]. anxiety for female Ugandan youth (β = 0.32, p < 0.001) Two studies focused on HIV serostatus as an outcome. [40]. Experiencing sexual violence was related to PTSD Depression (p < 0.001) but not PTSD (p = 0.06) was re- in a study among Ugandan adults (β = 3.75, p < 0.05) lated to HIV positive serostatus among female survivors [41]. Similarly, in another study of Ugandan adults, those of the Rwandan genocide in a cross-sectional, non- who experienced three or more sexual trauma events randomly selected cohort [30]. In a randomly selected had greater odds of PTSD (AOR = 2.02, CI 1.08, 3.67) cross-sectional study in Uganda, men and women with [34]. For Ugandan adults living in camps for internally depression (adjusted odds ratio (AOR) = 1.89, CI 1.28, 2. displaced people, those who reported rape had greater 80) and PTSD (AOR = 1.44, CI 1.06, 1.96) had greater odds of PTSD (AOR = 1.76, CI 1.01, 2.75), but not depres- odds of testing positive for HIV [38]. sion [42]. Amongst adult and adolescent Guatemalan Two studies reported STIs other than HIV as an out- refugees in Mexico, sexual abuse was independently come. Female survivors of the Rwandan genocide who associated with anxiety, but did not remain significant had a non-HIV STI had greater odds of elevated depres- after controlling for other variables [43]. Among Kosovar sion scores (AOR = 1.64, CI 1.01, 2.65), but not PTSD ethnic Albanian women, rape was not related to PTSD (OR = 1.07, CI 0.77, 1.50) [30]. Female USA veterans of (AOR = 1.68, CI 0.69, 4.08) [44]. It is notable that in Koegler and Kennedy Conflict and Health (2018) 12:20 Page 17 of 22 this study, rape was only identified in 4.4% of women longer significant when protective factors were added (N = 60), [44]. [22]. Rape was the only time-invariant predictor found to Three studies by the same study team examined the be related to depression/anxiety (considered together) relationship of sexual violence with depression and after adjusting for hardship and protective factors (b = 4.34, PTSD in different countries. The studies utilized differ- p = 0.04) [25]. Combined, these studies demonstrate that ent methodologies and had different outcomes. Sexual rape (as a more extreme war event) is a strong predictor of violence was related to PTSD and depression in Liberian anxiety over time amongst war-affected adolescents. adults, with prevalence of PTSD higher for male (81% Three publications examined sexual assault and men- vs. 46%, p < 0.001) and female (74% vs. 44%, p = 0.005) tal health in conflict-affected American military popula- former combatants who experienced sexual violence; de- tions. Female veterans who returned from deployment pression was higher for male former combatants who in the Persian Gulf were more likely to experience PTSD experienced sexual violence compared to those who did if they had been sexually assaulted during their service not (64% vs.42%, p = 0.003) [45]. In the Democratic Re- compared to women who did not experience harass- public of Congo, prevalence of PTSD (70.2% vs. 40.3%, ment, who only experienced verbal harassment, and who p < 0.001) and depression (60.4% vs. 30.7%, p < 0.001) experienced physical harassment that was not sexual as- were higher for participants who experienced sexual vio- sault [49]. In cross-sectional analyses, female veterans lence compared to those who did not; prevalence of with PTSD compared to those without PTSD were more PTSD and depression were higher for females who likely to have experienced sexual assault while in the experienced conflict-related sexual violence compared military (43% vs. 5.1%, p < 0.001) [50]. Sexual assault re- to females who did not, respectively (75.9% vs. 44.4%, ports were significantly associated with PTSD for men p < 0.001 and 67.7% vs. 30.3%, p < 0.001) [46]. During (OR = 6.21, CI 2.26, 17.04) and women (OR = 5.41, CI 3. the 2007 election violence in Kenya, the prevalence of 19, 9.17) after controlling for other variables [51]. Find- sexual violence was not significantly related to major ings from these studies supported a relationship between depressive disorder or PTSD [47]. sexual assault and poor mental health. The Liberian study had the strongest methodological Overall, studies reported strong associations between quality (utilizing a population based multistage random sexual violence and all three mental health outcomes. household cluster survey) and the study in the Demo- Sexual violence was most strongly associated with PTSD cratic Republic of Congo had the weakest quality of the [41, 42, 45, 46, 49–51], then depression [22, 24, 25, 40, three (non random accessible population cluster). Al- 45, 46], and then anxiety mainly amongst male survivors though sexual violence in Kenya was not significantly re- [22–25, 40, 43]. lated to major depressive disorder and PTSD, sexual violence was significantly related to suicidal ideation and HIV acquisition/disease progression and mental health suicide attempt [47]. outcomes Sexual torture and being forced to marry were not re- Factors related to HIV acquisition and disease progres- lated to depression, anxiety, and PTSD amongst formerly sion were associated with mental health outcomes in war abducted adolescents in Uganda [48]. However, the two studies that reported HIV sexual risk behavior, one study may not have had the statistical power to detect sig- that reported CD4 count, one that reported receiving nificance forsuchanalysisasthe primaryobjective wasto HIV treatment, two that reported HIV status, and two compare formerly abducted and non-abducted adolescents. that compared HIV-positive and HIV-negative partici- No trauma event had a significant relationship with mental pants and rape during genocide. health outcomes in this randomly selected population. In Uganda, formerly conflict-abducted young women Four studies reported data from a non-randomly se- who recently exchanged sex for money or resources lected prospective cohort of child soldiers in Sierra were assessed for mental health with a locally developed Leone. In two studies with overlapping samples (one measure; mean depression and anxiety scores were 12.84 cross-sectional with 273 participants and one longitu- and 8.76, respectively [52]. Cut off scores defining symp- dinal with 156 participants), rape was significantly re- tomatic versus asymptomatic depression and anxiety were lated to anxiety (b = 2.85, p = 0.05; b = 4.06, p < 0.05), not provided in this study so no conclusion can be made especially in males, but not depression after controlling regarding the association between HIV sexual risk with for other variables [23, 24]. In a slightly different sample, depression and anxiety. In another study in Uganda, rape was associated with increased anxiety after control- adults with high risk sexual behavior did not have higher ling for discrimination and protective factors (b = 5.35, odds of depression (AOR = 1.37, CI 0.91, 2.09) [33]. p < 0.001); rape significantly predicted an increase in Lower CD4 count indicates more advanced HIV disease depression over time, but this relationship lessened progression and potentially increases transmissibility. In when perceived discrimination was added and was no one non-randomly selected population in Rwanda, there Koegler and Kennedy Conflict and Health (2018) 12:20 Page 18 of 22 was no difference in depression scores for adults with increased PTSD (β =0.35, p < 0.001) and depression (β =0. CD4 counts at or below 200 compared to adults with 29, p < 0.001). CD4 counts over 200 [53]. Receiving HIV treatment Overall, studies in this category demonstrated incon- was associated with depression (AOR = 3.22, CI 1.08, sistent associations between mental health and HIV ac- 9.57) among a random selection of adults in Uganda quisition and disease progression. [33]. Reporting an HIV-positive status was associated with both PTSD (OR = 2.09, CI 1.48, 2.95) [34]and Discussion depression (OR = 1.83, CI 1.22, 2.74) [33]among Overall, the thirty-three studies included in this review Ugandan adults. suggest that symptoms of poor mental health are associ- In two publications from another non-randomly selected ated with increased sexual risk behaviors and HIV cohort in Rwanda, depression and PTSD in HIV-positive markers, and HIV exposures and HIV-related health sta- and negative women were examined both cross-sectionally tus are associated with poor mental health symptoms in and longitudinally. At baseline, HIV-positive women were conflict-affected populations. Only five studies found no more likely than HIV-negative women to have depression association between mental health and factors related to symptoms (81% vs. 65%, p < 0.0001) and HIV-positive HIV acquisition and disease progression. Associations women with CD4 counts below 200 were the most likely were strongest for mental health and HIV serostatus/ to have depression symptoms (OR = 4.97, CI 2.93, 8.45) HIV-related outcomes and sexual violence and mental [26]. Women who experienced rape during the genocide health outcomes. Associations between HIV acquisition/ were more likely to have PTSD (OR = 1.63, CI 1.23, 2.15) disease progression and mental health outcomes and and depression (OR = 1.56, CI 1.12, 2.16) in unadjusted outcomes that examined other associations were more analysis [26]. Over three follow-up visits at six, twelve, and inconsistent. This could be partially attributed to the eighteen months, HIV-positive and HIV-negative women greater variety in the types of associations measured in had reduced PTSD scores; this improvement was not re- these categories. For example, studies that examined lated to antiretroviral therapy (ART) use, but was related sexual violence and mental health outcomes only in- to HIV-positive status at follow-up three, and was related cluded associations between sexual violence and mental to rape during the genocide at follow-ups two and three health, given the large number of studies in this cat- [27]. Overall, studies examining mental health outcomes egory. However, studies that examined HIV acquisition/ and factors related to HIV acquisition and disease progres- disease progression and mental health outcomes looked sion reported both positive and null associations. at a variety of HIV-related correlates with mental health outcomes (e.g. CD4 count, sexual risk behaviors, HIV positive serostatus). Other associations between mental health and HIV There were slight inconsistencies across studies in as- acquisition and disease progression sociations. For articles that found an association between Three studies used other outcomes for their primary one or more mental health disorders and factors related analyses but provided data assessing the cross-sectional to HIV acquisition and disease progression, there were association between mental health and HIV acquisition other articles, or findings within the same article, that and disease progression, making less clear the authors’ failed to identify a relationship with one or more other expected directionality of the relationship. One study in disorders. These inconsistencies may represent real dif- Rwanda (with some study population overlap with other ferences in various associations or populations. Incon- included articles) examined quality of life amongst trauma- sistencies in the findings may also be attributed to less affected women with and without HIV. About half of the rigor among some studies that did not find associations: women in each group had experienced rape during the non-random selection and small sample size [37], small genocide. HIV-positive women experienced increased de- numbers of relevant participants within the sample [43, pression symptom scores compared to HIV-negative 48], and low rates of the HIV-related variable [35]. Some women (p < 0.001), but a greater percentage of HIV- studies that did not find associations had trends towards negative women experienced elevated symptoms of PTSD an association [47] or a marginal association [31]. Stud- [28]. In another study in Rwanda among HIV positive ies with strong rigor (large sample sizes of randomly se- women (with study population overlap) depression (p <0. lected participants or the entire population) found strong 001), but not PTSD (p = 0.60), was associated with CD4 positive associations both in conflict-affected USA [39, 50, count [29]. In the Democratic Republic of Congo, family 51] and African populations [31, 38]. rejection amongst survivors of sexual assault was more Our findings are similar to a review of HIV risk behav- strongly related to depression, but increased PTSD was iors and trauma amongst migrants from low and middle- more strongly related to sexual assault [54]. Rape or sexual income countries [18]. Our study differs from the review assault in the past ten years was associated with both by Michalopoulos et al. by including associations between Koegler and Kennedy Conflict and Health (2018) 12:20 Page 19 of 22 mental health and factors related to HIV acquisition and HIV focuses on this geographic area. Similarly, a wide disease progression in conflict-affected populations from range of conflict-affected populations were included in high-income countries, focusing on measurable mental this review, from soldiers fighting in their own country health disorders, and exclusively presenting quantitative to returned American combat veterans, from orphans to relationships. child soldiers, from abducted civilians to those living in We only included mental health symptoms docu- internally displaced persons camps. Unfortunately, there mented by validated scales to ensure that included stud- were not enough studies examining the same factors in ies met a minimal level of measurement rigor and to the range of conflict-affected regions or populations to increase comparability of findings across studies. Most draw strong conclusions as to how these relationships included articles meet the criteria for ‘protracted’ or might differ by region or population. ‘post-conflict’ [21, 47]. However, one study stands out as We used findings from this review along with existing it specified reporting election-related violence rather literature to develop a framework (Fig. 2). Mock et al. than war or conflict; with 1133 deaths over 59 days, the described how HIV prevalence could increase (via in- setting met criteria for an ‘explosive’ event [21, 47]. This creased military and civilian interaction, increased com- demonstrates the range of included conflict-affected set- mercial and casual sex, etc.) or decrease (via increased tings. Because HIV disproportionally affects sub-Saharan isolation, mortality of high-risk populations, etc.) in Africa, much of the literature discussing conflict and conflict-affected settings [13]. It is well established in the Fig. 2 Relationship Between Mental Health & HIV in Conflict Settings. This figure illustrates a framework of the relationship between mental health and HIV serostatus and HIV-related outcomes among conflict-affected populations, based on this review and existing literature. The outside blue panel presents existing knowledge of factors in conflict-affected settings. HIV prevalence could increase or decrease in conflict-settings through various mechanisms [13]. Multiple factors contribute to increased risk for poor mental health in conflict-affected settings. Additional factors adversely affect conflict-affected populations. The inside box presents the relationship between mental health and HIV-related outcomes. HIV can physiologically, psychologically, and socially increase risk for mental health disorders. Health status (being HIV-positive and lower CD4 count) is associated with poor mental health. Poor mental health can influence HIV risk exposures (sexual risk behaviors and STIs). Surviving sexual assault is associated with poor mental health and HIV-related outcomes. Demographic factors can influence each relationship Koegler and Kennedy Conflict and Health (2018) 12:20 Page 20 of 22 literature that conflict-affected populations are at risk health disorders using appropriate cross-culturally vali- for poor mental health [8, 12, 55], and often lack access dated tools. Referrals could then be made, assuming avail- to mental health services [56, 57]. Physiological, psycho- ability of evidence-based interventions. Similarly, programs logical, and social pathways can influence the relation- delivering mental health services should consider screening ship between HIV-related outcomes and mental health for HIV serostatus and associated risk factors. Since health disorders [7]. What is lacking in the literature, to the infrastructure is often limited in conflict-affected settings, best of our knowledge, is a framework that incorporates combining screening and services offers the potential to the relationship between mental health and factors asso- more systematically and holistically treat vulnerable indi- ciated with HIV acquisition and progression in conflict- viduals. An example where this has occurred is in Uganda, affected populations. where an organization that provided mental health inter- Our findings provide evidence for this framework. ventions for conflict survivors included HIV screening, re- Specifically, we identified studies that demonstrate ferrals, and services to meet the unique mental health health status (HIV-positive serostatus, lower CD4 count) needs of people living with HIV [61]. is associated with increased depression [26–30, 38], that At the policy level, by recognizing the relationships be- mental health disorders may influence HIV risk exposures tween mental health and factors related to HIV acquisition (sexual risk behaviors [30, 32, 36] and STIs [30, 39]), and and disease progression in conflict settings, infrastructure that surviving sexual assault may be associated with poor can be integrated to offer mental health and HIV-related mental health [22–25, 40, 41, 43, 45, 46, 49–51]. services simultaneously. Policy could also require moni- There were several limitations to this review. First, toring of results to recognize any differences in separate only one reviewer identified, screened, and extracted treatment compared to integrated treatment of mental data from included studies. Methodological rigor would health and HIV-related services. Future research should be strengthened if two reviewers had independently employ stronger methods where possible – specifically, completed each step and resolved any discrepancies. Al- random selection of participants to decrease bias and though it was not possible to have two reviewers conduct longitudinal studies to better determine directionality of all steps in the review process, a second, experienced re- the measured associations. Research should also examine viewer was consulted when specific questions arose during the associations of mental health with HIV acquisition and the process and a third reviewer verified the extracted disease progression with a wider range of conflict-affected data. Second, the search terms for ‘conflict’ focused only populations, as the relationship may vary depending on on conflict and war. By excluding terms such as refugee, the population and the possibilities for risk exposure. displaced persons, and asylum seekers we may have missed articles relevant to this review. Conclusions We examined only three mental health disorders, se- Existing literature demonstrates that depression, anxiety, lected because they are common and frequently mea- and PTSD have been quantifiably associated with four sured in conflict-affected populations. However, other factors related to HIV acquisition and disease progres- disorders may be relevant in conflict-affected popula- sion in conflict-affected populations: markers of HIV tions, specifically substance use, which has been shown risk (i.e. STIs), HIV-related health status (e.g. CD4 to be common, harmful, and related to HIV transmis- count), sexual risk behaviors, and HIV risk exposures (i. sion and risk in conflict-affected populations [58–60]. Fi- e. sexual violence). Specifically, poor mental health has nally, 23 of the 33 studies were cross-sectional, so the been associated with two outcomes, HIV markers and temporality of these relationships cannot be determined. sexual risk behaviors, while HIV risk exposures and No studies were identified in this review that examined health status have been associated with the outcome of the association between viral suppression and mental poor mental health. Additional research utilizing ran- health; future studies should examine these associations. dom selection and longitudinal design can further estab- Longitudinal studies should also be conducted examin- lish the strength of these associations and determine if ing associations between mental health and factors re- HIV and mental health services need to be integrated for lated to HIV acquisition and disease progression. Future conflict-affected populations. reviews could examine the associations between HIV ac- quisition or risk behaviors and substance use or other Additional file mental health disorders in conflict-affected populations. There are several implications from this review. Con- Additional file 1: Scoping review search terms for PubMed. (DOCX 14 kb) sidering the associations discussed in this paper, pro- grams delivering HIV or other reproductive health Abbreviations services for conflict-affected populations should consider AOR: Adjusted Odds Ratio; ART: Antiretroviral Therapy; CI: Confidence screening individuals who have HIV or STIs for mental Interval; HIV: Human Immunodeficiency Virus; HTQ: Harvard Trauma Koegler and Kennedy Conflict and Health (2018) 12:20 Page 21 of 22 Questionnaire; IDP: Internally Displaced Person; MDD: Major Depressive and displacement: a systematic review and meta-analysis. JAMA. 2009; Disorder; OR: Odds Ratio; PTSD: Posttraumatic Stress Disorder; SD: Standard 302(5):537–49. Deviation; SE: Standard Error; STI: Sexually Transmitted Infection; 13. Mock NB, et al. Conflict and HIV: a framework for risk assessment to prevent UNESCO: United Nations Educational, Scientific and Cultural Organization; HIV in conflict-affected settings in Africa. Emerg Themes Epidemiol. 2004;1(1):1. USA: United States of America 14. Salama P, Dondero TJ. HIV surveillance in complex emergencies. AIDS. 2001; 15:S4–S12. Acknowledgements 15. Spiegel PB, et al. Prevalence of HIV infection in conflict-affected and The authors thank Peggy Gross for assistance with generating the search terms. displaced people in seven sub-Saharan African countries: a systematic We also thank David Celentano for his extensive feedback on an earlier draft review. Lancet. 2007;369(9580):2187–95. and Jessica Haslag for verifying the extracted data. 16. Bennett BW, et al. Hiv incidence prior to, during, and after violent conflict in 36 sub-Saharan african nations, 1990-2012: an ecological study. PLoS One. Funding 2015;10(11):e0142343. This research was supported by NIMH NRSA F31MH095678. 17. Mulanga C, et al. Political and socioeconomic instability: how does it affect HIV? A case study in the Democratic Republic of Congo. AIDS. 2004;18(5):832–4. Availability of data and materials 18. Michalopoulos LM, Aifah A, El-Bassel N. A systematic review of HIV risk The data extracted from each included article is available upon request. behaviors and trauma among forced and unforced migrant populations from low and middle-income countries: state of the literature and future Authors’ contributions directions. AIDS Behav. 2015:1–19. EK extracted the data, coded included articles, analyzed the data, and was 19. Collins PY, et al. What is the relevance of mental health to HIV/AIDS care the primary writer of the manuscript. CK provided substantial contributions and treatment programs in developing countries? A systematic review. to the design of the study, analysis and interpretation of the data, and AIDS. 2006;20(12):1571. provided critical revisions throughout several drafts of the manuscript. Both 20. Moher D, et al. Preferred reporting items for systematic reviews and meta- authors read and approved the final manuscript. analyses: the PRISMA statement. Ann Intern Med. 2009;151(4):264–9. 21. Strand, H. and M. Dahl, Defining Conlict-affected Countries. 2010: Unesco. Ethics approval and consent to participate 22. Betancourt TS, et al. Past horrors, present struggles: the role of stigma in the Not applicable association between war experiences and psychosocial adjustment among former child soldiers in Sierra Leone. Soc Sci Med. 2010;70(1):17–26. Consent for publication 23. Betancourt TS, et al. Sierra Leone's child soldiers: war exposures and mental Not applicable health problems by gender. J Adolesc Health. 2011;49(1):21–8. 24. Betancourt TS, et al. Sierra Leone’s former child soldiers: a follow-up study Competing interests of psychosocial adjustment and community reintegration. Child Dev. 2010; The authors declare that they have no competing interests. 81(4):1077–95. 25. Betancourt TS, et al. Sierra Leone's former child soldiers: a longitudinal study of risk, protective factors, and mental health. J Am Acad Child Adolesc Publisher’sNote Psychiatry. 2010;49(6):606–15. Springer Nature remains neutral with regard to jurisdictional claims in published 26. Cohen MH, et al. Prevalence and predictors of posttraumatic stress disorder maps and institutional affiliations. and depression in HIV-infected and at-risk Rwandan women. J Women's Health (Larchmt). 2009;18(11):1783–91. Received: 24 October 2016 Accepted: 25 March 2018 27. Cohen MH, et al. Improvement in posttraumatic stress disorder in postconflict Rwandan women. J Women's Health (Larchmt). 2011;20(9):1325–32. 28. Gard TL, et al. The impact of HIV status, HIV disease progression, and post- References traumatic stress symptoms on the health-related quality of life of Rwandan 1. Nduna M, et al. Associations between depressive symptoms, sexual behaviour women genocide survivors. Qual Life Res. 2013;22(8):2073–84. and relationship characteristics: a prospective cohort study of young women 29. Sinayobye JdA, et al. Prevalence of shingles and its association with PTSD and men in the eastern cape, South Africa. J Int AIDS Soc. 2010;13(1):1. among HIV-infected women in Rwanda. BMJ Open. 2015;5(3):e005506. 2. Agardh A, Cantor-Graae E, Östergren PO. Youth, sexual risk-taking behavior, 30. Adedimeji AA, et al. Sexual behavior and risk practices of HIV positive and and mental health: a study of university students in Uganda. International HIV negative Rwandan women. AIDS Behav. 2015;19(7):1366–78. journal of behavioral medicine. 2012;19(2):208–16. 31. Kinyanda E, et al. Intimate partner violence as seen in post-conflict eastern 3. Cavanaugh CE, Hansen NB, Sullivan TP. HIV sexual risk behavior among low- Uganda: prevalence, risk factors and mental health consequences. BMC Int income women experiencing intimate partner violence: the role of Health Hum Rights. 2016;16(1):5. posttraumatic stress disorder. AIDS Behav. 2010;14(2):318–27. 32. Kinyanda E, et al. Psychiatric disorders and psychosocial correlates of high 4. Hutton HE, et al. HIV risk behaviors and their relationship to posttraumatic HIV risk sexual behaviour in war-affected eastern Uganda. AIDS Care. 2012; stress disorder among women prisoners. Psychiatr Serv. 2001;52(4):508–13. 24(11):1323–32. 5. Munroe CD, et al. The relationship between posttraumatic stress symptoms 33. Mugisha J, et al. Major depressive disorder seven years after the conflict and sexual risk: examining potential mechanisms. Psychological Trauma: in northern Uganda: burden, risk factors and impact on outcomes (the Theory, Research, Practice, and Policy. 2010;2(1):49. Wayo-Nero study). BMC psychiatry. 2015;15(1):48. 6. Rubin A, Gold M, Primack BA. Associations between depressive symptoms 34. Mugisha J, et al. Prevalence and factors associated with posttraumatic stress and sexual risk behavior in a diverse sample of female adolescents. J Pediatr disorder seven years after the conflict in three districts in northern Uganda Adolesc Gynecol. 2009;22(5):306–12. (the Wayo-Nero study). BMC psychiatry. 2015;15(1):170. 7. Prince M, et al. No health without mental health. Lancet. 2007;370(9590): 35. Talbot A, et al. Treating psychological trauma among Rwandan orphans is 859–77. associated with a reduction in HIV risk-taking behaviors: a pilot study. AIDS 8. De Jong JTVM, et al. Lifetime events and posttraumatic stress disorder in 4 Educ Prev. 2013;25(6):468–79. postconflict settings. JAMA. 2001;286(5):555–62. 9. Hamid AARM, Musa SA. Mental health problems among internally displaced 36. Adler AB, et al. Effect of transition home from combat on risk-taking and health-related behaviors. J Trauma Stress. 2011;24(4):381–9. persons in Darfur. Int J Psychol. 2010;45(4):278–85. 10. Kamau M, et al. Psychiatric disorders in an African refugee camp. 37. Svetlicky V, et al. Combat exposure, posttraumatic stress symptoms and International Journal of Mental health, Psychosocial Work and Counselling risk-taking behavior in veterans of the second Lebanon war. Isr J in Areas of Armed Conflict. 2004;2(2):84–9. Psychiatry Relat Sci. 2010;47(4):276–83. 11. Srinivasa Murthy R. Mass violence and mental health–recent 38. Malamba SS, et al. “The Cango Lyec Project-Healing the Elephant”:HIV epidemiological findings. Int Rev Psychiatry. 2007;19(3):183–92. related vulnerabilities of post-conflict affected populations aged 13–49 12. Steel Z, et al. Association of torture and other potentially traumatic events years living in three Mid-Northern Uganda districts. BMC Infect Dis. with mental health outcomes among populations exposed to mass conflict 2016;16(1):690. Koegler and Kennedy Conflict and Health (2018) 12:20 Page 22 of 22 39. Cohen BE, et al. Reproductive and other health outcomes in Iraq and Afghanistan women veterans using VA health care: association with mental health diagnoses. Womens Health Issues. 2012;22(5):e461–71. 40. Amone-P’Olak K, et al. Cohort profile: mental health following extreme trauma in a northern Ugandan cohort of war-affected youth study (the WAYS study). SpringerPlus. 2013;2(1):1–11. 41. Nakimuli-Mpungu E, et al. Implementation and scale-up of psycho-trauma centers in a post-conflict area: a case study of a private-public Partnership in Northern Uganda. PLoS Med. 2013;10(4) 42. Roberts B, et al. Factors associated with post-traumatic stress disorder and depression amongst internally displaced persons in northern Uganda. BMC psychiatry. 2008;8(1):38. 43. Sabin M, et al. Factors associated with poor mental health among Guatemalan refugees living in Mexico 20 years after civil conflict. JAMA. 2003;290(5):635–42. 44. Cardozo BL, et al. Mental health, social functioning, and attitudes of Kosovar Albanians following the war in Kosovo. JAMA. 2000;284(5):569–77. 45. Johnson K, et al. Association of combatant status and sexual violence with health and mental health outcomes in postconflict Liberia. JAMA. 2008; 300(6):676–90. 46. Johnson K, et al. Association of sexual violence and human rights violations with physical and mental health in territories of the eastern Democratic Republic of the Congo. JAMA. 2010;304(5):553–62. 47. Johnson K, et al. A national population-based assessment of 2007–2008 election-related violence in Kenya. Confl Health. 2014;8:2. 48. Okello J, Onen T, Misisi S. Psychiatric disorders among war-abducted and non-abducted adolescents in Gulu district, Uganda: a comparative study. Afr J Psychiatry (Johannesbg). 2007;10(4):225–31. 49. Wolfe J, et al. Sexual harassment and assault as predictors of PTSD symptomatology among US female Persian gulf war military personnel. Journal of Interpersonal Violence. 1998;13(1):40–57. 50. Washington DL, et al. PTSD risk and mental health care engagement in a multi-war era community sample of women veterans. JGIM: Journal of General Internal Medicine. 2013;28(7):894–900. 51. Kang H, et al. The role of sexual assault on the risk of PTSD among gulf war veterans. Ann Epidemiol. 2005;15(3):191–5. 52. Muldoon KA, et al. After abduction: exploring access to reintegration programs and mental health status among young female abductees in northern Uganda. Confl Health. 2014;8:5. 53. Epino HM, et al. Reliability and construct validity of three health-related self-report scales in HIV-positive adults in rural Rwanda. AIDS Care. 2012;24(12):1576–83. 54. Kohli A, et al. Risk for family rejection and associated mental health outcomes among conflict-affected adult women living in rural eastern Democratic Republic of the Congo. Health Care for Women International. 2014;35(7–9):789–807. 55. Miller KE, Andrew R. War exposure, daily stressors, and mental health in conflict and post-conflict settings: bridging the divide between trauma- focused and psychosocial frameworks. Soc Sci Med. 2010;70(1):7–16. 56. Mollica RF, et al. Mental health in complex emergencies. Lancet. 2004; 364(9450):2058–67. 57. Tol WA, et al. Mental health and psychosocial support in humanitarian settings: linking practice and research. Lancet. 2011;378(9802):1581–91. 58. Ezard N. Substance use among populations displaced by conflict: a literature review. Disasters. 2012;36(3):533–57. 59. Ezard N, et al. Six rapid assessments of alcohol and other substance use in populations displaced by conflict. Confl Health. 2011;5(1):1. 60. Friedman SR, Rossi D, Braine N. Theorizing 'Big Events' as a potential risk environment for drug use, drug-related harm and HIV epidemic outbreaks. Submit your next manuscript to BioMed Central International Journal of Drug Policy. 2009;20(3):283–91. and we will help you at every step: 61. Sharer, M. and M. Gutmann, Prioritizing HIV in mental health services delivered in post-conflict settings. 2011. • We accept pre-submission inquiries � Our selector tool helps you to find the most relevant journal � We provide round the clock customer support � Convenient online submission � Thorough peer review � Inclusion in PubMed and all major indexing services � Maximum visibility for your research Submit your manuscript at www.biomedcentral.com/submit http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Conflict and Health Springer Journals

A scoping review of the associations between mental health and factors related to HIV acquisition and disease progression in conflict-affected populations

Free
22 pages

Loading next page...
 
/lp/springer_journal/a-scoping-review-of-the-associations-between-mental-health-and-factors-H1pC0Y1yXa
Publisher
Springer Journals
Copyright
Copyright © 2018 by The Author(s).
Subject
Medicine & Public Health; Medicine/Public Health, general; Emergency Medicine; Conflict Studies
eISSN
1752-1505
D.O.I.
10.1186/s13031-018-0156-y
Publisher site
See Article on Publisher Site

Abstract

The association between poor mental health and factors related to HIV acquisition and disease progression (also referred to as HIV-related factors) may be stronger among conflict-affected populations given elevated rates of mental health disorders. We conducted a scoping review of the literature to identify evidence-based associations between mental health (depression, anxiety, and post-traumatic stress disorder [PTSD]) and factors related to HIV acquisition and progression in conflict-affected populations. Five electronic databases were searched on October 10, 2014 and updated on March 7, 2017 to identify peer-reviewed publications presenting primary data from January 1, 1994 to March 7, 2017. Articles were included if: 1) depression, anxiety, and/or PTSD was assessed using a validated scale, 2) HIV or HIV-related factors were a primary focus, 3) quantitative associations between depression/anxiety/PTSD and HIV or HIV-related factors were assessed, and 4) the study population was conflict-affected and from a conflict-affected setting. Of 714 citations identified, 33 articles covering 110,818 participants were included. Most were from sub-Saharan Africa (n = 25), five were from the USA, and one each was from the Middle East, Europe, and Latin America. There were 23 cross-sectional, 3 time-series, and 7 cohort studies. The search identified that mental health has been quantitatively associated with the following categories of HIV-related factors in conflict-affected populations: markers of HIV risk, HIV-related health status, sexual risk behaviors, and HIV risk exposures (i.e. sexual violence). Further, findings suggest that symptoms of poor mental health are associated with sexual risk behaviors and HIV markers, while HIV risk exposures and health status are associated with symptoms of poor mental health. Results suggest a role for greater integration and referrals across HIV and mental health programs for conflict-affected populations. Keywords: Mental health, HIV, Conflict settings, Depression, Anxiety, PTSD Background and socially increases risk for neuropsychiatric conditions The relationship between mental health and HIV acqui- [7]. In this paper we consider a broad range of factors as- sition and disease progression (also referred to as HIV- sociated with HIV acquisition and disease progression, related factors) is bi-directional. Having symptoms of such as markers of HIV risk, HIV-related health status, post-traumatic stress disorder (PTSD), depression, and/ sexual risk behaviors, and other potential HIV risk expo- or anxiety has been linked to HIV risk factors in various sures not under individual control (i.e. sexual violence). populations both prospectively [1] and cross-sectionally As conflict-affected populations often have elevated [2–6]. Being HIV-positive physiologically, psychologically, rates of PTSD, depression, and anxiety [8–12], the associ- ation between poor mental health and risk for HIV acqui- sition and disease progression may be stronger among * Correspondence: koeglere@health.missouri.edu these populations. Conflict can shape population move- Department of Health Sciences, University of Missouri, 512 Clark Hall, ments, opportunities for sexual partnering, and mortality Columbia, MO 65211, USA Department of International Health, Johns Hopkins Bloomberg School of patterns in ways that might increase or decrease HIV Public Health, 615 N. Wolfe Street Room E5547 Baltimore, Baltimore, MD prevalence [13, 14]. Epidemiological evidence suggests 21205, USA © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Koegler and Kennedy Conflict and Health (2018) 12:20 Page 2 of 22 elevated HIV prevalence in the five years prior to conflict, measures. This review was conducted following PRISMA but an overall decrease in HIV prevalence during and just guidelines [20]. after conflict [15, 16]. However, vulnerable populations may remain at elevated risk for HIV acquisition during Definition of terms political and socioeconomic instability [14, 17]. A seminal We sought to illuminate the ways a broad range of factors paper discussing population vulnerability to HIV trans- related to HIV acquisition and disease progression have mission in conflict-affected settings discusses health fac- been examined in relationship to mental health. There- tors but does not detail the ways poor mental health can fore, we defined factors related to HIV acquisition and impact population vulnerability to HIV [13]. Poor mental disease progression to include factors such as: markers health, conflict, and being HIV positive are independently of HIV risk (i.e. sexually transmitted infections (STIs)); related to morbidity and mortality. Co-occurrence of these HIV-related health status (i.e. HIV seropositive status, factors can contribute to increased vulnerability to mor- CD4 count); sexual risk behaviors (e.g. unprotected sex, bidity and mortality. multiple sexual partners, exchange sex, etc.); and other Other reviews that have examined associations be- potential HIV exposures not under individual control (i.e. tween mental health and HIV risk behaviors or care and sexual assault). A range of factors related to HIV were in- treatment programs have focused on migrant popula- cluded in order to provide a comprehensive understand- tions [18] and populations from developing countries ing of how researchers have quantitatively examined the [18, 19]. It is yet unknown how the vulnerabilities of relationships between specific mental health disorders and poor mental health and factors related to HIV acquisi- HIV-related factors. tion and disease progression operate in conflict-affected Conflict-affected settings were defined according to populations. Understanding how mental health is associ- UNESCO as areas with ‘explosive’ (over 200 battle-related ated with HIV acquisition and disease progression in deaths in a year) or ‘protracted’ (over 1000 battle-related conflict-affected populations can inform program and deaths over ten years) events [21]. Both active conflict and policy work in these settings. The aim of our study was post-conflict settings were included. Since not all popula- to conduct a scoping review of the literature to identify tions in conflict-affected countries are directly affected by evidence-based associations between common mental conflict, the study population had to be affected by con- health conditions (depression, anxiety, and PTSD) and flict and described as such by the article’sauthors.All factors related to HIV acquisition and disease progression combat-affected populations were considered for inclu- in conflict-affected populations. We sought to understand sion, both combatants and civilians. Conflict-affected the bi-directional associations between these mental populations across the economic spectrum were consid- health conditions and various measures of HIV-related ered for inclusion to examine the relationship between factors, and to examine the strength and directionality of mental health and HIV-related factors in a variety of associations to offer suggested directions for future re- economic situations. search, policy, and interventions. Search strategy Methods Five electronic databases (PubMed, PsycINFO, SCOPUS, Peer-reviewed publications that presented primary data CINAHL, and EMBASE) were searched first on October from January 1, 1994 to March 7, 2017 were included in 10, 2014 and updated on March 7, 2017. Search terms this review if they met the following inclusion criteria: 1) included combinations of terms for mental health, HIV one or more of three common mental health conditions risk, and conflict-affected settings (Additional file 1). We (depression, anxiety, PTSD) was a primary or substantive also searched reference lists of included articles and focus of the article and was assessed using a validated hand searched the table of contents of Conflict and scale, 2) HIV serostatus or factors related to HIV acquisi- Health and Medicine Conflict and Survival. Only articles tion and disease progression (defined below) was a pri- with an abstract in English were screened. mary or substantive focus of the article, 3) the quantitative relationship between HIV or HIV-related factors and the Data extraction and management mental health condition(s) was discussed, and 4) the study Articles were screened and data extracted by one reviewer reported that participants were conflict-affected and from (EK), with uncertainty resolved through discussion with a a conflict-affected setting. All age groups were included in second reviewer (CK). A third reviewer verified all data this review. All study designs were considered as long as presented in Tables 1 and 2. First, titles and abstracts iden- the four inclusion criteria were met. Articles were ex- tified through the search strategy were screened. Full text cluded if they measured HIV-related factors on a war articles were obtained for all selected abstracts. An eligibil- events scale but did not present data for the relationship ity form was completed to determine final study selection. between the HIV-related factor alone and mental health Data were extracted using a standardized data extraction Koegler and Kennedy Conflict and Health (2018) 12:20 Page 3 of 22 Table 1 Description of included studies Primary author Country Study design Sampling strategy Sample size and Participant characteristics and Year participation rate Mental health and HIV serostatus/HIV-related outcomes Adedimeji et Rwanda Cross-sectional Non random selection N = 928 Women over age 15 who experienced the al., 2015 Baseline data from 2005 RWISA HIV+ and HIV- women approximately 50% 99% of N = 936 included 1994 genocide, 76% HIV+ prospective cohort of whom experienced rape during the 20.5% < 30 years genocide 48.4% 30–40 years 31.1% 40+ years 100% female Adler et al., USA Time-series Non random selection N = 647 Active duty USA soldiers in a brigade combat 2011 Time 1: 4 months after return from Part of a larger study on post deployment 39% of N = 1651 included team who had returned from a 15-month deployment transition who completed both deployment in Iraq Time 2: 4 months later assessments Age not reported 96% male 4% female Kinyanda et Uganda Cross-sectional Random selection N = 1560 Vulnerable (widows, orphans, single mothers) al., 2012 Nested in study on HIV-related psychiatric Multistage sampling to include vulnerable 98.5% of N = 1584 included and non-vulnerable individuals in a war- & psychosocial vulnerabilities in war- and non-vulnerable individuals who completed the affected community affected community interview Aged 15 years and older 56% were aged between 18 and 44 years 43% male 57% female Kinyanda et Uganda Cross-sectional Random sampling N = 1110 Vulnerable (widowed, divorced, orphan, al., 2016 Nested in study addressing HIV-related Multistage sampling for representative 71.2% with complete data suffered torture, mental illness, etc.) and non- psychiatric and psychosocial vulnerabilities sample of vulnerable and non-vulnerable included of N = 1560 vulnerable individuals in a war-affected in the war-affected community individuals community Aged 15 years and older 56% were aged between 18 and 44 years 43% male 57% female Malamba et Uganda Cross-sectional Random selection N = 2388 Conflict affected individuals aged 13–49 al., 2016 Baseline data from a longitudinal cohort Two-stage stratified sampling for 97.5% who had HIV results 29.1% 13–19 years study to determine HIV prevalence and risk representative sample included of N = 2449 20.2% 20–24 years factors to inform program development consenting individuals 19.6% 25–29 years 12.4% 30–34 years 18.6% 35+ years 40% male 60% female Svetlicky et al., Lebanon Cross-sectional Non random selection N = 180 Mean age = 29.95 years (SD = 5.82; range = 20 2010 Collected 6 months post-conflict, collected Combat reserve soldiers who sought 65.7% of N = 274 included to 54 years). for 4 months. treatment in the Combat Reaction Unit in who completed 100% male the wake of the Second Lebanon War questionnaires Most were Israeli-born (82.8%) Talbot et al., Rwanda Time-series Random selection N = 120 94% were orphaned from the genocide 2013 Collected at baseline, 5, 9, and 12 months Orphans selected via random number 95% of N = 120 completed Mean age = 18 years (range 15–25) generation from a list of all eligible orphans all 4 assessments; all Male 47% enrolled in program participants were included Female 53% in analysis Koegler and Kennedy Conflict and Health (2018) 12:20 Page 4 of 22 Table 1 Description of included studies (Continued) Primary author Country Study design Sampling strategy Sample size and Participant characteristics and Year participation rate B.E. Cohen et USA Retrospective cohort Non random selection N = 71,504 Veterans of Operations Enduring and Iraqi al., 2012 From a roster of all USA veterans from 2 Separated USA veterans who were new Freedom operations users of Department of Veterans Affairs Mean age = 28.5 to 29.5 healthcare 100% female Sexual violence and mental health outcomes Amone P’Olak Uganda Cross-sectional Random selection N = 539 Aged between 18 and 25 years et al., 2013 Baseline data nested in a before and after War-affected youth who had been 83% of N = 650 who were 61% male study abducted and lived in rebel captivity for at invited to the study 39% female least 6 months 86% Acholi ethnic group Roberts et al., Uganda Cross-sectional Random selection N = 1210 Adults living in camps for internally displaced 2008 Multi-stage cluster sampling of camps, persons administrative zones, and individuals Mean age = 35.3 years 40% male 60% female 91% Acholi ethnic group Nakimuli- Uganda Time series Non random selection N = 375 Demographic data reported all patients N = 2868, Mpungu et al., Collected at baseline, 3, and 6 months Analysis included only adults with a history 59% of N = 631 included many of whom were not included in the main 2013 of war-related traumatic experiences who were present for at analysis least 2 visits Mean age adult men = 34.5 Mean age adult women = 37.3 47% male 53% female Okello et al., Uganda Case control Random selection N = 153 War affected adolescents 2007 Cross- sectional, unmatched Systematic recruitment, every 3rd name at Formerly abducted Boys mean age = 15.5 years Cases were formerly abducted youth 2 sites: a children’s support organization N =82 Girls mean age = 15.2 years Controls were non abducted youth (case) and a mixed boarding school Non-abducted N =71 Cases: 64% male; 36% female (control) Controls: 61% male; 39% female 100% of controls in secondary school, 12.2% of cases in secondary school Betancourt, Sierra Leone Prospective cohort Non random selection N = 152 Former child soldiers Agnew-Blais et Collected at baseline and time 2 Two stage method: 1) master list of youth in 60% of N = 260 interviewed Mean age = 17.4 years al., 2010 care 2) Invited youth between ages 10–18 at both times 89% male with contact information 11% female Betancourt et Sierra Leone Cross-sectional Non random selection N = 273 Former child soldiers al., 2011 Partially nested in a longitudinal study Longitudinal participants from those who N = 146 from longitudinal Mean age = 16.55 (SD 2.61) participated in one follow up visit, new study and 71% male participants recruited with NGO outreach N = 127 newly recruited for 29% female lists study (50% male, 50% female) Betancourt, Sierra Leone Prospective cohort Non random selection N = 156 Former child soldiers Borisova et al., Collected at baseline and time 2, Two stage method: 1) master list of youth 60% of N = 260 interviewed Mean age = 15.13 years 2010 approximately 2 years later in care 2) Youth aged 10–18 who did not at both times 88% male have a severe disability participated 12% female Sierra Leone Prospective cohort Non random selection N = 260 Former child soldiers Koegler and Kennedy Conflict and Health (2018) 12:20 Page 5 of 22 Table 1 Description of included studies (Continued) Primary author Country Study design Sampling strategy Sample size and Participant characteristics and Year participation rate Betancourt, Collected at baseline (2002), time 2 (2004), Sample from a master list of youth assisted 56.5% (N = 147) assessed at Mean age at time 1 = 15.13 (SD = 2.22) Brennan et al., and time 3 (2008) by program. Youth aged 10–17 with time 2 89% male 2010 contact information invited to participate. 68.8% (N = 179) assessed at 11% female time 3 Johnson et al., Liberia Cross-sectional Random selection N = 1666 Adults in Liberia; 1/3 were former combatants 2008 Population based multi stage random 98.2% of N =1696 Mean age = 41 years cluster of households attempted interviews 47.2% male 52.8% female Johnson et al., Democratic Cross-sectional Non random selection N = 998 Adults in conflict-affected provinces and 2010 Republic of Accessible population based cluster (some 98.9% of N =1005 districts Congo originally selected villages were inaccessible households surveyed Mean age = 40.1 years due to weather and security concerns) 40.6% male 59.4% female Johnson et al., Kenya Cross-sectional Random sampling N = 916 Adults in Kenya 2014 Systematic sampling of 90 villages and 10 95.8% of N = 956 households Mean age = 37.3 years households to assess election-related samples 40% male violence 60% female Cardozo et al., Kosovo Cross-sectional Random selection N = 1358 Kosovar ethnic Albanians aged 15+ years 2000 Two-stage cluster sampling Only women included in 45.3% 15–34 relevant analysis, N = 825 34.1% 35–54 10.9% 55–64 9.7% 65+ 37.7% male 62.3% female Sabin et al., Guatemalan Cross-sectional Non random selection N = 170 Adults and children in Mayan refugee camps 2003 refugees Convenience sample of 5 camps; all 93% of N = 183 households Mean age = 37.9 years living in households sampled in 4 camps, every 3rd 42% male Mexico house in 1 camp 58% female Wolfe et al., USA Retrospective cohort Non random selection N = 160 Returned veterans of the Persian Gulf War 1998 Nested in longitudinal study. Baseline Included women who completed the 66.7% of N = 240 women Mean age = 28.2 years (SD = 6.8) within 5 days of return from deployment, mailed sexual harassment questionnaire assessed at baseline 100% female time 2 18–24 months later. Washington et USA Cross-sectional Random selection N = 3598 Veterans who had been called to duty al., 2013 Pertinent result presented as case (PTSD) Population-based stratified sample 99.6% of N =3611 Mean age = 46.8 (SD = 17.3) for PTSD positive control (no PTSD) Included those who completed the PTSD and 57.4 (SD = 17.0) for PTSD negative screener women 100% female Kang et al., USA Case control Random selection N = 11,441 Gulf War veterans 2005 Nested data from a population based Stratified sample to include each subgroup 76.3% of N = 15,000 sampled Mean age: survey of military personnel Females: with PTSD = 39.1; without PTSD = 38.1 Cases: PTSD Males: with PTSD 40.4; without PTSD 39.6 Controls: did not meet criteria for PTSD 81.4% male 18.6% female HIV acquisition/disease progression and mental health outcomes Koegler and Kennedy Conflict and Health (2018) 12:20 Page 6 of 22 Table 1 Description of included studies (Continued) Primary author Country Study design Sampling strategy Sample size and Participant characteristics and Year participation rate Epino et al., Rwanda Cross-sectional Non random selection N = 610 HIV-positive adults who initiated lifelong ART 2012 From a prospective cohort Patients from clinics Mean age = 38 (SD = 10) 38% male 62% female Mean CD4 count =214 (SD = 92) Mugisha, Uganda Cross-sectional Random selection N = 2361 Adult residents of 3 of the most war affected Muyinda, Baseline data from a project delivering a Two-stage cluster sample stratified at the 98% with complete data of districts Wandiembe et kinship intervention for post-conflict sub-county N = 2406 23.5% 18–24 years al., 2015 mental health 27.3% 25–34 years 20.8% 35–44 years 28.5% 45–54 years 37.5% male 62.5% female Mugisha, Uganda Cross-sectional Random selection N = 2361 Adult residents of 3 of the most war affected Muyinda, Nested in project delivering a kinship Multistage sampling for a representative 98% who had complete districts Malamba et intervention for post-conflict mental health sample from 3 districts data included of N = 2406 23.8% 18–24 years al., 2015 27.1% 25–34 years 20.7% 35–44 years 28.4% 45+ years 37.5% male 62.5% female Muldoon et Uganda Cross-sectional Non random selection N = 129 Formerly abducted by the Lords Resistance al., 2014 From a larger community-based study of Recruited through peer/sex worker led Army sex workers outreach in bars and hotels, and Median age = 22 years (IQR:20–26) community-led outreach to former IDP 100% female camps 96.1% from Acholi tribe M.H. Cohen et Rwanda Cross-sectional Non random selection N = 850 HIV-positive and HIV-negative women al., 2009 Baseline data from a prospective cohort Mainly recruited by Rwandan women’s 91% of N = 936 with About half of each group experienced study associations available mental health data genocidal rape Mean age = 36.4 100% female M.H. Cohen et Rwanda Prospective cohort Non random selection N = 698 HIV-positive and HIV-negative women al., 2011 Baseline, 6, 12, and 18 months later Recruited from Rwandan women’s 74.6% of N = 936 who 50% of each group experienced genocidal associations and HIV clinics in Kigali completed baseline HTQ rape and at least 1 post-baseline Mean age = 36.7 (SD = 8.3) HTQ 100% female Other associations between mental health and HIV acquisition and disease progression Gard et al., Rwanda Cross-sectional Non random selection N = 922 HIV-positive and HIV-negative women 2013 Baseline data nested in a prospective Recruited Rwandan women’s associations 98.5% of N = 936 women 50% of each group experienced genocidal cohort study and clinical sites for HIV patients who completed the Health- rape Related Quality of Life 20.8% under 30 years measure 48.4% aged 30–40 years 30.8% over 40 years 100% female Cross-sectional Non random selection N = 315 Conflict-affected adult women Koegler and Kennedy Conflict and Health (2018) 12:20 Page 7 of 22 Table 1 Description of included studies (Continued) Primary author Country Study design Sampling strategy Sample size and Participant characteristics and Year participation rate Kohli et al., Democratic Baseline data from a randomized Included if provided family rejection 1.9% 16–19 years 2014 Republic of community trial information and had experienced at least 1 14.6% 20–24 years Congo traumatic event in the past 10 years 28.25% 25–34 years 22.54% 35–44 years 29.52% 45–60 years 3.17% over 60 years 100% female Sinayobye et Rwanda Cross-sectional Non random selection N = 710 HIV+ women over age 15, ART naïve al., 2015 Baseline data from 2005 RWISA HIV+ women, approximately 50% of whom Mean age = 34.9 ± 7.0 prospective cohort experienced rape during the genocide 100% female ART Antiretroviral therapy, HIV Human Immunodeficiency Virus, HTQ Harvard trauma questionnaire, IDP Internally displaced person, PTSD Post traumatic stress disorder, USA United States of America Koegler and Kennedy Conflict and Health (2018) 12:20 Page 8 of 22 Table 2 Study outcomes for association between mental health and HIV risk First author & Mental Mental health scales HIV risk measures Results Year health disorders Mental health and HIV serostatus/HIV-related outcomes Adedimeji et al., Depression Center for Epidemiologic Studies Depression HIV serostatus Depression (p < 0.001) but not PTSD (p = 0.06) was related 2015 Rwanda PTSD Scale (CES-D) Had sex last 6 months to HIV serostatus Harvard Trauma Questionnaire (HTQ) Condom use at least 50% of time last 6 months Depression (p = 0.002) but not PTSD (p = 0.09) was History of ever exchanging sex for cash or help related to sex in the last 6 months; women who had History of a non-HIV STI sex did not have different odds of depression scores between 16 and 26 (OR = 0.88, CI 0.64, 1.22) but had decreased odds of depression scores 27+ (OR = 0.57, CI 0.04, 0.81) and no different odds of symptomatic PTSD (OR = 0.78, CI 0.60, 1.03) Depression (p = 0.04) and PTSD (p = 0.006) were related to 50% condom use in the last 6 months; women who used condoms had greater odds of depression scores between 16 and 26 (OR = 1.84, CI 1.20, 2.82) but not scores 27+ (OR = 1.36, CI 0.87, 2.54) and decreased odds of symptomatic PTSD (OR = 0.60, CI 0.42, 0.86) Depression (p = 0.02) and PTSD (p = 0.003) were related to exchange sex; women who had exchanged sex had greater odds of depression scores between 16 and 26 (OR = 1.82, CI 1.19, 2.77) and 27+ (OR = 1.74, CI 1.10, 2.76) and greater odds of being symptomatic for PTSD (OR = 1.68, CI 1.19, 2.36) Depression (p = 0.04) but not PTSD (p = 0.74) related to history of a non-HIV STI; women with a non-HIV STI had greater odds of depression scores between 16 and 26 (OR = 2.02, CI 1.39, 3.09; AOR = 1.64, CI 1.01, 2.65) but not depression scores of 27+ (OR = 1.50, CI 0.94, 2.41; AOR = 1.11, CI 0.65, 1.89) nor symptomatic PTSD (OR = 1.07, CI 0.77, 1.50) Adler et al., PTSD PTSD Checklist (PCL) Risked STD by having unprotected sex PTSD at time 1 predicted sex without a condom four 2011 USA months later (OR = 1.57, CI 1.20, 2.04) Kinyanda et al., Depression Hopkins Symptom Checklist (HSCL-15) High risk sexual behaviors: High-risk sexual behavior was marginally related to MDD 2012 Uganda sex outside marriage; amongst males in univariate analysis (OR = 1.61, 95% CI sex in exchange for gifts; 0.99–2.62, p = 0.06) but not females (OR = 1.17, 95% CI sex in exchange for money; 0.68–2.01, p = 0.57). sex in exchange for protection; High-risk sexual behavior was related to MDD amongst sex with an older person; males (OR = 1.70, 95% CI 1.01–2.86, p = 0.05) in sex with someone known for less than a day; multivariable analysis but not females (OR = 1.03, 95% CI sex with uniformed personnel; 0.59–1.80, p = 0.91). sex with more than one partner Kinyanda et al., Depression HSCL-25 Sexual intimate partner violence (IPV) (‘force you to have Females who experienced sexual IPV had greater odds of 2016 Uganda sex when you don’t want to’) probable MDD (AOR = 4.20, CI 1.54, 11.46) Malamba et al., Depression HSCL-25 HIV serostatus Those with MDD symptoms had greater odd of testing 2016 Uganda PTSD HTQ positive for HIV (UOR = 2.70, CI 1.95, 3.75; AOR = 1.89, CI 1.28, 2.80) Koegler and Kennedy Conflict and Health (2018) 12:20 Page 9 of 22 Table 2 Study outcomes for association between mental health and HIV risk (Continued) First author & Mental Mental health scales HIV risk measures Results Year health disorders Those with PTSD symptoms had greater odds of testing positive for HIV (UOR = 1.90, CI 1.30, 2.78; AOR = 1.44, CI 1.06, 1.96) Svetlicky et al., PTSD PTSD Inventory Risky sexual activities (3 items including sex without No relationship was found between PTSD and risky sexual 2010 Lebanon protection against sexually transmitted diseases) activities Talbot et al., PTSD PCL Laboratory STI testing Rates of STI were too low to evaluate associations with 2013 Rwanda HIV risk taking behavior: PTSD make any conclusions. Exchanging sex for drugs, money, or favors; Higher PTSD symptoms correlated with increased HIV risk- Having sex with an HIV-infected or status unknown taking behavior (r = 0.24, p = 0.006) at baseline. partner; PTSD symptoms were related to baseline HIV risk (0.01, Having two or more sexual partners within the past p = 0.002) in a growth model; for each 1 point increase 3 months of trauma symptoms there was a 0.01 unit increase in baseline HIV risk B.E. Cohen et al., Depression ICD-9-CM diagnostic codes Sexually transmitted infections: cervical dysplasia; All STIs except chlamydia were associated with PTSD. 2012 USA PTSD genital herpes; Cervical dysplasia AOR = 1.86 (CI 1.61–2.16), Comorbid genital warts; Genital herpes AOR = 1.69 (CI 1.36–2.08), depression chlamydia; Genital warts AOR = 1.83 (CI 1.45–2.31), and PTSD gonorrhea; Chlamydia AOR = 1.66 (CI 0.93–2.96), trichomonas; Gonorrhea AOR = 3.12 (CI 1.51–6.44), and other STIs Trichomonas AOR = 1.60 (CI 1.08–2.39), Other STIs AOR = 1.83 (CI 1.52–2.21) All STIs were associated with depression. Cervical dysplasia AOR = 2.35 (CI 2.12–2.59), Genital herpes AOR = 2.51 (CI 2.20–2.87), Genital warts AOR = 2.44 (CI 2.09–2.86), Chlamydia AOR = 2.21 (CI 1.49–3.27), Gonorrhea AOR = 3.99 (CI 2.38–6.71), Trichomonas AOR = 2.38 (CI 1.85–3.06), Other STIs AOR = 2.21 (CI 1.95–2.53) All STIs were most strongly associated with comorbid PTSD and depression. Cervical dysplasia AOR = 2.65 (CI 2.41–2.91), Genital herpes AOR = 2.55 (CI 2.24–2.91), Genital warts AOR = 2.97 (CI 2.56–3.43), Chlamydia AOR = 2.58 (CI 1.80–3.70), Gonorrhea AOR = 4.74 (CI 2.91–7.71), Trichomonas AOR = 3.75 (CI 3.01–4.66), Other STIs AOR = 2.92 (CI 2.59–3.28) Sexual violence and mental health outcomes Amone-P’olak et Depression Acholi Psychosocial Assessment Instrument Sexual abuse measured by one item in the War Trauma Sexual abuse (β = 0.32, SE = 0.16, p < 0.001) predicted al., 2013 and anxiety (APAI) Screening scale symptoms of depression and anxiety for female but not Uganda male youths in multivariate analysis. Roberts et al., PTSD HTQ Rape or sexual abuse Those who reported rape or sexual abuse had greater 2008 Uganda odds of PTSD symptoms (AOR = 1.76, CI 1.01, 2.75) but not depression symptoms (NR) Koegler and Kennedy Conflict and Health (2018) 12:20 Page 10 of 22 Table 2 Study outcomes for association between mental health and HIV risk (Continued) First author & Mental Mental health scales HIV risk measures Results Year health disorders Nakimuli- Depression Self- reporting questionnaire (SRQ-20) Experienced sexual violence Experiencing sexual violence was significantly related to Mpungu et al., PTSD HTQ HIV serostatus PTSD symptom scores (β = 3.75, SE = 1.01, p < 0.05) but 2013 Uganda not depression symptom scores (β = 0.54, SE = 0.45). Being HIV-positive was not significantly related to depression (β = 0.51, SE = 0.43) or PTSD (β = −1.41, SE = 0.94) scores. Okello et al., Depression MINI-KID Sexual torture (undefined) Quantitative results not presented in a table, but the 2007 Uganda Anxiety Being forced to marry stated that no trauma event (including sexual torture and PTSD being forced to marry) showed any significant relationship with any diagnosis of PTSD, major depression and generalized anxiety disorder. Betancourt, Depression A measure developed by the Oxford Refugee Rape as part of Child War Trauma Questionnaire Surviving rape predicted an increase in depression over time Agnew-Blais, et and anxiety Studies Program for use among former child (b =2.58, p = 0.01) after controlling for demographic and al., 2010 Sierra soldiers includes a subscale for anxiety, war-related experiences. When perceived discrimination was Leone depression, and hostility included, the strength of the relationship between rape and depression is reduced, (b = 1.65, p = 0.08). When protective factors were added, there was no longer a relationship between rape and depression. Surviving rape was significantly associated with higher levels of anxiety (b = 5.35, p < 0.001) even after perceived discrimination and protective factors were controlled for. Betancourt et Depression HSCL-25 Rape as part of Child War Trauma Questionnaire No significant relationship between rape and al., 2011 Sierra and anxiety depression after controlling for multiple variables b =2.42 Leone (CI -0.99, 5.84). Rape was significantly related to anxiety b = 2.85 (CI 0.45, 5.26, p = 0.05). A smaller percentage of boys experienced rape (5%) compared to girls (44%), but the effect of rape on anxiety was significant among male child soldiers and not for females (b = −6.42, p = 0.05). Betancourt, Depression Oxford Refugee Studies Program measure for Rape as part of Child War Trauma Questionnaire Rape was correlated to depression symptoms (r = 0.24, Borisova, et al., and anxiety use among former child soldiers p ≤ 0.01) and anxiety symptoms (r = 0.38, p ≤ 0.001). 2010 Sierra Rape was not predictive of depression at T2, adjusting for Leone all covariates (b = 1.74, CI -0.53, 4.00). Rape was the strongest predictor of anxiety at T2 controlling for anxiety levels at T1 (b = 4.06, CI 1.49, 6.62, p < 0.05) and adjusting for all other covariates. Betancourt, Depression Oxford Refugee Studies Program measure for Rape as part of Child War Trauma Questionnaire Rape was associated with higher baseline levels of Brennan et al., and anxiety use among former child soldiers internalizing problems (depression/anxiety) (b =4.60, p < 0.05). 2010 Sierra After adjusting for all hardship and protective factors, among Leone time-invariant predictors, only being raped remained significantly related to depression/ anxiety (b =4.34, p = 0.039). Johnson et al., Depression Patient Health Questionnaire 9 Sexual violence defined as any violence, physical or Adults who experienced sexual violence were more likely 2008 Liberia PTSD PTSD Symptom Scale Interview (1 month psychological, carried out through sexual means or by to meet criteria for PTSD (69% vs. 38%, p < 0.001) and recall) targeting sexuality and included rape and attempted rape, MDD (57% vs. 37%, p = 0.002) compared to adults who molestation, sexual slavery, being forced to undress or did not experience sexual violence. Koegler and Kennedy Conflict and Health (2018) 12:20 Page 11 of 22 Table 2 Study outcomes for association between mental health and HIV risk (Continued) First author & Mental Mental health scales HIV risk measures Results Year health disorders being stripped of clothing, forced marriage, and insertion The weighted prevalence of PTSD (81% vs. 46%, p < 0.001) of foreign objects into the genital opening or anus, forcing and MDD (64% vs.42%, p = 0.003) was higher among male 2 individuals to perform sexual acts on one another or former combatants who had experienced sexual violence harm one another in a sexual manner, or mutilating a compared to those who had not. person’s genitals. The weighted prevalence of PTSD (74% vs. 44%, p = 0.005) was higher but not MDD (63% vs.55%, p = 0.51) among female former combatants who experienced sexual violence compared to those who had not. Noncombatant sexual violence was not related to MDD (32% vs. 29%, p = 0.73) nor PTSD (39% vs. 36%, p = 0.74) for men nor MDD (48% vs. 36%, p = 0.15) nor PTSD (56% vs.36%, p = 0.09) for women. Those who experienced lifetime sexual violence had 1.39 (p = 0.04) the odds of MDD and 2.67 (p < 0.001) the odds of PTSD compared to those who did not experience sexual violence. Johnson et al., Depression Patient Health Questionnaire–9 Sexual violence – defined above The prevalence of MDD was significantly higher for those 2010 PTSD PTSD Symptom Scale Interview (PSS-I) who experienced sexual violence (60.4%) compared to those Democratic who did not experience sexual violence (30.7%, p < 0.001); Republic of The prevalence of PTSD was significantly higher for those Congo who experienced sexual violence (70.2%) than those who did not experience sexual violence (40.3%, p < 0.001) . The prevalence of MDD for females who experienced conflict-related sexual violence was significantly higher (67.7%) than for those who did not experience conflict- related sexual violence (30.3, p <0.001). The prevalence of PTSD for females who experienced conflict-related sexual violence was significantly higher (75.9%) than for those who did not experience conflict- related sexual violence (44.4%, p < 0.001). There were no differences in the prevalence of MDD (47.5% vs. 36.3%, p = 0.18) or PTSD (56% vs. 41.7%, p =0.17) for men who did and did not experience conflict-related sexual violence. There were no differences in the prevalence of MDD (50.7% vs. 38.4%, p = 0.38) or PTSD (61.5% vs. 44.1%, p = 0.34) for men nor of MDD (72.9% vs. 40.1%, p = 0.07) or PTSD (83.6% vs. 52.4%, p = 0.06) for women who experienced community based sexual violence. Johnson et al., Depression Patient Health Questionnaire–9 Sexual violence – defined above 31% of those who experienced sexual violence had 2014 Kenya PTSD PSS-I anxiety and depression before the 2007 election, 45% who experienced sexual violence had anxiety and depression during the election, and 33.7% who experienced sexual violence had anxiety and depression after the 2007 election. The weighted prevalence of MDD (41.0%, CI 27, 55 vs. 35.0%, CI 29.2, 40.8) and PTSD (40.1%, CI 28.6, 51.6 vs. Koegler and Kennedy Conflict and Health (2018) 12:20 Page 12 of 22 Table 2 Study outcomes for association between mental health and HIV risk (Continued) First author & Mental Mental health scales HIV risk measures Results Year health disorders 30.9%, CI 25, 36.8) were not significantly different between those who reported sexual violence and those who did not report sexual violence. Cardozo et al., PTSD HTQ Rape Rape was not related to PTSD symptoms: 2000 Kosovo 21.6% or women who reported rape had symptoms of PTSD vs. 16.92% of women who did not report rape, p =0.49; AOR = 1.68, CI 0.69, 4.08 Sabin et al., Depression HSCL-25 Sexual abuse or rape reported as traumatic event Sexual abuse or rape was independently associated with 2003 Anxiety HTQ anxiety (p = 0.02) but sexual abuse did not remain Guatemalan PTSD significant in the full model. refugees living All rape survivors (N = 6, 100%) experienced anxiety. in Mexico Sexual abuse or rape was not related to PTSD or depression. Wolfe et al., PTSD Mississippi Scale for Combat-related PTSD Sexual assault defined as a sexual experience that was Women who were sexually assaulted experienced a 1998 USA unwanted and involved the use or threat of force significant 18.9 point increase in PTSD scores (M = 91.83, (attempted or completed rape) either by strangers or SD = 22.69) compared to women with no sexual people you knew harassment (M = 71.36, SD = 17.53). Women who were sexually assaulted had increased risk for PTSD compared to women who were only physically (12.5 point difference) or verbally (15.9 point difference) harassed. Washington et PTSD 7-item screen for DSM IV PTSD History of military sexual assault Women with PTSD were significantly more likely to have al., 2013 USA had experienced sexual assault in military (43% vs. 5.1%, p <0.001). Kang et al., 2005 PTSD PCL Sexual assault Among female (AOR = 5.41; 95% CI 3.19, 9.17) and male USA (AOR = 6.21 CI 2.26, 17.04) veterans, sexual assault was significantly associated with PTSD even while controlling for other covariates. HIV acquisition/disease progression and mental health outcomes Epino et al., Depression HSCL-15 CD4 count There was not a significant difference in depression for 2012 Rwanda those with <=200 CD4 cell count (25.5) and > 200 CD4 count (26) (p = 0.58). Mugisha, PTSD Mini-International Neuropsychiatric Interview HIV status Those reporting HIV+ status had greater odds of having Muyinda, (MINI) Sexual trauma events PTSD (UOR = 2.09, CI 1.48, 2.95) Wandiembe et Those who experienced 1–2 sexual trauma events had al., 2015 greater odds of having PTSD in the unadjusted (UOR = 2.6, Uganda CI = 1.63, 4.15) but not the adjusted (AOR = 1.23, CI 0.73, 2.07) model Those who experienced 3+ sexual trauma events had greater odds of having PTSD (UOR = 5.65, CI 3.33, 9.61; AOR = 2.02, CI 1.08, 3.76) Mugisha, Depression MINI HIV status Muyinda, High risk sexual behaviors Koegler and Kennedy Conflict and Health (2018) 12:20 Page 13 of 22 Table 2 Study outcomes for association between mental health and HIV risk (Continued) First author & Mental Mental health scales HIV risk measures Results Year health disorders Malamba et al., Receiving HIV treatment HIV+ status was related to MDD (UOR = 2.85, CI 2.04, 3.96), 2015 Uganda after adjusting for sex and age (AOR = 2.63, CI 1.87, 3.70), and in the multivariate model (OR = 1.83, CI 1.22, 2.74) High risk sexual behavior was not related to MDD in the unadjusted (UOR = 1.13, CI 0.77, 1.67) or adjusted model (AOR = 1.37, CI 0.91, 2.09) Receiving HIV treatment was related to MDD in the adjusted model (AOR = 3.22, CI 1.08, 9.57) but not the unadjusted model (UOR = 2.03, CI 0.85, 4.85) Muldoon et al., Depression APAI All participants had exchanged sex for money or resources For all participants the mean score for the depression 2014 Uganda and anxiety in the previous 30 days sub-scale was 12.84 (SD = 4.79) and the mean score for the anxiety sub-scale was 8.76 (SD = 5.14). No cut off score is defined for symptomatic for either subscale. M.H. Cohen et Depression CES-D About 50% of participants in each group of HIV-positive Women with HIV infection were more likely than HIV- al., 2009 PTSD HTQ and HIV-negative experienced genocidal rape negative women to have clinically significant Rwanda CD4 cell counts depression (81% vs. 65%, p < 0.0001) and MDD (31% vs. 23%, p <0.047). Women with more advanced HIV, indicated by CD4 cell counts < 200 = mL (OR 4.97, CI 2.93, 8.45), were the most likely to have depressive symptoms. Women who had experienced genocidal rape were more likely to have PTSD in unadjusted analyses (OR = 1.63, CI 1.23, 2.15). Depressive symptoms were higher in women who had a history of genocidal rape (OR = 1.56, CI 1.12, 2.16). M.H. Cohen et Depression CES-D About 50% of participants in each group of HIV-positive HIV-positive status was related to increased symptoms of al., 2011 PTSD HTQ and HIV-negative experienced genocidal rape depression (81.5% vs. 63.8%, p < 0.0001), marginally related Rwanda to symptoms of PTSD (59.6 vs. 67.5%, p = 0.081), and not related to MDD (29.2% vs. 22.7%, p = 0.11) compared to HIV-negative status at baseline. There was a continued reduction in PTSD at each follow-up visit for both HIV-positive and HIV-negative groups (6 month change = −0.78, p < 0.0001; 12 month change = − 0.9, p < 0.0001; 18 month change = − 0.84, p < 0.0001). HIV-positive status (b = 0.03, p = 0.38) was not related to PTSD improvement from baseline to 18-month follow up. All participants had fewer depressive symptoms at 18 months follow up compared to baseline (77% vs. 57%). In changes from baseline to visit 4, experiencing genocidal rape was significantly associated with reduced PTSD. Other associations between mental health and HIV acquisition and disease progression Gard et al., 2013 Depression CES-D About 50% of participants in each group of HIV-positive Rwanda PTSD HTQ and HIV-negative experienced genocidal rape Koegler and Kennedy Conflict and Health (2018) 12:20 Page 14 of 22 Table 2 Study outcomes for association between mental health and HIV risk (Continued) First author & Mental Mental health scales HIV risk measures Results Year health disorders HIV-positive women had higher depression scores than HIV- negative participants (23.67, SD = 9.19 vs. 20.79, SD = 9.60, p <0.001). More HIV-positive women met criteria for depression than HIV-negative women (81.46% vs.64.58%, p < 0.001). There was no difference in PTSD scores between HIV- positive and HIV-negative women (2.31, SD = 0.69 vs. 2.4 SD = 0.67, p = 0.09). A greater percentage of HIV-negative compared to HIV- positive women experienced elevated PTSD scores (65.63% vs. 57.8%, p = 0.05). Kohli et al., 2014 Depression HSCL-15 Rape Rape or sexual assault in the past 10 years was related to Democratic PTSD HTQ increased symptoms of PTSD (β = 0.35, p < 0.001) and Republic of depression (β = 0.29, p < 0.001) in multivariate regression. Congo Sinayobye et al., Depression CES-D CD4 count Depression scores were associated with CD4 count (p < 0.001) 2015 Rwanda PTSD HTQ with: CD4 counts > 350 having a mean depression score of 22.4 ± 9.3; CD4 count 200–350 having a mean depression score of 23.0 ± 8.2; CD4 count < 200 having a mean depression score of 25.8 ± 9.1 PTSD scores were not associated with CD4 count (p =0.60) with: CD4 counts > 350 having a median (IQR) PTSD score of 2.1 (1.7–2.7) CD4 count 200–350 having a median (IQR) PTSD score of 2.1 (1.8–2.8) CD4 count < 200 having a median (IQR) PTSD score of 2.2 (1.8–2.8) AOR adjusted odds ratio, ART Antiretroviral therapy, CI confidence interval, MDD major depressive disorder, NR not reported, OR odds ratio, SE Standard Error, SD standard deviation Effect size data are reported where available; textual descriptions of results are reported when that was all that the authors present Koegler and Kennedy Conflict and Health (2018) 12:20 Page 15 of 22 spreadsheet. For each included study the following infor- considered the outcome. Results are organized by the mation was extracted where applicable: citation; location, outcome variable reported by the authors (using the au- setting and target group; study design; sample size; age thors’ expected directionality of the association) as fol- range; gender; random or non-random selection of partici- lows: mental health and HIV serostatus/HIV-related pants; length of follow up; outcome measures; comparison outcomes; sexual violence and mental health outcomes; groups; effect sizes; confidence intervals; significance HIV acquisition/disease progression and mental health levels; measures of HIV risk, mental health conditions and outcomes; and other associations. Under other associa- measures; funding source; and study limitations. To assess tions, we include studies that met the inclusion criteria study quality, we extracted data on study design, sampling but that did not specify the expected directionality of strategy, sample size, and participant characteristics, the associations. Twenty-five studies were from sub- presented in Table 1. These factors were then consid- Saharan Africa, five were from the United States of ered in relation to study quality as presented in the America (USA), and one each was from the Middle East, results and discussion. We did not conduct meta- Europe, and Latin America. Most studies were cross- analysis due to the diversity of populations, study de- sectional, though three time-series studies and seven co- signs, and measured outcomes. hort studies reported longitudinal data. Most studies had non-random sampling of participants, though 14 studies Results utilized a form of random sampling. Overall, 110,818 Of 714 citations identified through the search strategy, 33 participants were included across articles. However, publications were included in this review (Tables 1 and 2). there was overlap in participants across four papers dis- Figure 1 presents a flowchart of the search and screening cussing child soldiers in Sierra Leone [22–25], in five pa- process. Eighteen articles were identified via database pers discussing HIV-positive and negative women, half searching, thirteen through reference searching, and two of whom survived rape during the genocide in Rwanda through searching journal table of contents. [26–30], in two papers from Uganda [31, 32], and in two Table 1 provides information on location, study design, papers from the Wayo-Nero Study in Uganda [33, 34]. sampling strategy, study size, and participant characteris- The smallest study included 120 participants and the lar- tics for each included article. Although most studies gest study included 71,504 participants. were cross-sectional, many articles presented the ex- Table 2 presents each of the included studies by author pected directionality of the relationship based on which and year, country, mental health disorders and measure- variable was considered the exposure and which was ment scales, HIV risk measures, and the relationship Fig. 1 Flow Diagram of review. This flow diagram depicts the search and screening process for the review. Of the 714 records identified through database searching and other sources, 33 articles were ultimately included in this review Koegler and Kennedy Conflict and Health (2018) 12:20 Page 16 of 22 found between mental health and HIV-related outcomes. conflict in Iraq and Afghanistan who had PTSD were Most studies reported mental health outcomes either in significantly more likely to have six of seven different relation to sexual violence or other factors related to STIs compared to veterans without any mental health HIV acquisition and disease progression. Eight studies diagnosis; female veterans with depression were more reported HIV serostatus or HIV risk outcomes. Three likely to have all seven STIs compared to those without studies reported other outcomes but included analyses any mental health diagnosis [39]. Those with comorbid of the association between mental health and HIV ac- depression and PTSD were even more likely to have all quisition and disease progression. In several studies, seven STIs compared to those without any mental health the relationship between HIV-related factors with mental diagnosis with adjusted OR between 2.55 and 4.74. health was not the main objective, but rather one of many This study had the largest population amongst the results reported. included articles (N = 71,504) and represented the entire population of female veterans who met inclu- Mental health and HIV serostatus/HIV-related outcomes sion criteria. Of the eight studies that reported HIV serostatus or Finally, one study from Uganda reported intimate part- HIV-related outcomes, five included HIV sexual risk ner sexual violence as an outcome. In this a cross-sectional behaviors, two reported HIV serostatus, two reported study, females who experienced sexual violence from an in- other STIs, and one reported intimate partner sexual timate partner had greater odds of probable depression violence. (AOR = 4.20, CI 1.54, 11.46) [31]. Four HIV sexual risk behavior studies found some re- Overall, studies reported strong associations between lationship between mental health and HIV risk behav- depression and PTSD and HIV serostatus/HIV-related iors. A study from Uganda found a relationship between outcomes in African and American conflict-affected depression and at least one of eight sexual risk behaviors populations. amongst males in multivariate analysis in a war-affected community (odds ratio (OR) = 1.70, p = 0.05) [32]. In one Rwandan study, increased symptoms of PTSD were cor- Sexual violence and mental health outcomes related with at least one of three HIV risk-taking behav- Twenty-two studies reported mental health outcomes: iors (r = 0.24, p = 0.0006) [35]. In another Rwandan two reported depression only; six PTSD only; six depres- study, women who had exchanged sex had greater odds sion and anxiety; six depression and PTSD; and two de- of depression (OR = 1.74, confidence interval (CI) 1.10, pression, anxiety, and PTSD. A large number of studies 2.76) and PTSD (OR = 1.68, CI 1.19, 2.36) [30]. Women in this review (n = 16) were conducted in sub-Saharan who used condoms at least 50% of the time had de- Africa and reported the relationship between mental creased odds of PTSD (OR = 0.60, CI 0.42, 0.86), but in- health and sexual violence, rape, or sexual abuse (as an creased odds of elevated depression scores (OR = 1.84, HIV-related measure out of the victims’ control). The CI 1.20, 2.82) [30]. Women who had sex in the last six remaining studies in this category were conducted in the months had decreased odds of depression (OR = 0.57, CI USA (n = 3), with Kosovar ethnic Albanians (n = 1), and 0.04, 0.81) [30]. A longitudinal study from the USA with Guatemalan refugees in Mexico (n = 1). showed that PTSD predicted unprotected sex four Most studies found a positive association between sex- months later (OR = 1.57, CI 1.20, 2.04) [36]. In Lebanon, ual violence and poor mental health; three studies found no relationship was found between PTSD and risky sex- no association. Sexual abuse predicted depression and ual activities [37]. anxiety for female Ugandan youth (β = 0.32, p < 0.001) Two studies focused on HIV serostatus as an outcome. [40]. Experiencing sexual violence was related to PTSD Depression (p < 0.001) but not PTSD (p = 0.06) was re- in a study among Ugandan adults (β = 3.75, p < 0.05) lated to HIV positive serostatus among female survivors [41]. Similarly, in another study of Ugandan adults, those of the Rwandan genocide in a cross-sectional, non- who experienced three or more sexual trauma events randomly selected cohort [30]. In a randomly selected had greater odds of PTSD (AOR = 2.02, CI 1.08, 3.67) cross-sectional study in Uganda, men and women with [34]. For Ugandan adults living in camps for internally depression (adjusted odds ratio (AOR) = 1.89, CI 1.28, 2. displaced people, those who reported rape had greater 80) and PTSD (AOR = 1.44, CI 1.06, 1.96) had greater odds of PTSD (AOR = 1.76, CI 1.01, 2.75), but not depres- odds of testing positive for HIV [38]. sion [42]. Amongst adult and adolescent Guatemalan Two studies reported STIs other than HIV as an out- refugees in Mexico, sexual abuse was independently come. Female survivors of the Rwandan genocide who associated with anxiety, but did not remain significant had a non-HIV STI had greater odds of elevated depres- after controlling for other variables [43]. Among Kosovar sion scores (AOR = 1.64, CI 1.01, 2.65), but not PTSD ethnic Albanian women, rape was not related to PTSD (OR = 1.07, CI 0.77, 1.50) [30]. Female USA veterans of (AOR = 1.68, CI 0.69, 4.08) [44]. It is notable that in Koegler and Kennedy Conflict and Health (2018) 12:20 Page 17 of 22 this study, rape was only identified in 4.4% of women longer significant when protective factors were added (N = 60), [44]. [22]. Rape was the only time-invariant predictor found to Three studies by the same study team examined the be related to depression/anxiety (considered together) relationship of sexual violence with depression and after adjusting for hardship and protective factors (b = 4.34, PTSD in different countries. The studies utilized differ- p = 0.04) [25]. Combined, these studies demonstrate that ent methodologies and had different outcomes. Sexual rape (as a more extreme war event) is a strong predictor of violence was related to PTSD and depression in Liberian anxiety over time amongst war-affected adolescents. adults, with prevalence of PTSD higher for male (81% Three publications examined sexual assault and men- vs. 46%, p < 0.001) and female (74% vs. 44%, p = 0.005) tal health in conflict-affected American military popula- former combatants who experienced sexual violence; de- tions. Female veterans who returned from deployment pression was higher for male former combatants who in the Persian Gulf were more likely to experience PTSD experienced sexual violence compared to those who did if they had been sexually assaulted during their service not (64% vs.42%, p = 0.003) [45]. In the Democratic Re- compared to women who did not experience harass- public of Congo, prevalence of PTSD (70.2% vs. 40.3%, ment, who only experienced verbal harassment, and who p < 0.001) and depression (60.4% vs. 30.7%, p < 0.001) experienced physical harassment that was not sexual as- were higher for participants who experienced sexual vio- sault [49]. In cross-sectional analyses, female veterans lence compared to those who did not; prevalence of with PTSD compared to those without PTSD were more PTSD and depression were higher for females who likely to have experienced sexual assault while in the experienced conflict-related sexual violence compared military (43% vs. 5.1%, p < 0.001) [50]. Sexual assault re- to females who did not, respectively (75.9% vs. 44.4%, ports were significantly associated with PTSD for men p < 0.001 and 67.7% vs. 30.3%, p < 0.001) [46]. During (OR = 6.21, CI 2.26, 17.04) and women (OR = 5.41, CI 3. the 2007 election violence in Kenya, the prevalence of 19, 9.17) after controlling for other variables [51]. Find- sexual violence was not significantly related to major ings from these studies supported a relationship between depressive disorder or PTSD [47]. sexual assault and poor mental health. The Liberian study had the strongest methodological Overall, studies reported strong associations between quality (utilizing a population based multistage random sexual violence and all three mental health outcomes. household cluster survey) and the study in the Demo- Sexual violence was most strongly associated with PTSD cratic Republic of Congo had the weakest quality of the [41, 42, 45, 46, 49–51], then depression [22, 24, 25, 40, three (non random accessible population cluster). Al- 45, 46], and then anxiety mainly amongst male survivors though sexual violence in Kenya was not significantly re- [22–25, 40, 43]. lated to major depressive disorder and PTSD, sexual violence was significantly related to suicidal ideation and HIV acquisition/disease progression and mental health suicide attempt [47]. outcomes Sexual torture and being forced to marry were not re- Factors related to HIV acquisition and disease progres- lated to depression, anxiety, and PTSD amongst formerly sion were associated with mental health outcomes in war abducted adolescents in Uganda [48]. However, the two studies that reported HIV sexual risk behavior, one study may not have had the statistical power to detect sig- that reported CD4 count, one that reported receiving nificance forsuchanalysisasthe primaryobjective wasto HIV treatment, two that reported HIV status, and two compare formerly abducted and non-abducted adolescents. that compared HIV-positive and HIV-negative partici- No trauma event had a significant relationship with mental pants and rape during genocide. health outcomes in this randomly selected population. In Uganda, formerly conflict-abducted young women Four studies reported data from a non-randomly se- who recently exchanged sex for money or resources lected prospective cohort of child soldiers in Sierra were assessed for mental health with a locally developed Leone. In two studies with overlapping samples (one measure; mean depression and anxiety scores were 12.84 cross-sectional with 273 participants and one longitu- and 8.76, respectively [52]. Cut off scores defining symp- dinal with 156 participants), rape was significantly re- tomatic versus asymptomatic depression and anxiety were lated to anxiety (b = 2.85, p = 0.05; b = 4.06, p < 0.05), not provided in this study so no conclusion can be made especially in males, but not depression after controlling regarding the association between HIV sexual risk with for other variables [23, 24]. In a slightly different sample, depression and anxiety. In another study in Uganda, rape was associated with increased anxiety after control- adults with high risk sexual behavior did not have higher ling for discrimination and protective factors (b = 5.35, odds of depression (AOR = 1.37, CI 0.91, 2.09) [33]. p < 0.001); rape significantly predicted an increase in Lower CD4 count indicates more advanced HIV disease depression over time, but this relationship lessened progression and potentially increases transmissibility. In when perceived discrimination was added and was no one non-randomly selected population in Rwanda, there Koegler and Kennedy Conflict and Health (2018) 12:20 Page 18 of 22 was no difference in depression scores for adults with increased PTSD (β =0.35, p < 0.001) and depression (β =0. CD4 counts at or below 200 compared to adults with 29, p < 0.001). CD4 counts over 200 [53]. Receiving HIV treatment Overall, studies in this category demonstrated incon- was associated with depression (AOR = 3.22, CI 1.08, sistent associations between mental health and HIV ac- 9.57) among a random selection of adults in Uganda quisition and disease progression. [33]. Reporting an HIV-positive status was associated with both PTSD (OR = 2.09, CI 1.48, 2.95) [34]and Discussion depression (OR = 1.83, CI 1.22, 2.74) [33]among Overall, the thirty-three studies included in this review Ugandan adults. suggest that symptoms of poor mental health are associ- In two publications from another non-randomly selected ated with increased sexual risk behaviors and HIV cohort in Rwanda, depression and PTSD in HIV-positive markers, and HIV exposures and HIV-related health sta- and negative women were examined both cross-sectionally tus are associated with poor mental health symptoms in and longitudinally. At baseline, HIV-positive women were conflict-affected populations. Only five studies found no more likely than HIV-negative women to have depression association between mental health and factors related to symptoms (81% vs. 65%, p < 0.0001) and HIV-positive HIV acquisition and disease progression. Associations women with CD4 counts below 200 were the most likely were strongest for mental health and HIV serostatus/ to have depression symptoms (OR = 4.97, CI 2.93, 8.45) HIV-related outcomes and sexual violence and mental [26]. Women who experienced rape during the genocide health outcomes. Associations between HIV acquisition/ were more likely to have PTSD (OR = 1.63, CI 1.23, 2.15) disease progression and mental health outcomes and and depression (OR = 1.56, CI 1.12, 2.16) in unadjusted outcomes that examined other associations were more analysis [26]. Over three follow-up visits at six, twelve, and inconsistent. This could be partially attributed to the eighteen months, HIV-positive and HIV-negative women greater variety in the types of associations measured in had reduced PTSD scores; this improvement was not re- these categories. For example, studies that examined lated to antiretroviral therapy (ART) use, but was related sexual violence and mental health outcomes only in- to HIV-positive status at follow-up three, and was related cluded associations between sexual violence and mental to rape during the genocide at follow-ups two and three health, given the large number of studies in this cat- [27]. Overall, studies examining mental health outcomes egory. However, studies that examined HIV acquisition/ and factors related to HIV acquisition and disease progres- disease progression and mental health outcomes looked sion reported both positive and null associations. at a variety of HIV-related correlates with mental health outcomes (e.g. CD4 count, sexual risk behaviors, HIV positive serostatus). Other associations between mental health and HIV There were slight inconsistencies across studies in as- acquisition and disease progression sociations. For articles that found an association between Three studies used other outcomes for their primary one or more mental health disorders and factors related analyses but provided data assessing the cross-sectional to HIV acquisition and disease progression, there were association between mental health and HIV acquisition other articles, or findings within the same article, that and disease progression, making less clear the authors’ failed to identify a relationship with one or more other expected directionality of the relationship. One study in disorders. These inconsistencies may represent real dif- Rwanda (with some study population overlap with other ferences in various associations or populations. Incon- included articles) examined quality of life amongst trauma- sistencies in the findings may also be attributed to less affected women with and without HIV. About half of the rigor among some studies that did not find associations: women in each group had experienced rape during the non-random selection and small sample size [37], small genocide. HIV-positive women experienced increased de- numbers of relevant participants within the sample [43, pression symptom scores compared to HIV-negative 48], and low rates of the HIV-related variable [35]. Some women (p < 0.001), but a greater percentage of HIV- studies that did not find associations had trends towards negative women experienced elevated symptoms of PTSD an association [47] or a marginal association [31]. Stud- [28]. In another study in Rwanda among HIV positive ies with strong rigor (large sample sizes of randomly se- women (with study population overlap) depression (p <0. lected participants or the entire population) found strong 001), but not PTSD (p = 0.60), was associated with CD4 positive associations both in conflict-affected USA [39, 50, count [29]. In the Democratic Republic of Congo, family 51] and African populations [31, 38]. rejection amongst survivors of sexual assault was more Our findings are similar to a review of HIV risk behav- strongly related to depression, but increased PTSD was iors and trauma amongst migrants from low and middle- more strongly related to sexual assault [54]. Rape or sexual income countries [18]. Our study differs from the review assault in the past ten years was associated with both by Michalopoulos et al. by including associations between Koegler and Kennedy Conflict and Health (2018) 12:20 Page 19 of 22 mental health and factors related to HIV acquisition and HIV focuses on this geographic area. Similarly, a wide disease progression in conflict-affected populations from range of conflict-affected populations were included in high-income countries, focusing on measurable mental this review, from soldiers fighting in their own country health disorders, and exclusively presenting quantitative to returned American combat veterans, from orphans to relationships. child soldiers, from abducted civilians to those living in We only included mental health symptoms docu- internally displaced persons camps. Unfortunately, there mented by validated scales to ensure that included stud- were not enough studies examining the same factors in ies met a minimal level of measurement rigor and to the range of conflict-affected regions or populations to increase comparability of findings across studies. Most draw strong conclusions as to how these relationships included articles meet the criteria for ‘protracted’ or might differ by region or population. ‘post-conflict’ [21, 47]. However, one study stands out as We used findings from this review along with existing it specified reporting election-related violence rather literature to develop a framework (Fig. 2). Mock et al. than war or conflict; with 1133 deaths over 59 days, the described how HIV prevalence could increase (via in- setting met criteria for an ‘explosive’ event [21, 47]. This creased military and civilian interaction, increased com- demonstrates the range of included conflict-affected set- mercial and casual sex, etc.) or decrease (via increased tings. Because HIV disproportionally affects sub-Saharan isolation, mortality of high-risk populations, etc.) in Africa, much of the literature discussing conflict and conflict-affected settings [13]. It is well established in the Fig. 2 Relationship Between Mental Health & HIV in Conflict Settings. This figure illustrates a framework of the relationship between mental health and HIV serostatus and HIV-related outcomes among conflict-affected populations, based on this review and existing literature. The outside blue panel presents existing knowledge of factors in conflict-affected settings. HIV prevalence could increase or decrease in conflict-settings through various mechanisms [13]. Multiple factors contribute to increased risk for poor mental health in conflict-affected settings. Additional factors adversely affect conflict-affected populations. The inside box presents the relationship between mental health and HIV-related outcomes. HIV can physiologically, psychologically, and socially increase risk for mental health disorders. Health status (being HIV-positive and lower CD4 count) is associated with poor mental health. Poor mental health can influence HIV risk exposures (sexual risk behaviors and STIs). Surviving sexual assault is associated with poor mental health and HIV-related outcomes. Demographic factors can influence each relationship Koegler and Kennedy Conflict and Health (2018) 12:20 Page 20 of 22 literature that conflict-affected populations are at risk health disorders using appropriate cross-culturally vali- for poor mental health [8, 12, 55], and often lack access dated tools. Referrals could then be made, assuming avail- to mental health services [56, 57]. Physiological, psycho- ability of evidence-based interventions. Similarly, programs logical, and social pathways can influence the relation- delivering mental health services should consider screening ship between HIV-related outcomes and mental health for HIV serostatus and associated risk factors. Since health disorders [7]. What is lacking in the literature, to the infrastructure is often limited in conflict-affected settings, best of our knowledge, is a framework that incorporates combining screening and services offers the potential to the relationship between mental health and factors asso- more systematically and holistically treat vulnerable indi- ciated with HIV acquisition and progression in conflict- viduals. An example where this has occurred is in Uganda, affected populations. where an organization that provided mental health inter- Our findings provide evidence for this framework. ventions for conflict survivors included HIV screening, re- Specifically, we identified studies that demonstrate ferrals, and services to meet the unique mental health health status (HIV-positive serostatus, lower CD4 count) needs of people living with HIV [61]. is associated with increased depression [26–30, 38], that At the policy level, by recognizing the relationships be- mental health disorders may influence HIV risk exposures tween mental health and factors related to HIV acquisition (sexual risk behaviors [30, 32, 36] and STIs [30, 39]), and and disease progression in conflict settings, infrastructure that surviving sexual assault may be associated with poor can be integrated to offer mental health and HIV-related mental health [22–25, 40, 41, 43, 45, 46, 49–51]. services simultaneously. Policy could also require moni- There were several limitations to this review. First, toring of results to recognize any differences in separate only one reviewer identified, screened, and extracted treatment compared to integrated treatment of mental data from included studies. Methodological rigor would health and HIV-related services. Future research should be strengthened if two reviewers had independently employ stronger methods where possible – specifically, completed each step and resolved any discrepancies. Al- random selection of participants to decrease bias and though it was not possible to have two reviewers conduct longitudinal studies to better determine directionality of all steps in the review process, a second, experienced re- the measured associations. Research should also examine viewer was consulted when specific questions arose during the associations of mental health with HIV acquisition and the process and a third reviewer verified the extracted disease progression with a wider range of conflict-affected data. Second, the search terms for ‘conflict’ focused only populations, as the relationship may vary depending on on conflict and war. By excluding terms such as refugee, the population and the possibilities for risk exposure. displaced persons, and asylum seekers we may have missed articles relevant to this review. Conclusions We examined only three mental health disorders, se- Existing literature demonstrates that depression, anxiety, lected because they are common and frequently mea- and PTSD have been quantifiably associated with four sured in conflict-affected populations. However, other factors related to HIV acquisition and disease progres- disorders may be relevant in conflict-affected popula- sion in conflict-affected populations: markers of HIV tions, specifically substance use, which has been shown risk (i.e. STIs), HIV-related health status (e.g. CD4 to be common, harmful, and related to HIV transmis- count), sexual risk behaviors, and HIV risk exposures (i. sion and risk in conflict-affected populations [58–60]. Fi- e. sexual violence). Specifically, poor mental health has nally, 23 of the 33 studies were cross-sectional, so the been associated with two outcomes, HIV markers and temporality of these relationships cannot be determined. sexual risk behaviors, while HIV risk exposures and No studies were identified in this review that examined health status have been associated with the outcome of the association between viral suppression and mental poor mental health. Additional research utilizing ran- health; future studies should examine these associations. dom selection and longitudinal design can further estab- Longitudinal studies should also be conducted examin- lish the strength of these associations and determine if ing associations between mental health and factors re- HIV and mental health services need to be integrated for lated to HIV acquisition and disease progression. Future conflict-affected populations. reviews could examine the associations between HIV ac- quisition or risk behaviors and substance use or other Additional file mental health disorders in conflict-affected populations. There are several implications from this review. Con- Additional file 1: Scoping review search terms for PubMed. (DOCX 14 kb) sidering the associations discussed in this paper, pro- grams delivering HIV or other reproductive health Abbreviations services for conflict-affected populations should consider AOR: Adjusted Odds Ratio; ART: Antiretroviral Therapy; CI: Confidence screening individuals who have HIV or STIs for mental Interval; HIV: Human Immunodeficiency Virus; HTQ: Harvard Trauma Koegler and Kennedy Conflict and Health (2018) 12:20 Page 21 of 22 Questionnaire; IDP: Internally Displaced Person; MDD: Major Depressive and displacement: a systematic review and meta-analysis. JAMA. 2009; Disorder; OR: Odds Ratio; PTSD: Posttraumatic Stress Disorder; SD: Standard 302(5):537–49. Deviation; SE: Standard Error; STI: Sexually Transmitted Infection; 13. Mock NB, et al. Conflict and HIV: a framework for risk assessment to prevent UNESCO: United Nations Educational, Scientific and Cultural Organization; HIV in conflict-affected settings in Africa. Emerg Themes Epidemiol. 2004;1(1):1. USA: United States of America 14. Salama P, Dondero TJ. HIV surveillance in complex emergencies. AIDS. 2001; 15:S4–S12. Acknowledgements 15. Spiegel PB, et al. Prevalence of HIV infection in conflict-affected and The authors thank Peggy Gross for assistance with generating the search terms. displaced people in seven sub-Saharan African countries: a systematic We also thank David Celentano for his extensive feedback on an earlier draft review. Lancet. 2007;369(9580):2187–95. and Jessica Haslag for verifying the extracted data. 16. Bennett BW, et al. Hiv incidence prior to, during, and after violent conflict in 36 sub-Saharan african nations, 1990-2012: an ecological study. PLoS One. Funding 2015;10(11):e0142343. This research was supported by NIMH NRSA F31MH095678. 17. Mulanga C, et al. Political and socioeconomic instability: how does it affect HIV? A case study in the Democratic Republic of Congo. AIDS. 2004;18(5):832–4. Availability of data and materials 18. Michalopoulos LM, Aifah A, El-Bassel N. A systematic review of HIV risk The data extracted from each included article is available upon request. behaviors and trauma among forced and unforced migrant populations from low and middle-income countries: state of the literature and future Authors’ contributions directions. AIDS Behav. 2015:1–19. EK extracted the data, coded included articles, analyzed the data, and was 19. Collins PY, et al. What is the relevance of mental health to HIV/AIDS care the primary writer of the manuscript. CK provided substantial contributions and treatment programs in developing countries? A systematic review. to the design of the study, analysis and interpretation of the data, and AIDS. 2006;20(12):1571. provided critical revisions throughout several drafts of the manuscript. Both 20. Moher D, et al. Preferred reporting items for systematic reviews and meta- authors read and approved the final manuscript. analyses: the PRISMA statement. Ann Intern Med. 2009;151(4):264–9. 21. Strand, H. and M. Dahl, Defining Conlict-affected Countries. 2010: Unesco. Ethics approval and consent to participate 22. Betancourt TS, et al. Past horrors, present struggles: the role of stigma in the Not applicable association between war experiences and psychosocial adjustment among former child soldiers in Sierra Leone. Soc Sci Med. 2010;70(1):17–26. Consent for publication 23. Betancourt TS, et al. Sierra Leone's child soldiers: war exposures and mental Not applicable health problems by gender. J Adolesc Health. 2011;49(1):21–8. 24. Betancourt TS, et al. Sierra Leone’s former child soldiers: a follow-up study Competing interests of psychosocial adjustment and community reintegration. Child Dev. 2010; The authors declare that they have no competing interests. 81(4):1077–95. 25. Betancourt TS, et al. Sierra Leone's former child soldiers: a longitudinal study of risk, protective factors, and mental health. J Am Acad Child Adolesc Publisher’sNote Psychiatry. 2010;49(6):606–15. Springer Nature remains neutral with regard to jurisdictional claims in published 26. Cohen MH, et al. Prevalence and predictors of posttraumatic stress disorder maps and institutional affiliations. and depression in HIV-infected and at-risk Rwandan women. J Women's Health (Larchmt). 2009;18(11):1783–91. Received: 24 October 2016 Accepted: 25 March 2018 27. Cohen MH, et al. Improvement in posttraumatic stress disorder in postconflict Rwandan women. J Women's Health (Larchmt). 2011;20(9):1325–32. 28. Gard TL, et al. The impact of HIV status, HIV disease progression, and post- References traumatic stress symptoms on the health-related quality of life of Rwandan 1. Nduna M, et al. Associations between depressive symptoms, sexual behaviour women genocide survivors. Qual Life Res. 2013;22(8):2073–84. and relationship characteristics: a prospective cohort study of young women 29. Sinayobye JdA, et al. Prevalence of shingles and its association with PTSD and men in the eastern cape, South Africa. J Int AIDS Soc. 2010;13(1):1. among HIV-infected women in Rwanda. BMJ Open. 2015;5(3):e005506. 2. Agardh A, Cantor-Graae E, Östergren PO. Youth, sexual risk-taking behavior, 30. Adedimeji AA, et al. Sexual behavior and risk practices of HIV positive and and mental health: a study of university students in Uganda. International HIV negative Rwandan women. AIDS Behav. 2015;19(7):1366–78. journal of behavioral medicine. 2012;19(2):208–16. 31. Kinyanda E, et al. Intimate partner violence as seen in post-conflict eastern 3. Cavanaugh CE, Hansen NB, Sullivan TP. HIV sexual risk behavior among low- Uganda: prevalence, risk factors and mental health consequences. BMC Int income women experiencing intimate partner violence: the role of Health Hum Rights. 2016;16(1):5. posttraumatic stress disorder. AIDS Behav. 2010;14(2):318–27. 32. Kinyanda E, et al. Psychiatric disorders and psychosocial correlates of high 4. Hutton HE, et al. HIV risk behaviors and their relationship to posttraumatic HIV risk sexual behaviour in war-affected eastern Uganda. AIDS Care. 2012; stress disorder among women prisoners. Psychiatr Serv. 2001;52(4):508–13. 24(11):1323–32. 5. Munroe CD, et al. The relationship between posttraumatic stress symptoms 33. Mugisha J, et al. Major depressive disorder seven years after the conflict and sexual risk: examining potential mechanisms. Psychological Trauma: in northern Uganda: burden, risk factors and impact on outcomes (the Theory, Research, Practice, and Policy. 2010;2(1):49. Wayo-Nero study). BMC psychiatry. 2015;15(1):48. 6. Rubin A, Gold M, Primack BA. Associations between depressive symptoms 34. Mugisha J, et al. Prevalence and factors associated with posttraumatic stress and sexual risk behavior in a diverse sample of female adolescents. J Pediatr disorder seven years after the conflict in three districts in northern Uganda Adolesc Gynecol. 2009;22(5):306–12. (the Wayo-Nero study). BMC psychiatry. 2015;15(1):170. 7. Prince M, et al. No health without mental health. Lancet. 2007;370(9590): 35. Talbot A, et al. Treating psychological trauma among Rwandan orphans is 859–77. associated with a reduction in HIV risk-taking behaviors: a pilot study. AIDS 8. De Jong JTVM, et al. Lifetime events and posttraumatic stress disorder in 4 Educ Prev. 2013;25(6):468–79. postconflict settings. JAMA. 2001;286(5):555–62. 9. Hamid AARM, Musa SA. Mental health problems among internally displaced 36. Adler AB, et al. Effect of transition home from combat on risk-taking and health-related behaviors. J Trauma Stress. 2011;24(4):381–9. persons in Darfur. Int J Psychol. 2010;45(4):278–85. 10. Kamau M, et al. Psychiatric disorders in an African refugee camp. 37. Svetlicky V, et al. Combat exposure, posttraumatic stress symptoms and International Journal of Mental health, Psychosocial Work and Counselling risk-taking behavior in veterans of the second Lebanon war. Isr J in Areas of Armed Conflict. 2004;2(2):84–9. Psychiatry Relat Sci. 2010;47(4):276–83. 11. Srinivasa Murthy R. Mass violence and mental health–recent 38. Malamba SS, et al. “The Cango Lyec Project-Healing the Elephant”:HIV epidemiological findings. Int Rev Psychiatry. 2007;19(3):183–92. related vulnerabilities of post-conflict affected populations aged 13–49 12. Steel Z, et al. Association of torture and other potentially traumatic events years living in three Mid-Northern Uganda districts. BMC Infect Dis. with mental health outcomes among populations exposed to mass conflict 2016;16(1):690. Koegler and Kennedy Conflict and Health (2018) 12:20 Page 22 of 22 39. Cohen BE, et al. Reproductive and other health outcomes in Iraq and Afghanistan women veterans using VA health care: association with mental health diagnoses. Womens Health Issues. 2012;22(5):e461–71. 40. Amone-P’Olak K, et al. Cohort profile: mental health following extreme trauma in a northern Ugandan cohort of war-affected youth study (the WAYS study). SpringerPlus. 2013;2(1):1–11. 41. Nakimuli-Mpungu E, et al. Implementation and scale-up of psycho-trauma centers in a post-conflict area: a case study of a private-public Partnership in Northern Uganda. PLoS Med. 2013;10(4) 42. Roberts B, et al. Factors associated with post-traumatic stress disorder and depression amongst internally displaced persons in northern Uganda. BMC psychiatry. 2008;8(1):38. 43. Sabin M, et al. Factors associated with poor mental health among Guatemalan refugees living in Mexico 20 years after civil conflict. JAMA. 2003;290(5):635–42. 44. Cardozo BL, et al. Mental health, social functioning, and attitudes of Kosovar Albanians following the war in Kosovo. JAMA. 2000;284(5):569–77. 45. Johnson K, et al. Association of combatant status and sexual violence with health and mental health outcomes in postconflict Liberia. JAMA. 2008; 300(6):676–90. 46. Johnson K, et al. Association of sexual violence and human rights violations with physical and mental health in territories of the eastern Democratic Republic of the Congo. JAMA. 2010;304(5):553–62. 47. Johnson K, et al. A national population-based assessment of 2007–2008 election-related violence in Kenya. Confl Health. 2014;8:2. 48. Okello J, Onen T, Misisi S. Psychiatric disorders among war-abducted and non-abducted adolescents in Gulu district, Uganda: a comparative study. Afr J Psychiatry (Johannesbg). 2007;10(4):225–31. 49. Wolfe J, et al. Sexual harassment and assault as predictors of PTSD symptomatology among US female Persian gulf war military personnel. Journal of Interpersonal Violence. 1998;13(1):40–57. 50. Washington DL, et al. PTSD risk and mental health care engagement in a multi-war era community sample of women veterans. JGIM: Journal of General Internal Medicine. 2013;28(7):894–900. 51. Kang H, et al. The role of sexual assault on the risk of PTSD among gulf war veterans. Ann Epidemiol. 2005;15(3):191–5. 52. Muldoon KA, et al. After abduction: exploring access to reintegration programs and mental health status among young female abductees in northern Uganda. Confl Health. 2014;8:5. 53. Epino HM, et al. Reliability and construct validity of three health-related self-report scales in HIV-positive adults in rural Rwanda. AIDS Care. 2012;24(12):1576–83. 54. Kohli A, et al. Risk for family rejection and associated mental health outcomes among conflict-affected adult women living in rural eastern Democratic Republic of the Congo. Health Care for Women International. 2014;35(7–9):789–807. 55. Miller KE, Andrew R. War exposure, daily stressors, and mental health in conflict and post-conflict settings: bridging the divide between trauma- focused and psychosocial frameworks. Soc Sci Med. 2010;70(1):7–16. 56. Mollica RF, et al. Mental health in complex emergencies. Lancet. 2004; 364(9450):2058–67. 57. Tol WA, et al. Mental health and psychosocial support in humanitarian settings: linking practice and research. Lancet. 2011;378(9802):1581–91. 58. Ezard N. Substance use among populations displaced by conflict: a literature review. Disasters. 2012;36(3):533–57. 59. Ezard N, et al. Six rapid assessments of alcohol and other substance use in populations displaced by conflict. Confl Health. 2011;5(1):1. 60. Friedman SR, Rossi D, Braine N. Theorizing 'Big Events' as a potential risk environment for drug use, drug-related harm and HIV epidemic outbreaks. Submit your next manuscript to BioMed Central International Journal of Drug Policy. 2009;20(3):283–91. and we will help you at every step: 61. Sharer, M. and M. Gutmann, Prioritizing HIV in mental health services delivered in post-conflict settings. 2011. • We accept pre-submission inquiries � Our selector tool helps you to find the most relevant journal � We provide round the clock customer support � Convenient online submission � Thorough peer review � Inclusion in PubMed and all major indexing services � Maximum visibility for your research Submit your manuscript at www.biomedcentral.com/submit

Journal

Conflict and HealthSpringer Journals

Published: Jun 1, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off