Background
Mental health disorders among adolescents constitute a significant problem worldwide, with estimates ranging from 10 to 20%.Reference Kieling, Baker-Henningham, Belfer, Conti, Ertem and Omigbodun1 Although the prevalence rates are less well-established in low- and middle-income countries (LMICs), estimates ranging from 13% in Brazil and India to 18% in Ethiopia and 15% in South AfricaReference Patel, Flisher, Hetrick and McGorry2 indicate similarly high prevalence rates. With almost 90% of the world's children and adolescents living in LMICs, failure to address mental health problems, which often initiate in adolescence, has wide-reaching implications for other health and developmental concerns such as lower educational achievement, substance use, violence and risk-inducing reproductive and sexual health.Reference Patel, Flisher, Hetrick and McGorry2 Adding to this burden is the growing number of young people living with HIV in LMICs: of the estimated 1.8 million adolescents living with HIV, 85% live in sub-Saharan Africa.3 Prevalence rates of mental health disorders in LMICs are poorly established because of a dearth of studies and of the few that exist, most are cross sectional. Findings from a study in the USA indicate 53% had a psychiatric diagnosis before HIV treatment and 44% experienced ongoing depressive disorders. However, there may be higher rates of undiagnosed depression in youth who are HIV-positive.Reference Mellins, Elkington, Leu, Santamaria, Dolezal and Wiznia4
Relationship between HIV and mental health disorders
Even though mental health disorders are prevalent in adolescents living with HIV, a recent review indicated that not much has been done to measure the impact of these disorders or the interventions that could form a healthcare response to mental health issues.Reference Vreeman, McCoy and Lee5 The relationship between HIV and mental health disorders is bidirectional. HIV exacerbates psychological distress among young people, and young people with mental health disorders are more likely to acquire HIV.Reference Cluver, Orkin, Gardner and Boyes6,Reference Pence7 Not all adolescents living with HIV will have mental health difficulties,Reference Betancourt, Meyers-Ohki, Charrow and Hansen8,Reference Bhana, Mellins, Small, Nestadt, Leu and Petersen9 but a disproportionate number experience emotional and behavioural problems at higher than expected rates, including that of other high-risk groups.Reference Mellins, Elkington, Leu, Santamaria, Dolezal and Wiznia4,Reference Elkington, Robbins, Bauermeister, Abrams, McKay and Mellins10,Reference Mellins and Malee11 adolescents living with HIV who have mental health difficulties are less likely to achieve viral suppression,Reference Yehia, Stephens-Shield, Momplaisir, Taylor, Gross and Dube12 have lower odds of retention in careReference Mayston, Patel, Abas, Korgaonkar, Paranjape and Rodrigues13,Reference Rooks-Peck, Adegbite, Wichser, Ramshaw, Mullins and Higa14 and have an increased risk of AIDS-related mortalityReference Dow, Turner, Shayo, Mmbaga, Cunningham and O'Donnell15 than their peers and other high-risk groups.Reference Mellins, Elkington, Leu, Santamaria, Dolezal and Wiznia4,Reference Elkington, Robbins, Bauermeister, Abrams, McKay and Mellins10,Reference Mellins and Malee11
Adolescents living with HIV face a life-long illness and must confront numerous challenges, which evolve with time. Living with HIV has an impact on multiple areas of adolescent lives: they face difficulties associated with disclosure, stigma and fear of negative reactions from others, including being bullied.Reference Britto, Mehta, Thomas and Shet16,Reference Kim, Mazenga, Yu, Devandra, Nguyen and Ahmed17 They have to cope with the shock, fear, anger, guilt and shame of having a chronic health condition and ongoing stigmatising responses, which further has an impact on their mental health.Reference Remien and Mellins18
Namibian adolescents living with HIV had greater emotional, behavioural and conduct problems compared with case controls, even after controlling for sociodemographics. Notably, after controlling for orphanhood status, mental health outcomes between the two groups were no longer significant.Reference Gentz, Calonge-Romano, Martinez-Arias, Zeng and Ruiz-Casares19 Compared with matched controls, adolescents living with HIV in Cape Town, South Africa had significantly poorer functional competence and self-concept and higher levels of depression, anger and disruptive behaviour.Reference Hoare, Phillips, Brittain, Myer, Zar and Stein20 Mental health symptoms were mostly associated with sociodemographic factors and stressful life events and the loss of both parents was associated with disruptive behaviour. A recent study in Uganda of adolescents living with HIV revealed that even with antiretroviral treatment, 17% had at least one neurological disorder (enuresis/encopresis, motor/verbal tics, epilepsy), which in turn was associated with early onset of sexual intercourse.Reference Mpango, Rukundo, Muyingo, Gadow, Patel and Kinyanda21 These mild-to-moderate neurocognitive disorders persist over time despite being on highly active antiretroviral treatment.Reference Boivin, Ruiseñor-Escudero and Familiar-Lopez22 But despite the heightened risk of mental health difficulties among adolescents living with HIV, evidence suggests that parent–child involvement, as well as communication and social support from peers, parents and teachers, can act as a protective factor.Reference Mellins and Malee11
Adolescents who are not living with HIV but come from a household where someone else is, are also at risk for a range of mental health problems including depression, anxiety and social problems. The relationship between mental health and HIV is modified by multiple family and contextual risk influences such as the role of the caregiver,Reference Elkington, Robbins, Bauermeister, Abrams, McKay and Mellins10 levels of impoverishment, violence, substance misuse and neighbourhood disintegration. These adolescents also experience stigma by association with family members living with HIV, which can drive persistent mental health symptoms of anxiety and depression.Reference Boyes and Cluver23 A systematic review of psychosocial interventions focusing on mental health resilience of young people affected by HIV showed that parenting and family support were important to their mental health and psychosocial adjustment.Reference Betancourt, Meyers-Ohki, Charrow and Hansen8
Aims
Despite this growing evidence, key gaps in knowledge remain. First, the totality of the evidence base of mental health interventions for adolescents living with and affected by HIV remains unknown. Second, despite an understanding of the different mechanisms by which mental health difficulties occur for adolescents living with HIV (i.e. adolescents who were HIV-positive) versus adolescents affected by HIV (i.e. adolescents living in HIV-affected households or who are orphans because the caregiver has died from an AIDS-related illness), it is unknown whether and how the evidence base of interventions for these two populations differs. Third, a specific focus on LMICs is critical because of the limited number of mental health professionals in resource-limited settings, indicating a potential need for innovative and community-based interventions. This systematic review aims to address these gaps by examining the outcomes for interventions that have worked and that have not worked to support the mental health of adolescents in low- and middle-income contexts who are living with or are affected by HIV. Only quantitative studies regardless of the study design were included.
Method
This paper follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).Reference Moher, Liberati, Tetzlaff and Altman24 As the paper is a review of published findings, ethical approval for the review was unnecessary. The protocol for the systematic review was pre-published in PROSPERO in 2018 and is available at http://www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD42018103269.
Review scope
The scope of this review is restricted to intervention studies measuring a wide range of mental health outcomes among adolescents living with HIV or youth affected by HIV in LMICs between 2000 and 2018 and published in English, specifically as follows.
(a) Outcome: any term that refers to mental health or mental illness or emotional or psychological or adjustment or psychosocial or development or distress or trauma or self-harm or self-injury or well-being or social functioning or adaptive behaviours associated with any common or severe mental disorder, including substance use or misuse or drug misuse in adolescents living with HIV or affected by HIV.
(b) Population: adolescents and young people aged 10–24 years, either living with HIV or affected by HIV as a function of familial HIV.
(c) Geographic location: LMICs as defined by the World Bank.
(d) Study design: any type of quantitative evaluation of an intervention study regardless of study design.
Search strategy
The search strategy and actual searches were conducted in June to July 2018 by an information expert (A. Bullen) using electronic databases listed below. The search strategy was piloted in PubMed (which includes Medline), followed by searches in CINAHL; Psycharticles (EBSCO); Web of Knowledge Social Science and Emerging sources databases; Cochrane; SocIndex (EBSCO); Family and Society Studies Worldwide and Academic Search Complete (EBSCO) and covered the period 2000–2018.
In addition to searching online databases, additional hand searching and citation tracking was done. Grey literature searches included Global Health database, and websites of relevant conferences, including Conferences on Retroviruses and Opportunistic Infections; International AIDS Conference; International AIDS Society and AIDS Impact. In addition to reaching out to organisational contacts, researchers in adolescent mental health were contacted for any ongoing or unpublished work. The identified studies were shared with various experts locally and internationally and authors were contacted personally where studies referred to additional work (Supplementary Appendix 1, available online at https://doi.org/10.1192/bjo.2020.67).
Inclusion and exclusion criteria
Two reviewers (Mv.P., A.B.) independently screened titles, abstracts and full texts. Discrepancies were resolved through discussion and consensus. Papers were restricted to those reporting on quantitative evaluation. Evaluations were included if they reported on interventions with mental health outcomes among adolescents and young people aged 10–24 years living with HIV or affected by HIV in LMICs (defined by the World Bank classification). Affected by HIV includes any study of adolescents in which a caregiver is identified as living with HIV or had died because of an AIDS-related illness.
Full inclusion criteria were as follows.
(a) Study population: adolescents aged 10–24 years living with HIV, affected by HIV (as defined earlier).
(b) Study design: any quantitative study design IF they report on mental health outcomes, including randomised controlled trials, quasi-experimental designs, pre-post evaluations and post evaluations.
(c) Outcome measures: any measure of mental health among adolescents living with or affected by HIV, specifically measures on internalising problems (anxiety, depression), externalising problems (conduct problems, attention-deficit disorders and hyperactivity), including self-harm and substance abuse and coping or adjustment (emotional and psychological well-being, self-esteem, self-concept and resilience).
Exclusion criteria were as follows.
(a) Study population: adolescents aged 10–24 years whose HIV status is unknown or unspecified or the adolescent is an orphan, but the cause of the caregiver's death is unknown or not specified.
(b) Study design: qualitative studies; any quantitative study that does not assess any intervention effects.
(c) Outcome measures: measures of mental health not reported on at baseline and/or at follow-up.
(d) Non-English language.
Data extraction and quality appraisal
Included studies were subject to detailed data abstraction and analysis. We used a predetermined data extraction tool with information about the key aspects of the study design and findings as required by PRISMA guidelines (Fig. 1). Studies were critically appraised using the Standard Quality Assessment Criteria for Evaluating Primary Research Papers from a Variety of Fields. The assessment tool included 14 criteria covering objectivity, transparency and appropriateness of research methods used (Supplementary Appendix 2).25
Fig. 1 Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram.
Results
Of 4811 related titles, 16 mental health intervention studies were included in the review. The PRISMA flow diagram details the study exclusion process and reasons for exclusion (Fig. 1). Overall, 4133 individuals were represented in studies from China, Ethiopia, Kenya, Myanmar, Rwanda, South Africa, Uganda and Thailand.
Study quality characteristics
Thirteen of the studies focused on HIV-affected adolescents, and three studies focused on adolescents living with HIV.Reference Bhana, Mellins, Small, Nestadt, Leu and Petersen9,Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26,Reference Nestadt, Chutima, McKay, Torsak, Pardo and Sudrak27 Fifteen studies met 10 or more of the 14 quality criteria, most often related to inadequate participant selection details or insufficient details in the results. One study met 9 of the 14 criteria, primarily because it lacked a control group thereby increasing the chance of confounding.
Study design of included evaluations
Twelve studies used a randomised control trial (RCT) design.Reference Bhana, Mellins, Small, Nestadt, Leu and Petersen9,Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26–Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36 Five of these were cluster RCTs, clustered by village,Reference Li, Liang, Ji, Wu and Xiao34 township,Reference Mon, Liabsuetrakul and Htut32 schoolReference Kumakech, Cantor-Graae, Maling and Bajunirwe33,Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36 and church.Reference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35 One study used a quasi-experimental post-intervention design.Reference Mueller, Alie, Jonas, Brown and Sherr37 Two studies used a pre–post design, without control groups.Reference Jani, Vu, Kay, Habtamu and Kalibala38–Reference Betancourt, Ng, Kirk, Munyanah, Mushashi and Ingabire40
Nine studies employed a single follow-up period of either 3 or 6 months;Reference Bhana, Mellins, Small, Nestadt, Leu and Petersen9,Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26,Reference Bhana, Petersen, Mason, Mahintsho, Bell and McKay28,Reference Bell, Bhana, Petersen, McKay, Gibbons and Bannon29,Reference Mon, Liabsuetrakul and Htut32,Reference Kumakech, Cantor-Graae, Maling and Bajunirwe33,Reference Mueller, Alie, Jonas, Brown and Sherr37–Reference Thurman, Nice, Luckett and Visser39 six studies variously used two follow-up periods that ranged from 2, 3, 6, 9, 12 and 24 months,Reference Nestadt, Chutima, McKay, Torsak, Pardo and Sudrak27,Reference Betancourt, Ng, Kirk, Brennan, Beardslee and Stulac30,Reference Li, Liang, Ji, Wu and Xiao34–Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36,Reference Betancourt, Ng, Kirk, Munyanah, Mushashi and Ingabire40 and one study used three follow-up periods of 6, 12 and 18 months.Reference Eloff, Finestone, Makin, Boeving-Allen, Visser and Ebersohn31 The median follow-up period was 3 months. Only two studies specifically focused on orphansReference Kumakech, Cantor-Graae, Maling and Bajunirwe33,Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36 and one on migrant adolescents.Reference Jani, Vu, Kay, Habtamu and Kalibala38 In the remaining studies, the orphan status of the adolescent did not feature. Given that most studies used a family-strengthening approach, it is reasonable to assume that the adolescents lived with them as a family (not necessarily as a nuclear family).
Participants were recruited through schools,Reference Bhana, Petersen, Mason, Mahintsho, Bell and McKay28,Reference Bell, Bhana, Petersen, McKay, Gibbons and Bannon29,Reference Kumakech, Cantor-Graae, Maling and Bajunirwe33,Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36,Reference Mueller, Alie, Jonas, Brown and Sherr37 healthcare settings (for example clinics or hospitals),Reference Bhana, Mellins, Small, Nestadt, Leu and Petersen9,Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26,Reference Nestadt, Chutima, McKay, Torsak, Pardo and Sudrak27,Reference Betancourt, Ng, Kirk, Brennan, Beardslee and Stulac30,Reference Eloff, Finestone, Makin, Boeving-Allen, Visser and Ebersohn31,Reference Li, Liang, Ji, Wu and Xiao34,Reference Betancourt, Ng, Kirk, Munyanah, Mushashi and Ingabire40 churches,Reference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35 a township community sampleReference Mon, Liabsuetrakul and Htut32 and community-based or service delivery organisations.Reference Jani, Vu, Kay, Habtamu and Kalibala38,Reference Thurman, Nice, Luckett and Visser39
Measurement characteristics
The majority of reviewed studies (15/16) measured adolescent mental health and/or behaviour using scales such as the Child Problem Behaviour Checklist,Reference Bell, Bhana, Petersen, McKay, Gibbons and Bannon29,Reference Eloff, Finestone, Makin, Boeving-Allen, Visser and Ebersohn31 Child Depression InventoryReference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26,Reference Eloff, Finestone, Makin, Boeving-Allen, Visser and Ebersohn31,Reference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35–Reference Mueller, Alie, Jonas, Brown and Sherr37 and the Strengths and Difficulties Questionnaire.Reference Bhana, Mellins, Small, Nestadt, Leu and Petersen9,Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26,Reference Mon, Liabsuetrakul and Htut32,Reference Mueller, Alie, Jonas, Brown and Sherr37 Several studies also measured caregiver mental health or stress,Reference Bhana, Mellins, Small, Nestadt, Leu and Petersen9,Reference Eloff, Finestone, Makin, Boeving-Allen, Visser and Ebersohn31,Reference Thurman, Nice, Luckett and Visser39 as well as general health or well-being of adolescents,Reference Bell, Bhana, Petersen, McKay, Gibbons and Bannon29,Reference Betancourt, Ng, Kirk, Brennan, Beardslee and Stulac30,Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36 and experiencing abuse or violence.Reference Mueller, Alie, Jonas, Brown and Sherr37,Reference Jani, Vu, Kay, Habtamu and Kalibala38
Eight studies included a measure that focused on self-efficacy, self-concept or self-esteem of adolescents. For example, three studies used the Rosenberg Self-esteem Scale,Reference Li, Liang, Ji, Wu and Xiao34,Reference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35,Reference Mueller, Alie, Jonas, Brown and Sherr37 while another three used the Tennessee Self-concept Scale.Reference Bhana, Mellins, Small, Nestadt, Leu and Petersen9,Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26,Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36 Three studies also measured the coping or adaptive behaviour of adolescentsReference Bhana, Mellins, Small, Nestadt, Leu and Petersen9,Reference Eloff, Finestone, Makin, Boeving-Allen, Visser and Ebersohn31 and caregivers.Reference Eloff, Finestone, Makin, Boeving-Allen, Visser and Ebersohn31
Several studies (11/16) incorporated measures that focused on caregiver and adolescent communication and connectedness. Scales used to assess this included Communication Comfort and FrequencyReference Bhana, Mellins, Small, Nestadt, Leu and Petersen9,Reference Bhana, Petersen, Mason, Mahintsho, Bell and McKay28,Reference Bell, Bhana, Petersen, McKay, Gibbons and Bannon29 and Family Connectedness.Reference Betancourt, Ng, Kirk, Brennan, Beardslee and Stulac30,Reference Betancourt, Ng, Kirk, Munyanah, Mushashi and Ingabire40 In addition, 11 studies also measured parenting practices using adapted measures of parenting behaviourReference Bhana, Mellins, Small, Nestadt, Leu and Petersen9,Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26–Reference Eloff, Finestone, Makin, Boeving-Allen, Visser and Ebersohn31,Reference Li, Liang, Ji, Wu and Xiao34,Reference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35,Reference Thurman, Nice, Luckett and Visser39,Reference Betancourt, Ng, Kirk, Munyanah, Mushashi and Ingabire40 or scales such as the Alabama Parenting QuestionnaireReference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35 and Parental Bonding Instrument.Reference Li, Liang, Ji, Wu and Xiao34
Seven studies measured knowledge related to HIV/AIDS using scales such as the HIV Knowledge Questionnaire,Reference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35,Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36,Reference Jani, Vu, Kay, Habtamu and Kalibala38 AIDS Myth KnowledgeReference Bhana, Petersen, Mason, Mahintsho, Bell and McKay28,Reference Bell, Bhana, Petersen, McKay, Gibbons and Bannon29 and the South African Demographic Health Survey.Reference Thurman, Nice, Luckett and Visser39 Five studies also measured stigma related to HIV/AIDS.Reference Bhana, Mellins, Small, Nestadt, Leu and Petersen9,Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26,Reference Bhana, Petersen, Mason, Mahintsho, Bell and McKay28,Reference Bell, Bhana, Petersen, McKay, Gibbons and Bannon29,Reference Mueller, Alie, Jonas, Brown and Sherr37 Another four measured sexual behaviour using scales such as the Self-Efficacy to Refuse Sexual Behavior ScaleReference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35 and the Condom Use Self-efficacy Scale.Reference Thurman, Nice, Luckett and Visser39 One study measured adolescents living with HIV adherence to antiretroviral medication.Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26 Four studies looked at social support or capital using measures such as the Social Network Analysis ScaleReference Bell, Bhana, Petersen, McKay, Gibbons and Bannon29 and the Network of Relationships Inventory.Reference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35 One study measured educational outcomes (for example, school drop-out).Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36
Intervention characteristics
Eight studies used trained lay workers to deliver their programmes.Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26,Reference Bhana, Petersen, Mason, Mahintsho, Bell and McKay28–Reference Eloff, Finestone, Makin, Boeving-Allen, Visser and Ebersohn31,Reference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35,Reference Thurman, Nice, Luckett and Visser39,Reference Betancourt, Ng, Kirk, Munyanah, Mushashi and Ingabire40 Other studies used various levels of professional staff such as teachers, health educators, counsellors, social workers and mindfulness practitioners. One study used non-governmental organisation, community development, agricultural and financial officersReference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36 and another used trained youth ambassadors.Reference Mueller, Alie, Jonas, Brown and Sherr37
Intervention sites varied with five studies delivering the intervention at clinics or hospitals,Reference Bhana, Mellins, Small, Nestadt, Leu and Petersen9,Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26,Reference Nestadt, Chutima, McKay, Torsak, Pardo and Sudrak27,Reference Eloff, Finestone, Makin, Boeving-Allen, Visser and Ebersohn31,Reference Jani, Vu, Kay, Habtamu and Kalibala38 four were school based,Reference Bhana, Petersen, Mason, Mahintsho, Bell and McKay28,Reference Bell, Bhana, Petersen, McKay, Gibbons and Bannon29,Reference Kumakech, Cantor-Graae, Maling and Bajunirwe33,Reference Mueller, Alie, Jonas, Brown and Sherr37 two studies were community based (with one delivering the intervention after weekly church services),Reference Mon, Liabsuetrakul and Htut32,Reference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35 two studies both community and home basedReference Li, Liang, Ji, Wu and Xiao34,Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36 and three were only home based.Reference Betancourt, Ng, Kirk, Brennan, Beardslee and Stulac30,Reference Thurman, Nice, Luckett and Visser39,Reference Betancourt, Ng, Kirk, Munyanah, Mushashi and Ingabire40 Interventions were broadly grouped into family-level and group-based interventions. The section below describes these two types of interventions.
Family-level interventions for adolescents and caregivers
Twelve of the 16 reviewed studies focused on family-strengthening parenting interventions, all of which included both separate and joint sessions for caregivers and adolescents (Table 1).Reference Bhana, Mellins, Small, Nestadt, Leu and Petersen9,Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26–Reference Eloff, Finestone, Makin, Boeving-Allen, Visser and Ebersohn31,Reference Li, Liang, Ji, Wu and Xiao34–Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36,Reference Thurman, Nice, Luckett and Visser39,Reference Betancourt, Ng, Kirk, Munyanah, Mushashi and Ingabire40
Table 1 Characteristics of family-based interventions studiesa
RCT, randomised controlled trial.
a. Gender of participants: male and females.
b. Own calculation.
They all described culturally adapted parenting interventions. For example, the content and delivery of five studies (CHAMP, VUKA and VUKA Thailand) used the same methodology to meet the needs of the local contextReference Bhana, Mellins, Small, Nestadt, Leu and Petersen9,Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26–Reference Bell, Bhana, Petersen, McKay, Gibbons and Bannon29 as did the two Family-Strengthening Interventions (FSI-HIV)Reference Betancourt, Ng, Kirk, Brennan, Beardslee and Stulac30,Reference Betancourt, Ng, Kirk, Munyanah, Mushashi and Ingabire40 and an intervention delivered at local churches.Reference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35 The most common elements of the family-strengthening parenting interventions broadly encompassed narratives of sources of family resilience and protective influences for the child. To strengthen understanding, studies typically included psychoeducation related to HIV transmission, treatment knowledge, understanding puberty and sexual relationships, emotional experiences related to being HIV-positive, managing internalised and external stigma and adherence. To enhance protective influences, family-strengthening studies focused on developing more responsive (less authoritarian) parenting skills, emphasising supervision and monitoring and enhancing the frequency and quality of communication skills with children and adolescents. In addition, some family-strengthening studies included coping, problem-solving (including cognitive–behavioural therapy in some cases) and stress management skills for caregivers. All these studies used trained lay health workers or counsellors to deliver these interventions.Reference Bhana, Mellins, Small, Nestadt, Leu and Petersen9,Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26–Reference Bhana, Petersen, Mason, Mahintsho, Bell and McKay28,Reference Eloff, Finestone, Makin, Boeving-Allen, Visser and Ebersohn31,Reference Li, Liang, Ji, Wu and Xiao34,Reference Thurman, Nice, Luckett and Visser39 One study used multiple types of churches to enrol adolescents and their caregivers to participate in nine sessions of 2 h each to improve family relationships, communication skills, skills-based HIV prevention interventions, parent training and cognitive–behavioural therapies.Reference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35 One family-strengthening study focused specifically on economic empowerment seeking to enhance financial literacy skills, family microenterprise skills development and financial management skillsReference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36 and used both lay and professional staff. In this intervention, because non-governmental organisations use mentorship as part of standard care for orphaned children, an average of one mentorship meeting per month was introduced for all participants. Those in the intervention arm were offered an opportunity to open a bank account and any savings up to $10 a month per family would be matched by the project in addition to receiving ten 1- to 2-hour microenterprise development workshops (see Appendix).
Group-based interventions for adolescents
Three studies delivered an intervention that was group-based with some taking place in schoolsReference Kumakech, Cantor-Graae, Maling and Bajunirwe33,Reference Mueller, Alie, Jonas, Brown and Sherr37 or some form of group activity (Table 2).Reference Mon, Liabsuetrakul and Htut32 For example, school-based peer groups were used to deliver psychoeducation about fears and worries and skills development among orphans who lost one or both parents to manage their fears and increase self-esteem using individual and group problem-solving approaches to reinforce coping skills and improve self-esteem.Reference Kumakech, Cantor-Graae, Maling and Bajunirwe33 In another example, mindfulness training and practice in the form of individual and group sessions in meditation to alter emotional and cognitive processes formed the basis of developing skills around emotional awareness, self-regulation, developing non-judgmental attitudes, listening skills and empathy and was combined with family strengthening in the form of parent training skills to enhance positive parenting skills among caregivers in a community-based study.Reference Mon, Liabsuetrakul and Htut32
Table 2 Characteristics of group-based interventions studiesa
RCT, randomised controlled trial.
a. Gender of participants: male and females.
b. Own calculation.
Individual interventions for adolescents
Only one study took an individual-level approach to strengthen the mental health of adolescents affected by HIV using client-centred counselling, problem-solving therapy, group and creative therapies (music, art) as the basis of interventions among adolescent migrant workers in Ethiopia as they represented highly vulnerable mobile adolescents facing sexual abuse and exploitation (Table 3).Reference Jani, Vu, Kay, Habtamu and Kalibala38
Table 3 Characteristics of individual-based interventions studiesa
a. Gender of participants: male and females.
Intervention effectiveness
Effectiveness of family-focused interventions
Although some family-strengthening studies focused only on adolescent outcomes, others described both caregiver and adolescent outcomes separately and yet others the relative influence and association of one on the other emphasising parenting. Excluding the three pilot studies,Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26,Reference Bhana, Petersen, Mason, Mahintsho, Bell and McKay28,Reference Betancourt, Ng, Kirk, Brennan, Beardslee and Stulac30 8 of the 12 primary studies that measured mental health and behaviour outcomes among adolescents reported that their interventions had a significant impact on adolescent mental health.Reference Bhana, Mellins, Small, Nestadt, Leu and Petersen9,Reference Nestadt, Chutima, McKay, Torsak, Pardo and Sudrak27,Reference Betancourt, Ng, Kirk, Brennan, Beardslee and Stulac30,Reference Li, Liang, Ji, Wu and Xiao34,Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36,Reference Thurman, Nice, Luckett and Visser39,Reference Betancourt, Ng, Kirk, Munyanah, Mushashi and Ingabire40 In one study, caregiver reports indicated improvement in child externalising behaviours.Reference Eloff, Finestone, Makin, Boeving-Allen, Visser and Ebersohn31 Four studies reported no significant impact on adolescent mental health outcomes.Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26,Reference Bhana, Petersen, Mason, Mahintsho, Bell and McKay28,Reference Bell, Bhana, Petersen, McKay, Gibbons and Bannon29,Reference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35
In five primary studies, the interventions were associated with reduced mental health problems, including depression, anxiety, irritability and feelings of hopelessness.Reference Betancourt, Ng, Kirk, Brennan, Beardslee and Stulac30,Reference Kumakech, Cantor-Graae, Maling and Bajunirwe33,Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36,Reference Jani, Vu, Kay, Habtamu and Kalibala38,Reference Thurman, Nice, Luckett and Visser39 In one study, this was the case for female adolescents, but not for males.Reference Jani, Vu, Kay, Habtamu and Kalibala38 The authors explained that this was related to a higher prevalence of mental health problems among males at baseline and that males experienced greater social dislocation with most having lived or currently living on the street. While two studies reported no significant impact on conduct or behaviour problems,Reference Betancourt, Ng, Kirk, Brennan, Beardslee and Stulac30,Reference Mueller, Alie, Jonas, Brown and Sherr37 five reported an improvement in adolescent behaviour; specifically, externalising behaviourReference Eloff, Finestone, Makin, Boeving-Allen, Visser and Ebersohn31,Reference Mon, Liabsuetrakul and Htut32,Reference Li, Liang, Ji, Wu and Xiao34,Reference Jani, Vu, Kay, Habtamu and Kalibala38 and prosocial behaviour.Reference Betancourt, Ng, Kirk, Munyanah, Mushashi and Ingabire40 Using simple linear regression analysis, one family-strengthening intervention of adolescents living with HIV found that less youth depression was associated with higher caregiver education, greater caregiver supervision, more social support seeking, higher youth self-esteem, lower internalised stigma, and child use of resignation for coping.Reference Bhana, Mellins, Small, Nestadt, Leu and Petersen9 Out of the three studies that measured caregiver mental health, one reported a significant decrease in mental health issues among caregivers as a result of the intervention.Reference Thurman, Nice, Luckett and Visser39
One study found that fewer child difficulties were associated with lower caregiver depression.Reference Bhana, Mellins, Small, Nestadt, Leu and Petersen9 Another two studies reported an improvement in self-esteem among adolescents as a result of the intervention,Reference Li, Liang, Ji, Wu and Xiao34,Reference Betancourt, Ng, Kirk, Munyanah, Mushashi and Ingabire40 and one reported no significant difference.Reference Mueller, Alie, Jonas, Brown and Sherr37 However, in this same study self-efficacy improved. Lower self-efficacy was also associated with orphan status, violence at home, social connection and AIDS-related stigma. In one study both self-efficacy and self-concept improved,Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36 but in another there was no significant impact on self-concept,Reference Kumakech, Cantor-Graae, Maling and Bajunirwe33 thus indicating equivocal results when it comes to the effect of interventions on adolescents’ self-worth.
Caregiver and child communication and connectedness improved in seven of the eight studies that measured this. For instance, compared with the control groups, caregivers and adolescents in the intervention groups felt more comfortable communicating, and communicated more frequently.Reference Bhana, Mellins, Small, Nestadt, Leu and Petersen9,Reference Bell, Bhana, Petersen, McKay, Gibbons and Bannon29,Reference Thurman, Nice, Luckett and Visser39 Better family communication was also found among church-goers.Reference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35 In addition, prosocial behaviours were associated with greater caregiver-reported communication.Reference Bhana, Mellins, Small, Nestadt, Leu and Petersen9 However, one study found no significant difference in family connectedness.Reference Betancourt, Ng, Kirk, Brennan, Beardslee and Stulac30 With regards to parenting practices more generally, the results are mixed: in two studies parenting improved; specifically caregiver monitoringReference Bell, Bhana, Petersen, McKay, Gibbons and Bannon29 and disciplinary methods.Reference Betancourt, Ng, Kirk, Munyanah, Mushashi and Ingabire40 However, in three studies there was no significant difference when it came to parenting,Reference Betancourt, Ng, Kirk, Brennan, Beardslee and Stulac30,Reference Eloff, Finestone, Makin, Boeving-Allen, Visser and Ebersohn31,Reference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35 and in another, there was a significant difference for children but not for adolescents.Reference Li, Liang, Ji, Wu and Xiao34
Three out of the six studies that measured social network or support reported a significant impact on this outcome compared with control groups.Reference Bhana, Petersen, Mason, Mahintsho, Bell and McKay28,Reference Bell, Bhana, Petersen, McKay, Gibbons and Bannon29,Reference Betancourt, Ng, Kirk, Munyanah, Mushashi and Ingabire40 Family-level economic strengthening significantly enhanced the likelihood of taking primary leaving examinations to advance to post-primary schooling as well as influencing primary leaving examinations scores. In addition, the family-level economic strengthening arm was associated with lower levels of hopelessness among younger adolescents and a higher average self-concept compared with controls.Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36
Effectiveness of school and group-based interventions
A school-based group peer-support intervention revealed significant improvements in symptoms of depression, anxiety and anger among 10- to 15-year-old adolescents with HIV who were orphans.Reference Kumakech, Cantor-Graae, Maling and Bajunirwe33 Client-centred counselling combined with group and creative arts counselling showed improved knowledge of HIV transmission, uptake of HIV and sexual health services for both male and female adolescents, but a reduction in mental health problems/behaviours (social and attention problems; anxious/depressed and aggressive behaviours) occurred only among female adolescents.Reference Jani, Vu, Kay, Habtamu and Kalibala38 A monthly group mindfulness intervention among adolescents in Myanmar showed significant improvements in emotional and conduct behaviour problems but not social behaviours.Reference Mon, Liabsuetrakul and Htut32 A group-based intervention among HIV-affected adolescents did not show any significant differences in depression, emotional and behavioural problems or self-esteem, although this study did not have a baseline measure.Reference Mueller, Alie, Jonas, Brown and Sherr37
Associated non-mental health outcomes
There was a significant improvement in HIV knowledge in six of the seven studies that measured this outcome. For one study this occurred among caregivers;Reference Bhana, Petersen, Mason, Mahintsho, Bell and McKay28 in another among both adolescents and caregivers,Reference Bell, Bhana, Petersen, McKay, Gibbons and Bannon29 and four among adolescents.Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26,Reference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35,Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36,Reference Jani, Vu, Kay, Habtamu and Kalibala38 In several studies, there was an improvement in safe sex practices. For example, an increase in adolescent self-efficacy regarding condom use,Reference Thurman, Nice, Luckett and Visser39 and sexual risk-taking reduced in two studies.Reference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35,Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36 Adherence to antiretroviral medication improved in two out of the three studies that measured this outcome,Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26 as did internalisedReference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26 and externalised stigma.Reference Bell, Bhana, Petersen, McKay, Gibbons and Bannon29 In the client-centred study,Reference Jani, Vu, Kay, Habtamu and Kalibala38 the intervention was associated with an increased uptake of HIV and sexual health services, including HIV testing.
Discussion
This is the first comprehensive review of interventions aiming to support the mental health of adolescents living with and affected by HIV in LMICs. We identified 16 studies, of which 12 were family focused, 3 were group based and 1 focused on the individual adolescent.
Key findings
Family-strengthening interventions hold promise
Family-strengthening interventions were most often used with both adolescents living with HIV and those affected by HIV (12 studies). Eight of these studies had a significant effect on mental health outcomes. An examination of the family-strengthening interventions that had a significant effect reveals that they were primarily focused on enhancing caregiver and adolescent resilience, improving communication and parenting skills (monitoring and supervision) and increasing social support and social networks to reduce social isolation.Reference Bhana, Mellins, Small, Nestadt, Leu and Petersen9,Reference Nestadt, Chutima, McKay, Torsak, Pardo and Sudrak27,Reference Betancourt, Ng, Kirk, Brennan, Beardslee and Stulac30,Reference Eloff, Finestone, Makin, Boeving-Allen, Visser and Ebersohn31,Reference Li, Liang, Ji, Wu and Xiao34,Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36,Reference Thurman, Nice, Luckett and Visser39 These elements were also most consistently reported as having significantly improved following an intervention. Nevertheless, similar elements constituted a focus of the four family-based interventions that did not work. Because studies varied in terms of sample size, study design, intervention controls and measurement, it is difficult to know what may have accounted for the non-significant findings. In part, this may be because two of these studies were pilot studies focused on establishing feasibility.Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26,Reference Bhana, Petersen, Mason, Mahintsho, Bell and McKay28 In the other two studies, mental health outcomes were deemed to be long-term outcomes secondary to family-strengthening efforts.Reference Bell, Bhana, Petersen, McKay, Gibbons and Bannon29,Reference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35
Many of the identified studies focused on strengthening families through efforts to improving parent–child relationships (parenting styles), communication and increasing knowledge related to HIV and its transmission. These studies also clearly identified the importance of caregiver well-being as well as good psychosocial support in mitigating the effects of living with or being affected by HIV. An important strength of family-focused studies is the emphasis on families and adolescents living in low-resource settings. Parental monitoring and supervision were also highlighted as important positive influences (although there is inconsistency in positive findings in the studies). This was often combined with attempts to increase knowledge around HIV transmission among both caregivers and adolescents. In addition, interventions variously included developing skills related to self-esteem, self-concept, ways of coping, problem-solving, stress management, as well as elements of cognitive–behavioural therapy. Together, it was expected that these elements would help reduce internalising stigma as well as reducing prejudice towards others affected by HIV. The overarching intervention approach was to strengthen family interactions emphasising improved parent–child communication around topics related to potential adolescent risk behaviours such as sexual debut, drug and alcohol use, negative peer influences, stigma and understanding HIV itself. The central role of the family in interventions may also be related to the targeted adolescent age ranges with eight studies recruiting samples 16 years and younger, in one from 6 to 18 years, in another from 12 to 17 years and in two from 7 to 17 years. Mental health programming that is integrated with family-strengthening efforts is likely to enhance mental health outcomes.
It is difficult to identify the elements of family-strengthening programmes that mediate improved adolescent mental health. Change may have come about through the focus on enhancing communication, promoting caring and supportive relationships and on training parents and other caregivers in skills that would be more sensitive to the emotional needs of adolescents. Most trained the adolescent and their families in problem-solving, which is an evidence-based way to improve depression. Several programmes also provided sessions to enhance knowledge and understanding of sexual and reproductive health, which, within the context of fears about HIV transmission, may positively influence mental health outcomes. The ecological model of risk and resilience extends beyond individual factors to emphasising the importance of social and cultural factors in reducing sexual risk and highlighting the protective role of family and community environments for long-term benefits, including positive effects of family-based and parenting interventions, improved communication, parent–child relationships, decreased conflict, reducing internalising and externalising symptoms, including risk behaviour. Individual well-being is dependent on both individual behaviour changes as well as contextual and cultural factors in this model.
What might work, but only worked in a few studies?
The four studies that did not use a family-strengthening approach used very different modalities of intervention within and between them (one of which had no significant effects on mental healthReference Mueller, Alie, Jonas, Brown and Sherr37). Client-centred psychosocial counselling was used to deliver individual, group and creative therapies such as music art and drama as part of the intervention package. Aside from a basic single counselling session, everyone was assessed on an ongoing basis to determine further individual or group counselling. Deeply rooted emotional issues would be referred for group counselling that encompassed art, music or drama therapy.Reference Jani, Vu, Kay, Habtamu and Kalibala38 The principle of ‘conscientisation’, i.e., sharing experiential knowledge, feelings, dialogue, participation and development of critical awareness and empowerment to facilitate the transformation of oppressive experiences into liberating ones formed the basis of a school-based peer-group support interventions among adolescents orphaned by AIDS. These took the form of 16 psychosocial exercises implemented by selected teachers over 10 weeks lasting 1 h each and resulted in significant improvements in depression, anger, and anxiety but not self-concept.Reference Kumakech, Cantor-Graae, Maling and Bajunirwe33 A mindfulness intervention involving monthly group sessions for 3 consecutive months conducted by an experienced trainer led to significant emotional and conduct behaviour improvements at 6 months but not social behaviours. Each adolescent was exposed to eight sessions lasting about 2 h. The sessions content included general principles of mindfulness meditation followed by actual practice. Each session began with mindfulness meditation increasing from 10 to 15 min with logged mindfulness practice at home three to four times per week. Guardians were only involved in providing feedback about the adolescent's practice.Reference Mon, Liabsuetrakul and Htut32
Knowledge gaps and recommendations for research
Although mental health was an important outcome, attempts to directly influence mental health outcomes were found only in a minority of the studies reviewed. Further, the association between HIV and specific mental disorders is relatively understudied among adolescents living with HIV and those affected by HIV. This may be because mental health tends to be conceptualised broadly for adolescents living with HIV, as part of psychosocial well-being, alongside other socioeconomic, educational, sexual health and family-related vulnerabilities. For instance, many interventions focus on sexual risk given this was perceived as very important by study participants. Measurement bias constitutes another important gap as few studies used culturally adapted measures of mental health outcomes.
The strong focus on family-based interventions meant that assessments concentrated on caregiver and adolescent communication, and connectedness and parenting practices as these are seen to mitigate risk influences.
Aside from the relatively more frequent use of measures such the Strengths and Difficulties Questionnaire and the Child Depression Inventory, measurement characteristics of the studies varied greatly and included measures of self-esteem (Tennessee Self-concept Scale; Rosenburg Self-esteem Scale), knowledge related to HIV/AIDS, stigma and sexual behaviour. Adherence to medication was a prominent feature of the few studies in which the sample of adolescents were HIV-positive.
More importantly, well-known mental health interventions for adolescents such as cognitive–behavioural therapy or problem-solving therapy did not feature prominently, and consideration should be given to how these approaches could be incorporated with the family-strengthening elements of most interventions.
It is difficult to separate the individual effects of the more individual-strengthening elements from the family-strengthening elements as these are often reported as part of the overall intervention (even though their individual effects may be reported) and no doubt also part of the intervention. To some extent, these individual elements were added to address specific issues related to the study sample to influence individual risk as well as the risk posed by negative environments. It is also not possible to determine whether family-strengthening studies worked better with adolescents living with HIV compared with those affected by HIV, in part because only three studiesReference Bhana, Mellins, Small, Nestadt, Leu and Petersen9,Reference Bhana, Mellins, Petersen, Alicea, Myeza and Holst26,Reference Nestadt, Chutima, McKay, Torsak, Pardo and Sudrak27 focused on adolescents living with HIV (two of which used the same intervention with one being a pilot). There is also a significant dearth of intervention studies targeted to orphans living or affected by HIV. Only two studies included orphans as participants with significant positive effects.Reference Kumakech, Cantor-Graae, Maling and Bajunirwe33,Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36
While every study included both female and male adolescents in their study sample, there did not appear to be any specific attempts to tailor any of the interventions to any gender. It is likely that adolescent girls and boys not only experience qualitative differences in relation to their families but may also experience specific forms of vulnerability related to their gender. Examining gender-specific vulnerabilities and its relation to mental outcomes among adolescents living with HIV and those affected by HIV should form an important part of the next generation of studies in this area.
Despite an emphasis on interventions meeting the needs of adolescents living with HIV or those affected by HIV more holistically by linking these to social protection, caregiver health, levels of household employment and sources of income,Reference Cluver, Pantelic, Orkin, Toska, Medley and Sherr41 for the most part, these elements did not form a primary focus of the interventions except in one study focused on economic empowerment.Reference Ssewamala, Karimli, Torsten, Wang, Han and Ilic36 With the advent of antiretrovirals being made more widely available none of the studies reported on the effects of this treatment, on self-concept or self-esteem. Greater emphasis needs to be placed on understanding the bidirectional relationship between antiretroviral treatment and mental health.
What was encouraging is that just over a third of the studies employed multiple extended follow-up assessments ranging from 1 to 3 months, while others included follow-up periods of 6, 12, 18 and 24 months. While only a few studies used a follow-up period of one year or longer,Reference Bhana, Mellins, Small, Nestadt, Leu and Petersen9, Reference Betancourt, Ng, Kirk, Brennan, Beardslee and Stulac30, Reference Puffer, Green, Sikkema, Broverman, Ogwang-Odhiambo and Pian35 the positive intervention effects on mental health outcomes in these studies is encouraging.
Public mental health services for adolescents and their families living with or affected by HIV is an important policy issue as these remain unaddressed in current health services. Access to antiretroviral therapy alone is insufficient to ensure a good quality of life. Using implementation science approaches to scale up innovative community-based interventions that can be delivered by non-specialists, given the shortage of mental health professionals in LMICs, should be a consideration among policymakers. Given overburdened healthcare services and that most family-strengthening interventions used lay health workers, integrating interventions with current testing services should be explored.
Limitations
Despite most of the interventions being described as family-strengthening interventions, each of these interventions used a variety of methods to deliver their interventions and used a heterogeneous group of individuals varying from trained lay counsellors to trained healthcare professionals, and it is therefore not possible to determine the scalability of these interventions. Even though most studies employed randomised study designs, the variations in sample size, study sites from school to healthcare settings, churches and community-based service delivery organisations, as well as the varying measures make direct comparisons between studies difficult. Information related to the quality and fidelity of implementation, as well as costs of the implementation are lacking. The absence of any studies specifically focused on the treatment of mental disorders is a huge research gap.
Implications
This systematic review provides a comprehensive overview of the evidence on mental health outcomes related to adolescents living with HIV and affected by HIV. The number of evidence-based interventions that address mental health for adolescents in low- and middle-income contexts who are living with HIV is negligible and in need of urgent attention given that an increasing number of adolescents are on antiretroviral treatment. Findings among both adolescents living with or affected by HIV indicates that family-strengthening interventions are favoured over school or group-based or individual interventions. Most family-strengthening initiatives focus on enhancing caregiver parenting, communication and social connectedness than targeting any specific mental health outcome. This is primarily a function of the emphasis on building resilience in vulnerable populations from low-resource contexts as a protective mechanism for adolescents. Although the heterogeneity in study design and methods of implementation and evaluation make it difficult to directly compare the interventions, what this group of studies shows is that it is possible to influence mental health outcomes of HIV-positive adolescents or adolescents affected by HIV (see Appendix). There appears to be low uptake of these interventions, even though attempts to influence policy and practice are evident in some of the interventions. Interventions targeted at mental health concerns may need to extend the focus of family-strengthening to include a broader range of life issues associated with adolescents living with HIV and not only focus on health. This may mean that such interventions must be delivered through community-based agencies sensitive to local contexts.
Supplementary material
Supplementary material is available online at http://doi.org/10.1192/bjo.2020.67
Data availability
The authors confirm that the data supporting the findings of this study are available within the article and in the Supplementary material.
Acknowledgements
Alison Bullen provided professional library services undertaking the pilot and final database searches.
Cedric Nininahazwe from the YPLUSNetwork.org for his comments on the initial drafts of the manuscript.
Author contributions
A.B. had overall responsibility for the study design and management. A.B. and M.v.P. screened and finalised selection of the studies. JK assisted with the detailed table for each study. M.P., M.A.A. and L.A.M. independently reviewed the included studies. A.B., M.A.A., M.P., J.K., M.v.P. and L.A.M. reviewed and edited drafts and the final version of the manuscript.
Funding
The study was supported by the International HIV/AIDS Alliance.
Declaration of interest
None.
ICMJE forms are in the supplementary material, available online at https://doi.org/10.1192/bjo.2020.67.
Appendix
Intervention types: common characteristics
Open access
eLetters
No eLetters have been published for this article.