Evidence from high-income countries on the effectiveness of psychosocial interventions to improve mental health, wellbeing and quality of life for adults living with HIV: a systematic review and meta-analysis

Abstract

Introduction: There is a need to synthesize recent evidence on the effectiveness of psychosocial interventions to improve mental health, quality of life and wellbeing in adults living with HIV in high-income countries. A systematic review and meta-analysis was conducted to address this research gap.

Methods: Medline, Embase, Psychinfo and Web of science were searched (from 2008 to December 2023). In total, 67 randomized controlled trials (RCTs) of psychosocial intervention among adults living with HIV in high-income countries were eligible.

Results: In the meta-analysis, there was an overall positive effect of interventions on reducing depression (N = 40; standardized mean difference [SMD] -0.19 [95% CI: -0.29, -0.10]), anxiety (N = 15; SMD -0.12 [-0.23, -0.02]), stress (N = 13; SMD -0.22 [-0.41, -0.04]), and other measures of poor wellbeing (N = 19; SMD -0.18 [-0.35, -0.02]) and increasing levels of coping/self-efficacy (N = 8; SMD 0.17 [0.04, 0.31]). For depression, interventions that used symptom screening above a threshold score to identify eligible individuals were more effective than those without such an eligibility criterion (SMD -0.29 vs. -0.10, p = 0.023). Interventions compared to standard care controls had a greater effect on depression versus interventions compared to not standard care controls, when the latter category included standard care controls that received intentional support (SMD -0.28 vs. -0.11, p = 0.035). There was also weak evidence of an overall positive effect on: reducing stigma (N = 7; SMD -0.17 [-0.35, 0.02]), and improving social support/participation (N = 6; SMD 0.17 [-0.02, 0.35]), mental health quality of life (N = 12; SMD 0.09 [-0.01, 0.19]), physical health quality of life (N = 11; SMD 0.07 [-0.02, 0.16]) and quality of social life (N = 6; SMD 0.10 [-0.04, 0.24]). There was no evidence found for an effect on loneliness, although data were limited.

Discussion: Pooled effect estimates were small or small tomoderate. In line with previous literature, there was no evidence of differential effects on depression according to the intervention type (psychotherapeutic vs. other).

Conclusions: Evidence from RCTs suggest that psychosocial interventions are effective in improving mental health for adults living with HIV in high-income settings. Interventions were more effective at reducing depression when targeted at those screening positive for mental health symptoms and when compared to a standard care only control group. There was some evidence that longer, more intensive interventions were more effective.

Keywords: HIV; RCT; anxiety; depression; intervention; psychosocial.

Figure 1

Flow chart of study inclusion and exclusion. ^aA systematic search was undertaken of the electronic databases Embase, Medline, PsycInfo and Web of Science to identify RCTs of psychosocial interventions for people living with HIV in high‐income countries published from 2008 up to 14th December 2023. Search terms included those related to: (a) HIV, (b) psychosocial interventions (definition based on a previous review [11], e.g. psychotherapy, counselling, coaching, social support, Cognitive behavioural therapy (CBT), meditation), (c) RCTs (based on a Cochrane recommended search strategy [45]) and (d) high‐income countries (based on World Bank Data [46, 47]). Terms related to mental health/wellbeing/quality of life outcomes were not specified in the search strategy. This was to identify studies that did not mention these outcomes in the title, abstract or key words but did report on this data in the paper. Search terms specifically related to studies of people without HIV (e.g. HIV self‐testing, PrEP, PEP) and pharmaceutical interventions (e.g. pharmacokinetic, immunogenicity) were used to remove ineligible studies. ^bStage 1: The title and abstract of all identified studies was reviewed for eligibility. Two reviewers (ARM and JS) independently assessed a sample of 100 studies (first and last 50 studies when sorted by author name). Thereafter, one reviewer (ARM) completed the abstract review. ^cStage 2: Two reviewers (ARM and/or JS) reviewed the full‐text of all studies selected in stage 1. Disagreement about study selection was resolved via discussion with one or more co‐authors. Of note, studies were included regardless of whether mental health, quality of life or wellbeing measures collected were the primary endpoint of the trial. ^dAuthors of recently published papers (2019−2023) were contacted. Forty‐six authors were contacted, of whom 27 responded, and five provided data. ^eTitle and abstract of all articles referenced in 67 eligible studies were imported into Endnote (N = 2499). All references with “HIV” in the title that were published after 2007 were reviewed (N = 804). Of these, 37 were deemed to be potentially eligible and the full‐text articles were reviewed. No articles were found to be eligible. Articles were excluded for the following reasons: not HIV‐positive sample (n = 2), not high‐income country (n = 1), not RCT (n = 8), not from 2007 onwards (n = 10), not >18 years of age (n = 2), no randomized comparison (n = 3), no mental health/wellbeing data collected (n = 5), and six articles were based on an eligible study already captured in the systematic review. Previous reviews and meta‐analyses on this subject were also searched to identify eligible studies—no eligible studies that were not already captured in the review were identified. ^fA protocol to extract the data from eligible studies was developed [14, 48]. One reviewer (ARM, SMR or JS) initially extracted the data from each study. A second blinded reviewer checked for data extraction errors (FCL, VC, FN). Authors were contacted to retrieve information that was not published.

Figure 2

Effect of psychosocial interventions on symptoms of depression among people living with HIV (N = 40 interventions).

Figure 3

Investigating publication bias in 40 studies that contributed to the main depression analysis (Egger's test p‐value = 0.640).

Figure 4

Effect of psychosocial interventions on symptoms of anxiety among people living with HIV (N = 15 interventions).

Figure 5

Effect of psychosocial interventions on stress (N = 13), stigma (N = 7), loneliness (N = 4) and other wellbeing measures (N = 19) among people living with HIV.

Figure 6

Effect of psychosocial interventions on measures of coping/self‐efficacy (N = 8) and social support (N = 6).

Figure 7

Effect of psychosocial interventions on measures of quality of life; general^a (N = 7), physical (N = 11), mental (N = 12) and social (N = 6).

Figure 8

Overall effects of psychosocial interventions on each outcome measure.

Figure 9

Risk of bias criteria assessed in 62^a studies included in analyses.