A Meta-analysis (MA) of the Effect of World-Wide Psychosocial HIV Interventions on the Sex Behaviors of Persons Who Use Drugs

Article type
Authors
Semaan S1, Meader N2, Des-Jarlais D3, Halton M2, Bhatti H2, Chan M2
1Centers for Disease Control and Prevention, National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention, 1600 Clifton Rd., E-07, Atlanta, GA, 30333
2National Collaborating Center for Mental Health, Royal College of Psychiatrists Research Unit, 21 Mansell Street, London, E1 8AA, UK
3Beth Israel Medical Center, Baron Edmond de Rothschild Chemical Dependency Institute, 160 Water St., Rm 2462, New York, New York, 10038
Abstract
Background: Users of opiates and stimulants are at sexual risk for HIV that can be reduced by psychosocial interventions. Objectives: We examined efficacy of 44 worldwide sexual risk reduction intervention studies with drug users (outcomes: unprotected sex, condom use, number of sex partners). We compared the results to our previous MA results (published in 2002) of 33 interventions conducted in the United States (US). The 44 interventions included the 33 interventions (published 1988 1999). Methods: We conducted comprehensive searches for sexual risk reduction intervention studies (experimental or quasi-experimental designs, 1988 August 2009). Effect sizes were converted to odds ratios (OR); an OR & 1.0 indicated less unsafe sexual behavior in experimental groups at follow-up. Results: Psychosocial interventions had modest additional benefit compared to shorter, control educational interventions (k = 44; OR, 0.87; 95% confidence interval [CI], 0.77 0.97), and larger positive effect compared to minimal interventions (k = 7; OR, 0.62; 95% CI, 0.47 0.83). Cumulative MA showed significant effect (OR = 0.88; 95% CI, 0.78 0.99) by 1996, similar to previous MA results. US (k = 38; OR, 0.87; 95% CI = 0.77 0.99) and non-US studies (k = 6; OR = 0.82; 95% CI = 0.63 1.06) had small effect, similar to previous MA results (k = 33; OR, 0.86; 95% CI = 0.76 0.98). Individual (k = 43; OR, 0.83; 95% CI, 0.73 0.94), network (k = 4, OR, 0.75; 95% CI, 0.47 1.18), and structural (k = 2; OR, 0.75; 95% CI, 0.60 0.94) interventions had similar effects; network and structural interventions seem promising. Meta-regression showed that HIV status (self-report or study test results; 0 20% vs. > 20%) was associated (B = 0.27, p = 0.04) with sexual risk reduction. Conclusion: Current and previous MA had similar results. Psychosocial interventions provide modest sexual risk reduction justifying offering them to drug users. Moderate-duration interventions may have diminishing returns. Stronger-effect interventions are needed.