Charania MR, Crepaz N, Guenther-Gray C, Henny K, Liau A, Willis LA, Lyles CM. Efficacy of structural-level condom distribution interventionns: A meta-analysis of U.S. and international studies, 1998-2007. AIDS Behav. 2010 Oct 1. [Epub ahead of print]
To examine the efficacy of structural-level condom distribution interventions (SLIs) on HIV risk behavior and STIs; to identify factors associated with intervention efficacy.
The CDC HIV/AIDS Prevention Research Synthesis (PRS) project cumulative HIV/AIDS/STI data base was reviewed.(1, 2) Although how this database is constructed was described, the authors did not specify how they searched the database. An additional automated search was performed in MEDLINE, EMBASE and PsycINFO using the keywords 'condom' and synonyms for 'distribution.' An additional search was performed of the Cochrane Library. English language reports and studies published from January 1988 through September 2007 were included if they involved: a) an HIV/AIDS/STI behavioral intervention focused on condom distribution as a 'structural component' targeting acceptability, availability, or accessibility of condoms, b) collection of data on at least one behavioral outcome, c) a trial or multiple cross-sectional sample design, and d) sufficient data for calculation of effect sizes. Data were extracted and coded by pairs of reviewers using standardized coding forms; methodological quality was assessed based on study design, type of comparison, retention rates and whether intention-to-treat analysis was used.
Out of 2097 behavioral prevention citations, 21 studies meeting eligibility criteria were selected, 14 of which were outside the US (5 in Africa, 4 in Asia, 2 in the Caribbean, and 3 in Central America). 10 studies were trials (4 RCTs and 6 non-RCT), among which the median retention rate was 86% (range 40-99%); 8 trials used intent-to-treat analyses. The remaining 11 studies compared data from independent cross-sectional samples; for these studies, the median followup time for behavioral outcomes was 14 months post-baseline, and the median follow-up for biological outcomes was 17.5months.
There were approximately 24,000 participants in the 21 studies; the median age was 22 (range 15-65). 8 studies targeted commercial sex workers or brothel managers; 8 targeted youth/young adults; 2 targeted clinic patients and the remainder targeted adults and those at risk of STIs.
SLI interventions were defined as those that addressed contextual factors impacting condom use, specifically: availability (e.g. condom bowls in clinics or 100% condom use brothel policies); accessibility (e.g. condom mass distribution); acceptability (e.g. changing social norms through public service announcements about condom benefits). SLIs could be directed at: the individual level (e.g. condom machines, flyers to teenagers); the organizational level (e.g. condoms in prison, television programming); or the environmental level (e.g. social marketing campaigns, policy changes, increased federal funding for condom availability). No interventions addressed acceptability as a sole strategy, and none were implemented only at the environmental level. The average length of intervention implementation was 10 months (US based studies) and 15 months (international studies). 10/14 international-based studies included 2 or more SLI components in the intervention; 14/21 included additional individual (ILI), group (GLI) or community level interventions (CLI) that were not structural in nature (e.g. teaching individuals how to use a condom, promoting a positive attitude about condoms, giving participants a condom after the intervention).
The effect size (ES) between the most theoretically potent intervention arm and the comparison arm (for trials), or between the longest post-intervention followup and baseline (for multiple cross-sectional studies) was calculated, using percentage of persons who 'did not engage in unprotected sex' as a measure of condom use. Other outcome measures included condom acquisition, sexual initiation, incidence of STIs and number of sex partners. . Effect sizes were estimated using odds ratios (OR) and an overall effect size across all studies was estimated. The magnitude of heterogeneity of the ES was evaluated using the Q statistic and Higgin's I2 index. Sensitivity analysis was conducted. Stratified analyses were conducted to examine intervention factors associated with outcome, including: the type of structural component (e.g. availability, acceptability, or accessibility), the level of implementation of this component (individual, organizational, or environmental), and whether additional (non-structural) individual, group, or community-level components were included. The Comprehensive Meta-Analysis software was used to perform analyses based on random- and mixed-effects models.
Significant intervention effects were found for the following outcomes: increased condom use (20 studies) (OR=1.8, 95% CI 1.5-2.2); increased condom acquisition/carrying (6 studies) (OR=5.4, CI 1.9-15.7); delayed sexual initiation or abstinence (5 studies) (OR=1.4, CI1.0=2.0); and reduced incidence of STIs (5 studies) (OR=-0.69, CI 0.53-0.91). There was no significant intervention effect on number of sex partners (7 studies) (OR 1.3, CI-0.9-1.9). There was considerable heterogeneity among studies, although no individual effect size exerted influence on overall heterogeneity. Both US-based and international studies showed efficacy, although those conducted in international settings were significantly more efficacious than those conducted in the US. Increasing availability or accessibility of condoms as a sole strategy, or those that combined both these approaches had greater efficacy. Interventions at all levels of implementation (individual, organizational or environmental) all had an impact on condom use behaviors, although those that were implemented at an individual level, whether alone or combined with organizational or environmental approaches, showed greater impact. Interventions that were only structural, and did not include non-structural ILI, GLI, or CLI also showed efficacy, but when combined with these additional components had significantly greater efficacy.
Structural level condom distribution interventions, evaluated as a whole, can increase condom use, reduce some HIV sexual risk behaviors and incident STIs. Those that combine both SLI with individual, community or small group approaches have the greatest impact.
This was a high quality meta-analysis in which search strategies and analytic approach were well described.
There is evidence that ILI, SLIs and CLIs demonstrate moderate to high success in promoting condom use.(3) A meta-analysis of HIV reduction indicated that educational, psychosocial or behavioral interventions can reduce high risk behavior among adolescents.(4) Another meta-analysis showed that safer sex programs providing condoms do not increase sexual frequency.(5) These studies were limited to small scale delivery (individual or group) and did not address the social, structural systems or the environment, factors that have been shown to be a critical influence on individual behaviors.(6, 7)
This systematic review supports the implementation of structural level condom distribution interventions to reduce HIV risk behavior. Efforts should continue to make condoms more universally available, accessible and acceptable, particularly in communities or venues targeting high risk individuals. However, SLIs alone work less well than those that also include some individualized components that address knowledge, attitudes, skills and behaviors For example, making condoms available at venues frequented by youth along with providing prevention messages through individual and small group sessions are likely to be the most useful. SLIs may also be able to prolong sexual initiation among youth.
- Lyles CM, Crepaz N, Herbst JH, Kay L, for the HIV/AIDS Research Synthesis (PRS) Project. Evidence-based HIV behavioral prevention from the perspective of CDC's HIV/AIDS Prevention Research Synthesis Team. AIDS Educ Prev. 2006;18(4Suppl A): S21-31.
- DeLuca JB, Mullins MM, Lyles C, Crepaz N, Kay LS, Thadiparti S. Developing a comprehensive search strategy for evidence based systematic review. Evid Based Libr Inf Prac. 2008;3(1):3-32.
- Noar S. Behavioral interventions to reduce HIV-related sexual risk behavior: review and synthesis of meta-analytic evidence. AIDS Behav. 2008;12(3):335-53.
- Johnson BT, Carey MP, Marsh KL, Levin KD, Scott-Sheldon LA. Interventions to reduce sexual risk for the human immunodeficiency virus in adolescents, 1985-2000: a research synthesis. Arch Pediatr Adolesc Med. 2003;157(4):381-8.
- Smoak ND, Scott-Sheldon LAJ, Johnson BT, Carey MP. Sexual risk reduction interventions do not inadvertently increase the overall frequency of sexual behavior: a meta-analysis of 174.
- Wohlfeiler D, Ellen JM. The limits of behavioral interventions for STD/HIV prevention. In: Cohen L, Chavez V, Chemini S, editors. Prevention is primary: strategies for community wellbeing. San Francisco: Jossey-Bass; 2008. p. 329-47.
- Blankenship KM, Bray SJ, Merson MH. Structural interventions in public health. AIDS. 2000;14(Suppl 1):S11-21. studies with 116, 735 participants. J Acquir Immune Defic Syndr. 2006;41(3):374-84.