|Efficacy of structural-level condom distribution interventions: a meta-analysis of U.S. and international studies, 1998-2007
|Charania MR, Crepaz N, Guenther-Gray C, Henny K, Liau A, Willis LA, Lyles CM
This review found that structural-level condom distribution interventions increased condom use and reduced HIV/STD risk. Although the review was generally well conducted, overall results may not be entirely reliable due to analyses being based on diverse populations, interventions and study types.
To evaluate the efficacy of structural-level condom distribution interventions
The authors searched the Centre for Disease Control HIV/AIDS/STI prevention database which covered a range of databases and handsearched journals from 1988. They also conducted a separate search of MEDLINE, EMBASE and PsycINFO in October 2007 with keywords documented in the report. The authors also searched The Cochrane Library and checked the references of all relevant studies.
Studies needed to report on a HIV/AIDS/STI behavioural intervention that focused on condom distribution as a structural component that targeted acceptability, availability or accessibility of condoms. Data needed to be collected on at least one behavioural outcome or biological outcome. Eligible studies could be randomised controlled trials or non-randomised controlled trials where control groups did not receive a structural-level condom distribution intervention, or before and after studies. Studies needed to be published between January 1988 and September 2007 in English and have sufficient data to calculate effect sizes.
Included studies took place across the world; over one third targeted commercial sex workers and either their clients or brothel managers. Over a third of studies targeted young people. The median age across the studies was 22 (range 15 to 65 years), where reported. Among the studies that reported participant gender, most either targeted exclusively or included a majority of women. Interventions included: condom provision in schools, local businesses, health clinics and other venues; information provision and media campaigns. Studies lasted from one month to 42 months or were ongoing. Length of follow-up ranged from two months to 51 months from baseline. Reported outcomes included condom use, number of sex partners, acquisition/condom carrying, sexual initiation or abstinence and STI.
The authors did not state how many reviewers were involved in study selection.
Assessment of study quality
Quality was assessed based on study design, type of comparison, retention rates and intention-to-treat.
Trials were independently coded by pairs of reviewers. The reviewers reconciled discrepancies and if no agreement was reached, a third reviewer helped to resolve the discrepancy.
For studies with multiple arms, the contrast between the most theoretically potent intervention arm and the comparison, which was typically a standard care or waiting list control arm, was extracted. For the condom use outcome, data from studies that reported unprotected sex were calculated to reflect the percentage that "did not engage in unprotected sex" to combine it with data from studies that reported condom use behaviour. For studies that reported multiple follow-up assessments, the longest post-intervention follow-up was selected. Data from adjusted models reported by study authors were used but if these were not provided, effect sizes were calculated for follow-up outcome data by adjusting for baseline differences. Effect sizes for each study were estimated using the odds ratio (OR). For studies that reported means and standard deviation values, standardised mean differences were calculated and converted into odds ratio values.
Trials were independently coded by pairs of reviewers using standardised forms. The reviewers reconciled discrepancies and if no agreement was reached a third reviewer helped to resolve the discrepancy.
Methods of synthesis
Overall odds ratios were calculated with their associated 95% confidence interval using a random-effects model. Study heterogeneity was examined using the Q statistic and Ι². Stratified analyses were conducted to assess the impact of study quality, study location, type of structural component, level of implementation of the structural-level component and whether individual, group or community interventions were included as part of the structural-level intervention. Sensitivity analyses were also conducted, excluding a study or set of studies at a time. Publication bias was investigated through funnel plots.
Results of the review
Twenty-one studies were included in the review (21,301 participants). Ten studies had a comparison arm not receiving a structural-level condom distribution intervention. The remaining eleven compared data from independent cross-sectional samples assessed before and after implementation of a structural-level condom distribution intervention.
Statistically significant effects of the interventions were found for increased condom use (OR 1.81, 95% CI 1.51 to 2.17; p<0.01; 20 studies), increased condom acquisition/condom carrying (OR 5.40, 95% CI 1.86 to 15.66; p<0.05; six studies), delaying sexual initiation/abstinence among youth (OR 1.43, 95% CI 1.01 to 2.03; p<0.05; five studies) and reduced incidence of STIs (OR 0.69, 95% CI 0.53 to 0.91; p<0.01; five studies). There was no significant effect for the number of sex partners (OR 1.28, 95% CI 0.89 to 1.85; p>0.18; seven studies). There was considerable unexplained heterogeneity among studies. There was no evidence of publication bias for any of the outcomes. Results of stratified analyses of intervention effects on condom use were also presented.
This review supported the structural-level condom distribution intervention as an efficacious approach to increasing condom use and reducing HIV/STD risk.
This review was based on defined inclusion criteria and was underpinned by a range of searching methods. Unpublished studies were not eligible for the review but the authors found no evidence of publication bias. Studies in languages other than English were not considered which could have lead to bias. The authors conducted a post-hoc search and believed that only two studies would have been missed which would not have changed their conclusions. Quality was broadly assessed and the impact of study quality on results assessed. More than one reviewer was involved in the processes of study selection, data extraction and quality assessment for the review which helped to minimise bias. It was unclear if pooling studies in meta-analyses was entirely appropriate given there was substantial unexplained heterogeneity in the overall results and stratified analyses. Readers should bear this in mind when considering results of the meta-analysis.
Implications of the review for practice and research
Practice: The review's findings supported the 2007 UNAIDS Guidelines Towards Universal Care which recommended universal and uninterrupted condom availability and integrated condom promotion into other health services as part of a comprehensive HIV prevention approach.
Research: Future studies should explore questions of implementation and costs of these programmes. They should also investigate the benefits of these interventions for high risk populations that were disproportionately affected by HIV and other STDs.
Charania MR, Crepaz N, Guenther-Gray C, Henny K, Liau A, Willis LA, Lyles CM. Efficacy of structural-level condom distribution interventions: a meta-analysis of U.S. and international studies, 1998-2007. AIDS and Behavior 2011; 15(7): 1283-1297
Subject indexing assigned by NLM
Adolescent; Adult; Africa; Asia; Condoms /statistics & Delivery of Health Care /organization & HIV Infections /ethnology /prevention & Health Knowledge, Attitudes, Practice; Humans; Male; Risk-Taking; Sexual Behavior; Sexually Transmitted Diseases /ethnology /prevention & United States; Vulnerable Populations; administration; control; control; numerical data /utilization
Date bibliographic record published
Date abstract record published
This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.