The review concluded that behavioural interventions targeted at sexually transmitted disease (STD) clinic patients reduced sexual risk behaviour and prevented HIV/STDs. Limitations in the analyses and uncertain quality of included studies mean the authors' conclusions should be interpreted with caution.

To determine the efficacy of behavioural interventions to reduce sexual risk behaviour and incidence of sexually transmitted diseases (STDs) among individuals attending STD clinics in the United States.

PubMed, PsycINFO, Dissertation Abstracts, ERIC, CINAHL and The Cochrane Library were searched up to February 2009; search terms were reported. NIH RePORTER was searched and references of relevant articles and professional journals were checked. Only articles published in English were included.

Randomised controlled trials (RCTs) or quasi-experimental studies of individual or group level behavioural interventions aimed at reducing sexual risk behaviour and risk of STDs in individuals attending STD clinics in USA that reported longer term outcomes (up to two years post intervention) were eligible for inclusion. Outcomes of interest included sexual risk behaviour (condom use, number of sexual partners) and STD acquisition (including HIV). Studies were required to report pre-test outcomes and sufficient data to enable calculation of effect sizes. Studies that included samples with greater than 50% HIV-positive patients or evaluated structural level interventions (such as mass media) were excluded. Where a study reported more than one comparator, the condition with the least contact (such as wait list) was used.

Intervention content included STD testing and counselling (98%), education (85%), skills training (85%) and motivational components. Forty per cent of interventions reported provision of condoms. Most studies included active comparators (STD/HIV education or brief form of the intervention); others used an assessment-only control. Most interventions occurred during the clinic visit (simultaneously or immediately after) and were conducted in one session (range one to seven) that lasted a median 44 minutes (range 12.5 to 210 minutes). Facilitators delivered the intervention to individuals and/or small groups. Timing of post-intervention assessments varied. First assessments occurred up to 64 weeks. Second and third assessments occurred at 13 and 26 weeks. Participants were largely restricted to those who had a STD (self-reported or diagnosed via clinical examination and/or laboratory test) and those who were at increased risk of contracting a STD. Forty-six per cent of the included participants were women and 72% were ethnically black. Mean age was 25.8 years (range 17 to 35).

The authors did not state how many reviewers performed the study selection.

Two reviewers independently assessed the methodological quality of the included studies against 12 items taken from two validated measures (Jadad and a measure described by Miller et al.). Possible scores ranged from zero to 17. Any disagreements were resolved through discussion.

Two reviewers independently extracted data to calculate effect sizes (d) for continuous outcomes and odds ratios (ORs) for dichotomous outcomes which were transformed into effect size using the Cox transformation. Where means and standard deviations were not available, other information (such as t or F tests) was used. Extreme effect sizes (more than two standard deviations from the mean) were recorded as equivalent to two standard deviations. Where no statistical information was available and the authors reported no significant between-group differences the effect size was estimated as zero. Authors of included trials were contacted for missing data. Any disagreements were resolved through discussion. Twenty studies were randomly selected to assess inter-rater reliability.

Pooled weighted mean effects and their associated 95% confidence intervals (CIs) were calculated using fixed-effect and random-effects models. Statistical heterogeneity was assessed using the Q-statistic and I². Effect sizes were clustered into three time intervals (short-term, intermediate and long-term); separate analyses were performed for each interval. The impact of various moderator variables was assessed using both univariate and multivariate regression analyses.

Where multiple effect sizes were calculated from individual studies (for each intervention or by sample characteristic) these were analysed as a separate study. Where a study reported multiple measures of the same outcome the effect sizes were averaged (corresponding sample sizes were also averaged).

Publication bias was assessed by visual inspection of funnel plots and quantified using Begg's adjusted rank and Egger's test. Rosenthal's fail safe N was applied when publication bias was found.

Thirty-two studies (48 separate interventions) were included in the review (n=67,538 participants): condom use (20 studies, 34 interventions), number of sexual partners (15 studies, 28 interventions), incidence of STD (22 studies, 40 interventions), incidence of HIV (five studies, six interventions). Methodological quality scores ranged from 7 to 15 (mean 10.5).

Small to medium improvements with risk reduction interventions were shown in condom use, reduced number of sexual partners and incidence of STDs (HIV included).

Short-term assessment (four to 13 weeks) showed significant benefit in favour of behavioural interventions was found for condom use (0.10, 95% CI 0.07 to 0.14, I²=59%; 31 interventions) and number of sexual partners (0.09, 95% CI 0.06 to 0.13, I²=0%; 24 interventions) compared to control groups. No between-group differences were found for incidence of STDs.

Intermediate assessment (22 to 39 weeks) showed significantly improved condom use (0.05, 95% CI 0.02 to 0.09, I²=0%; 26 interventions), reduced number of sexual partners (0.06, 95% CI 0.03 to 0.09, I²=0%; 22 interventions) and incidence of STDs (0.13, 95% CI 0.08 to 0.18, I²=75%; 21 interventions) reported with the behavioural intervention.

Long term assessment (≥52 weeks) showed a significantly reduced number of sexual partners (0.06, 95 CI 0.01 to 0.11, I²=69%; 11 interventions) and incidence of STDs (0.10, 95 CI 0.08 to 0.12, I²= 70%; 23 interventions), which included HIV (0.56, 95% CI 0.54 to 0.58, I²=99%; five interventions) was found with the behavioural intervention compared with controls. No significant between-group differences were found for condom use.

Eggers test indicated significant publication bias for number of sexual partners at short-term assessment (p=0.4). The fail-safe N indicated that 84 interventions assessing number of sexual partners within three months post intervention with non-significant results would be necessary to reverse the significant findings. Results of moderator analyses were reported.

Behavioural interventions targeted at STD clinic patients reduced sexual risk behaviour and prevented HIV/STDs. Widespread use of such interventions in STD clinics should be a public health priority.

The review question was supported by clearly defined inclusion criteria. Several relevant electronic databases were searched. The literature search was restricted by language (English), which risked language bias. Publication bias was found for only one outcome (number of sexual partners at short-term assessment). Appropriate steps were taken to minimise the likelihood of error or bias during data extraction and study quality assessment; it was unclear whether similar steps were taken for study selection. Study validity was assessed using appropriate criteria, but full results were not reported and this limited interpretation of the results. Studies were pooled using meta-analyses; significant statistical heterogeneity was found in most of the analyses performed. It appeared as if individuals may have been counted more than once in the analyses and that some effect sizes were averaged before pooling, which may not have been appropriate. The authors highlighted that many of the outcomes were self-reported and so vulnerable to a number of biases.

Limitations in the analyses and uncertain quality of included studies mean that the authors' conclusions should be interpreted with caution.

Practice: The authors state that sexual risk reduction behavioural interventions for younger patients should be routinely implemented in clinic settings. Widespread use of behavioural interventions in STD clinics to reduce sexual risk behaviour and incidence of STDs should be a public health priority.

Research: The authors suggest that future research should include patients most at risk, target content specifically for STD patients and include specific subgroups (such as young adults) as well as deliver intervention in longer doses. Investigation of the interactive effects of STD diagnosis and intervention efficacy was recommended. Exploration of motivational factors associated with STD diagnosis among men was recommended.

One reviewer was supported by National Institute of Health grant R01-MH068171.

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.