Behavioral Counseling Interventions to Prevent Sexually Transmitted Infections: A Systematic Evidence Review for the U.S. Preventive Services Task Force [Internet]

Review
Rockville (MD): Agency for Healthcare Research and Quality (US); 2020 Aug. Report No.: 19-05260-EF-1.

Excerpt

Objective: To support an update to the USPSTF’s previous recommendation on Behavioral Counseling Interventions to Prevent Sexually Transmitted Infections, we systematically reviewed evidence on behavioral counseling interventions aimed at primary risk reduction for adolescents and adults.

Data Sources: Studies from the previous review were evaluated for inclusion, and we performed a comprehensive search of MEDLINE, PubMed (publisher-supplied only), PsycINFO, and the Cochrane Collaboration Registry of Controlled Trials for studies published between January 2013 and June 2018. A research librarian developed and executed the search strategy.

Study Selection: Two researchers independently reviewed 4,649 abstracts and 273 full-text articles against prespecified inclusion criteria, then abstracted data from included studies. English-language controlled clinical trials published after 1999 were included to assess the direct health benefits, intermediate behavioral changes, and potential harms of interventions screening for STI prevention. Trials with control conditions of usual care, no treatment, attention control, or waitlist were included. Included study populations were pregnant and nonpregnant adolescents or adults of any sexual orientation, sexually active or not. Studies conducted in existing social networks (e.g., school classrooms, clubs, churches) were excluded, as were studies of interventions for HIV-positive individuals only. Included interventions included behavioral counseling to reduce STI risk, delivered alone or in combination with other sexual risk-reduction interventions, that were feasible to implement in or refer from primary care. Studies reporting incidence of STI or protective and risky sexual behaviors such as condom use or unprotected intercourse were eligible for inclusion if followup data at least 3 months from baseline were reported.

Data Analysis: Eligible studies assessed to have high risk of bias (e.g., > 40% loss to followup) were excluded for poor quality according to standard USPSTF procedures. Descriptions of the study populations and intervention characteristics were provided for all included studies. The most consistently reported health and behavioral outcomes were analyzed with random effects meta-analysis to calculate pooled intervention effects when data were sufficient. We conducted meta-analyses of STI incidence, condom use, and unprotected intercourse, using the Dersimonian and Laird method or, when fewer than 10 trials were available for pooling, restricted maximum likelihood (REML) models with the Knapp-Hartung correction for small samples. We also examined differences in intervention effects by population, intervention, and study design characteristics, using planned and exploratory metaregression to identify sources of statistical heterogeneity for the main outcome of STI incidence. Finally, funnel plots and Egger tests for small-study effects were generated for all pooled analyses that included 10 or more studies. Using established methods, we assessed the strength of evidence for each key question.

Results: We included 39 studies (54 intervention arms) reported in 63 articles (n = 65,888), including 17 new studies not in the previous review. All but five studies were conducted in the United States, and one-third were good-quality trials. Over half of the studies enrolled women or girls only; nearly two-thirds were among adolescents or young adults; and over two-thirds reported that more than half of participants identified as a racial or ethnic minority. The behavioral counseling interventions tested in the trials employed a range of behavior change strategies, with individual or group counseling a component of over two-thirds of the intervention arms. Eight trials tested interventions with low contact times (<30 minutes), and high contact time interventions (>120 minutes) were included in over half of the included trials. Behavior change techniques based on cognitive behavioral therapy or motivational interviewing were included in over one-third of the interventions, and the content of the interventions was often derived from concepts outlined in behavior change theories. Twenty-one studies reported STI outcomes. Meta-analysis spanning all studies providing adequate data showed that the interventions were associated with one-third lower odds of STI (19 RCTs; n=52,072; odds ratio [OR], 0.66 [95% CI, 0.54 to 0.81]; I2 74%) but statistical and clinical heterogeneity was high among studies. Behavioral outcomes were reported in 34 trials (n = 21,417), but inconsistency in measures used and outcomes reported limited meta-analysis. Eighteen studies reported condom use outcomes (n = 9,205). There was evidence that interventions were associated with increased condom use in meta-analysis of dichotomous reports such as consistent condom use (13 RCTs; n=5,253; odds ratio [OR], 1.31 [95% CI, 1.10 to 1.56]; I2 40%) and greater percentage times condoms were used with sexual intercourse (7 trials; n randomized = 2,920; mean difference [MD], 10.75 [95% CI, 1.01 to 20.50]; I2 79%). For studies reporting counts of unprotected intercourse (i.e., without a condom), measures were inconsistently defined with regard to the types of intercourse (vaginal, anal, any), time period assessed, and partnerships (e.g., any partner, non-primary partner). Behavioral interventions tested in studies that reported this outcome also were associated with a reduction in sexual risk behaviors. Intervention was associated with fewer unprotected sexual acts (14 RCTs; n = 9,183; mean difference [MD], −0.94 [95% CI, −1.40 to −0.48]; I2 16%). There was not evidence of small study effects. Meta-regression to assess potential sources of statistical heterogeneity was limited by correlated study features and the number of studies reporting STI outcomes. There was limited evidence suggesting potentially stronger effects for interventions that included group counseling or higher contact times (> 2 hours).

Limitations: Our search was limited to English-language literature. We excluded trials with active comparators (i.e., comparative effectiveness), conducted in existing social networks (e.g., clubs, schools, workplaces), interventions focused on reducing transmission among HIV-positive individuals, and people very high STI risk behaviors requiring specialized preventive care (e.g., IV drug using, engaged in sex work). Statistical and clinical heterogeneity were high for the included evidence. The direct evidence for STI prevention was consistent and precise, but the diversity of study populations, interventions, and effect sizes contributed to a wide range of effects. Factors explaining the range of effects were difficult to disentangle because they tended to be highly correlated. For behavioral outcomes, the outcomes reported and specific measures used were highly variable, and limited the estimation of pooled effects across studies such that only a subset of fewer than half of the included studies could be combined. Furthermore, self-reported behavioral outcomes, particularly those involving sexual practices, are likely to be affected by social desirability and recall bias. Men, older adults, gay, bisexual, and transgender people were not well-represented in the included study populations.

Conclusions: For individuals at increased risk for STI, there are effective behavioral counseling interventions for reducing STI incidence, by increasing condom use, and reducing sexual risk behaviors. There was no evidence of intervention harms, but few studies reported on potential adverse effects. Research is needed to address gaps in the literature on primary care feasible behavioral interventions that could apply to more heterogeneous populations seen in primary care and to address the needs of specific populations at heightened risk for STI not represented in this body of evidence.

Publication types

  • Review

Grants and funding

Prepared for: Agency for Healthcare Research and Quality, U.S. Department of Health and Human Services, 5600 Fishers Lane, Rockville, MD 20857; www.ahrq.govContract No. HHSA-290-2015-000017-I-5Prepared by: Kaiser Permanente Evidence-based Practice Center, Kaiser Permanente Center for Health Research, Portland, OR