Twelve RCTs (n=4,639) were included in the review. Follow-up duration ranged between one and 24 months. Retention rates (where reported) ranged from 37% to 95% (mean 70%).
Computer-based technology statistically significantly increased condom use (d=0.259, 95% CI 0.201 to 0.317; 12 RCTs). There was evidence of statistical heterogeneity (p=0.007). Sensitivity analysis did not significantly alter the results. Fail-safe N suggested that a large number of unpublished studies additional to those included in the meta-analysis would need to exist to nullify the reported effect size.
Computer-based technology had a statistically significant effect on reducing the frequency of sexual behaviour (d=0.427, 95% CI 0.251 to 0.602; three RCTs), incidence of sexually transmitted disease (d=0.140, 95% CI 0.035 to 0.245; three RCTs) and number of sexual partners (d=0.422, 95% CI 0.116 to 0.728; two RCTs).
Further analyses showed that there were statistically significant associations between effect size and gender (female-only interventions reported the largest effect size, p<0.01), individually tailored interventions (both p<0.001), stages of change model (p<0.001) and intervention dose (p<0.05).