Safetxt trial: a secondary analysis exploring the intervention effects

Problem

The safetxt trial was an individually randomised parallel group trial conducted among 6248 people aged 16-24 in the UK. The effect of the safetxt safer sex intervention on the cumulative incidence of chlamydia/ gonorrhoea at 12 months was odds ratio 1.13, 95% confidence interval 0.98 to 1.31 and on condom use at last sex was odds ratio1.14, 95% CI 1.01 to 1.28. Limited observational data suggests that the average duration of rectal chlamydia or gonorrhoea in 2-3 months. We conducted an secondary data analysis to explore if the slightly higher infections identified in the intervention group might be attributable to more cases being diagnosed and treated in clinics during the trial in the intervention group, with undiagnosed infections in the control group resolving before the 12 months postal STI test.

Approach

A secondary data analysis. We compared the prevalence of chlamydia or gonorrhoea reinfection at one year in each group using logistic regression. We adjusted the primary analysis regression for the prespecified baseline covariates (age, type of STI at baseline, sex or gender, sexual orientation, and ethnicity). We report the adjusted odds ratios with 95% confidence intervals. We also created kaplan-meier plots for time to STI test and time to diagnosis, and calculated the proportions of people undergoing a STI test during the trial in clinics.

Findings

The intervention effects on the prevalence of chlamydia and goorhoea at 12 months was odds ratio 1.04 (95% CI 0.83, 1.30)Although there were slightly more STI tests in the intervention group, and a slightly higher proportion of tests were positive these differences were not statistically significant. Time to test and diagnosis were similar in intervention and control groups.

Consequences

There was no difference in the prevalence of chlamydia/ gonorrhoea at 12 months between the intervention and control group and no statistically differences in STI testing. Our findings do not support a hypothesis that the slightly higher incidence of chlamydia/ gonorrhoea in the interventions group are attributable to more cases being identified and treated in the intervention compared to the control groups, with undiagnosed infections in the control group resolving prior to the 12 month postal chlamydia testing.