HIV sexual risk-reduction interventions for youth: a review and methodological critique of randomized controlled trials.

Pedlow et al. (2003) looked at every youth HIV risk-reduction trial they could find. They pulled 23 studies that randomly assigned teens to get lessons or not.

The team asked two simple questions. Did the program cut risky sex? And did the study hold up to basic science rules?

Just over half of the trials helped kids play it safer. The rest showed no clear win.

Big problem: most papers never said how large the benefit was. Few tracked kids after the classes ended, so we do not know if the gains stuck.

Tromans et al. (2018) saw the same mess in autism trials. They scanned 529 studies and found most had tiny samples, mirroring the youth HIV field.

Jonsson et al. (2016) added that even when ASD social-skills trials claimed success, they left out key details like where and how the lessons ran. This makes it hard to copy the program in real clinics, just like the HIV lessons.

Kok et al. (2026) later showed that single-case youth behavior studies also fade after treatment stops. The pattern is clear across topics: short-term wins, long-term question marks.

If you run or fund teen health programs, demand two things up front. First, ask authors for effect size numbers, not just p-values. Second, build follow-up probes at 3, 6, and 12 months into your grant. These simple steps turn a flashy demo into data you can trust and replicate.