Mobile Phone Life-Skills Program Reduced Adolescent Cannabis Use by 45% at 18 Months | Research

The takeaway

A text-message-based life-skills training program delivered to Swiss secondary school students reduced cannabis use prevalence by 45% and tobacco smoking by 33% at 18-month follow-up compared to controls.

School health program administrators; digital health developers; substance prevention researchers.

45% reduction in cannabis prevalence at 18 months (OR=0.55)

What the researchers found

At 18 months, the intervention group had significantly lower cannabis prevalence (OR=0.55, 95% CI 0.39-0.76) and tobacco prevalence (OR=0.67, 95% CI 0.47-0.96) compared to controls. Cannabis use days were also reduced (Cohen's d=-0.19). No effects were seen for problem drinking, alcohol quantity, or social skills.

Why it matters

Digital prevention programs that work are rare. This study shows a fully automated, text-message-based program can produce lasting reductions in cannabis and tobacco use, which is significant because such programs can be scaled at minimal cost.

The numbers in context

1,473 students, 89 school classes. 83.6% follow-up at 18 months. Cannabis prevalence: OR=0.55 (95% CI 0.39-0.76). Cannabis days: d=-0.19 (95% CI -0.29, -0.09). Tobacco: OR=0.67 (95% CI 0.47-0.96). No significant effects for problem drinking (OR=0.84), alcohol quantity, cigarette quantity, well-being, or social skills.

How the study worked

Two-arm cluster RCT with 1,473 students (mean age 15.4) across 89 Swiss school classes. The intervention consisted of 22 weeks of automated online feedback and individually tailored text messages based on social cognitive theory, addressing self-management, social skills, and substance resistance. Follow-up at 6 and 18 months.

What this study cannot tell us

Conducted in Swiss secondary schools, so cultural context may affect generalizability. Self-reported outcomes. The intervention period (22 weeks) is substantial, and engagement may have varied. Effect sizes for cannabis use days were small (d=-0.19).

How to read the evidence

Strong: cluster RCT with adequate sample size, 83.6% follow-up at 18 months, and pre-registered design.

When this study was published

Published in 2022, with students recruited in 2019/2020.

The bigger picture

The selective efficacy (cannabis and tobacco but not alcohol) mirrors other prevention programs and suggests that different substances may require different intervention strategies. The fully automated delivery makes this highly scalable across school systems.

Questions still open

Why does the program work for cannabis and tobacco but not alcohol? Would adaptation for other cultural contexts maintain efficacy? Do the effects persist beyond 18 months? What level of student engagement is needed for the program to be effective?

Common questions

How does a text message program prevent drug use?

The program delivered individually tailored messages over 22 weeks covering self-management skills, social skills, and resistance strategies. By meeting students on their phones rather than in classroom settings, it provided ongoing support and reinforcement.

Why did it work for cannabis but not alcohol?

The researchers did not identify a clear reason. One possibility is that cannabis and tobacco initiation may be more influenced by the self-management and resistance skills targeted by the program, while adolescent drinking patterns may be more socially embedded.

Read the original research

Longer-Term Efficacy of a Digital Life-Skills Training for Substance Use Prevention.

American journal of preventive medicine, 63(6), 944-953

View on PubMed ↗Publisher’s paper ↗

Citation

Paz Castro, Raquel; Haug, Severin; Wenger, Andreas; Schaub, Michael P. (2022). Longer-Term Efficacy of a Digital Life-Skills Training for Substance Use Prevention.. American journal of preventive medicine, 63(6), 944-953. https://doi.org/10.1016/j.amepre.2022.06.017