School Prevention Programs Work Differently for Kids with Different Risk Profiles | Research

The takeaway

A universal school prevention program reduced cannabis use across all risk levels, but the absolute effect was largest for high-risk children, whose predicted cannabis use probability dropped from 33% to 26%.

Read this if you work in youth prevention or are interested in personalized approaches to drug prevention.

High-risk children: probability dropped from 33% to 26% with intervention

What the researchers found

Researchers used a randomized controlled prevention trial with 1,874 sixth-graders (average age 11.8) to predict how a universal prevention program would affect individual children based on their specific risk profiles.

Using eight risk and protective factors, they calculated personalized probabilities of cannabis use for each child with and without the intervention. Low-risk children had predicted probabilities of 4.3% with intervention versus 6.5% without. Moderate-risk children: 10.9% versus 15.3%. High-risk children: 25.5% versus 32.6%.

School grades, thoughts of hurting oneself, and rule-breaking behavior were the three factors that most strongly distinguished high-risk from low-risk children.

Why it matters

This study demonstrates that universal prevention programs can be effective across all risk levels, but their impact varies substantially by individual risk profile. This personalized approach could help target resources more efficiently.

The numbers in context

1,874 students; 33-month follow-up; low-risk: 4.3% vs. 6.5%; moderate-risk: 10.9% vs. 15.3%; high-risk: 25.5% vs. 32.6%; significant differences in all risk groups

How the study worked

School-based randomized controlled trial over 33 months with 1,874 sixth-graders. Two-level random intercept logistic model for panel data. Eight risk/protective factors used to create individualized risk profiles and predict intervention effects.

What this study cannot tell us

Predicted probabilities are model estimates, not observed outcomes. The eight risk factors may not capture all relevant predictors. Czech Republic sample may not generalize to other cultural contexts. Long-term effects beyond 33 months unknown.

How to read the evidence

Randomized controlled trial with individualized risk modeling, though effects are predicted rather than directly observed at the individual level.

When this study was published

Published in 2015. Personalized prevention approaches have continued to develop.

The bigger picture

Moving from population-level to individual-level prediction of prevention effects represents an important advance. Understanding which children benefit most can inform resource allocation and program design decisions.

Questions still open

Can these personalized risk models be used in real-time to customize prevention delivery? Would additional risk factors improve prediction accuracy? Do the prevention effects persist into adolescence and adulthood?

Common questions

What makes a child high-risk for cannabis use?

The three most important factors were poor school grades, thoughts of hurting oneself, and rule-breaking behavior. The model used eight total risk and protective factors to create individualized profiles.

Should prevention only target high-risk kids?

The study found significant effects across all risk levels. Universal programs benefit everyone, but the absolute reduction in cannabis use probability is largest for high-risk children, suggesting both universal and targeted approaches have value.

Read the original research

Cannabis use in children with individualized risk profiles: Predicting the effect of universal prevention intervention.

Addictive behaviors, 50, 110-6

View on PubMed ↗Publisher’s paper ↗

Citation

Miovský, Michal; Vonkova, Hana; Čablová, Lenka; Gabrhelík, Roman. (2015). Cannabis use in children with individualized risk profiles: Predicting the effect of universal prevention intervention.. Addictive behaviors, 50, 110-6. https://doi.org/10.1016/j.addbeh.2015.06.013