ML-Based Prediction of Drinking Behavior

Using machine learning to predict post-treatment drinking behavior in patients with alcohol use disorder.

This project, funded by the Swiss National Science Foundation (SNSF, 2022–2025), investigates how machine learning models can be used to predict post-treatment drinking outcomes in patients with alcohol use disorder (AUD).

Objectives

  • Develop predictive models for post-treatment drinking behavior using clinical and experimental data from a multicenter randomized controlled trial.
  • Identify key predictors and their interactions that contribute to treatment success or relapse.
  • Translate findings into actionable insights for personalized treatment planning.

Methods

We leverage a range of supervised learning techniques — from regularized regression to ensemble methods — applied to rich longitudinal data collected during and after inpatient alcohol treatment.

  • Stein et al. (2022). Alcohol-Specific Inhibition Training in Patients with AUD: A Multicenter, Double-Blind, Randomized Clinical Trial. Addiction.
  • Jaeger et al. (2024). Antidepressants and AUD: A Multicenter Study on the Mediating Role of Depression Symptom Changes. Alcohol: Clinical and Experimental Research.