Leveraging Machine Learning to Advance Alcohol Research: Current Applications, Challenges, and Opportunities

Area of Science:

• Computational psychiatry
• Data science in healthcare
• Addiction research

Background:

• Machine learning (ML) is increasingly applied to alcohol research.
• Existing studies often face challenges with data limitations and model interpretability.
• Advancing personalized medicine for alcohol use disorder (AUD) requires overcoming these ML challenges.

Purpose of the Study:

• To survey ML approaches in the alcohol literature.
• To review challenges in applying ML to alcohol data.
• To explore how overcoming these challenges can advance personalized medicine for AUD.

Main Methods:

• Systematic literature search of PubMed, ScienceDirect, and EBSCO Academic Search Premier (2015-2025).
• Search terms focused on alcohol/drinking and machine learning/deep learning/prediction/classification.
• Included 110 original peer-reviewed human studies primarily analyzing alcohol consumption behaviors.

Main Results:

• Most ML predictions focused on alcohol consumption or AUD diagnosis in younger cohorts.
• Studies predominantly used conventional ML on single modalities and small datasets.
• Common limitations included small sample sizes and a lack of causal insights.

Conclusions:

• Conventional ML models offer accurate, transparent predictions on small AUD datasets.
• Future research needs to address data limitations and develop models providing causal insights.
• Fostering multidisciplinary teams is crucial for building robust ML models for AUD.