Addiction neuroscience · 9 authors, 4 centres
This summary was generated by AI from a single paper. It has not been reviewed by a clinician and is not clinical advice. Verify against the source before acting on it.
This systematic review analyzed 12 studies to assess the use of machine learning for identifying predictors of smoking cessation outcomes. The review found that while some predictors like delay discounting show promise, machine learning has not yet notably improved smoking cessation paradigms or clinical outcomes.
The systematic review followed PRISMA guidelines and included 12 studies identified from multiple databases through December 2022. Studies used various machine learning techniques to predict smoking cessation outcomes, which were measured by biochemical validation, self-report, or both. The review identified behavioral markers, biomarkers, and psychological factors (e.g., delay discounting, self-efficacy) as predictors, with some aligning with established non-machine learning literature. A key finding is that tree-based models, like classification and regression trees, offered high interpretability, with delay discounting correctly identifying smoking status in 80% of participants post-treatment in one study. Limitations include a reliance on self-reported measures, potential overfitting from small datasets, and a lack of integration of neurobehavioral mechanisms. The review concludes that machine learning approaches are still nascent and have yet to significantly enhance clinical outcomes for smoking cessation.
