Pre-treatment and Early Risk Stratification of Adverse Reactions During Oral Anti-obesity Drug Therapy
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Gastrointestinal adverse events and early discontinuation frequently limit the safe initiation and sustained use of oral anti-obesity medications, indicating that efficacy-oriented assessment alone is insufficient for treatment management. To address this gap, this study developed a patient-facing safety screening module that differs from conventional obesity-management tools in three aspects: it performs pre-treatment and early-phase joint risk stratification rather than efficacy-only estimation, integrates dynamic wearable and patient-reported signals in addition to baseline clinical variables, and translates calibrated risk outputs into patient-readable initiation, follow-up, and self-management recommendations. A total of 320 participants were followed for 12 weeks, during which moderate-to-severe adverse events and severity scores were recorded together with liver and renal function, metabolic indices, heart rate variability, sleep status, and patient-reported outcomes. An XGBoost model with Platt scaling was used for probability calibration, and the model was evaluated using stratified five-fold cross-validation with a held-out test set. The calibrated model achieved a ROC-AUC of 0.81 for predicting moderate-to-severe adverse events. SHAP analysis identified postprandial glucose variability, baseline BMI, gastrointestinal discomfort history, and sleep fragmentation as major contributors to risk. The module can operate in parallel with efficacy prediction to support risk-benefit combined screening and patient-centered decision support during oral anti-obesity therapy.
Fields of Study & Subjects
Pharmacology and Obesity Treatment
Artificial Intelligence in Healthcare
Machine Learning in Healthcare
Publication details
- Journal / source
- Zenodo (CERN European Organization for Nuclear Research)
- Publication date
- 28 March 2026
- Language
- EN
- Document type
- Preprint