Multi-expert models improve opioid overdose risk prediction accuracy
OverdoseMoE: A Multi-Expert Framework for Opioid Overdose Risk Prediction
Computation and Language
Summary
Opioid overdose is a serious health problem, and it’s important to identify people who might be at risk. The authors studied how using specialized models that focus on different diagnoses can better predict who might overdose in the next six months by looking at patients’ medical history. They created a new system called OVERDOSEMOE that combines expert models to make more accurate predictions than just one general model. This approach was tested on different groups of patient data and showed it could reliably find high-risk individuals.
What this means in practice
- •For hospital data teams: Identify patients at high risk for opioid overdose six months ahead by analyzing one-year clinical diagnostic histories with multi-expert language models.
- •For healthcare software developers: Integrate multi-scale, diagnosis-specific adapted language models into electronic health record systems to improve opioid overdose risk stratification accuracy.
Authors
Mingchen Li, Rohan Pandey, Junhui Qian, Feiyun Ouyang, Sunjae Kwon, Avijit Mitra, Zonghai Yao, Hong Yu
Abstract
Opioid overdose remains a major clinical and public health burden, highlighting the need for scalable approaches to identify patients at high risk. Here, we investigate diagnosis-specific adaptation for 180-day opioid overdose risk prediction from patients' preceding one-year longitudinal ICD histories. We develop OODMAMBA and OODQWEN through continued pretraining on longitudinal diagnostic sequences followed by task-specific fine-tuning. Building on the stronger Qwen-based predictors, we further propose OVERDOSEMOE, a multi-expert framework that integrates models of different scales using complementary expert-weighting strategies. Diagnosis-specific adaptation consistently improved predictive performance over general-purpose language-model baselines, with OODQWEN achieving an AUPRC of 24.47 and an AUROC of 68.56. OVERDOSEMOE further improved discrimination and precision, achieving an AUPRC of 25.17 and an AUROC of 69.49 while outperforming the strongest single-model baselines. Among patients ranked in the top 5% of predicted risk, OVERDOSEMOE identified substantially enriched overdose risk, achieving a PPV of 25.38% while retaining meaningful recall. Evaluation on an independent MIMIC-IV cohort further demonstrated cross-cohort robustness, with complementary weighting strategies showing advantages across different performance measures. These findings demonstrate that diagnosis-specific language-model adaptation combined with multi-expert integration can improve opioid overdose risk stratification and support more robust prediction across heterogeneous electronic health record populations.