Integrating multiple data sources to predict all-cause readmission or mortality in patients with substance misuse

Patients with substance misuse who are admitted to the hospital are at heightened risk for adverse outcomes, such as readmission and death. This study aims to develop methods to identify at-risk patients to facilitate timely interventions that can improve outcomes and optimize healthcare resources. To accomplish this, we leveraged the Substance Misuse Data Commons to predict 30-day death or readmission from hospital discharge in patients with substance misuse. We explored several machine learning algorithms and approaches to integrate information from multiple data sources, such as structured features from a patient’s electronic health record (EHR), unstructured clinical notes, socioeconomic data, and emergency medical services (EMS) data. Our gradient-boosted machine model, which combined structured EHR data, socioeconomic status, and EMS data, was the best-performing model (c-statistic 0.746 [95% CI: 0.732-0.759]), outperforming other machine learning methods and structured data source combinations. The addition of unstructured text did not improve performance, suggesting a need for further exploration of how to leverage unstructured data effectively. Feature importance plots highlighted the importance of prior hospital and EMS encounters and discharge disposition in predicting our primary outcome. In conclusion, we integrated multiple data sources that offer complementary information from data sources beyond the typically used EHRs for risk assessment in patients with substance misuse.Author summary: Our study leverages information from diverse data sources to develop predictive models to identify patients with substance use who are at increased risk of adverse outcomes, such as readmission and death, within 30 days of hospital discharge. We found that a gradient-boosted machine learning model that combined structured electronic health record data, socioeconomic status, and ambulance data achieved the highest performance in predicting these outcomes among patients with substance misuse. Multimodal non-deep and deep learning approaches designed to incorporate unstructured clinical notes did not enhance model performance. However, feature importance analysis revealed that prior ambulance encounters are a key predictor for our model, underlining the importance of using multiple data sources in clinical decision-making for patients with substance misuse.