Research published in the Canadian Medical Association Journal indicates that risk prediction tools can help identify patients at increased risk of overdose and death following a “before medically advised” (BMA) hospital discharge. Patients who choose to leave the hospital against medical advice face a significantly heightened risk, with studies showing they are approximately twice as likely to die and ten times more likely to experience an illicit drug overdose within the first 30 days post-discharge.
Each year, around 500,000 individuals in the United States and 30,000 in Canada opt for BMA discharges. Dr. Hiten Naik, from the University of British Columbia, along with co-authors, asserts that understanding a patient’s specific risk level—when combined with clinical judgment—can lead to more informed and patient-centered discussions regarding such decisions.
Risk Prediction Models Developed
The research team created two distinct risk prediction models. The first estimates the risk of death from any cause within 30 days after a BMA discharge, while the second targets patients with a history of substance use to assess the risk of illicit drug overdose. The study analyzed data from British Columbia, focusing on two cohorts: cohort A, comprising 6,440 adults who initiated a BMA discharge, and cohort B, including 4,466 individuals with a history of substance use.
Findings revealed that death occurred less frequently than anticipated in cohort A, with one death for every 63 BMA discharges. Key factors such as multimorbidity, heart disease, and cancer emerged as strong predictors of mortality within the 30-day post-discharge period.
In cohort B, several factors were identified as significant predictors of drug overdose. These included homelessness, reliance on income assistance, opioid use disorder, non-alcohol substance use disorder, a history of drug overdose within the past year, and discharge from surgical services. The data indicated that among patients with a substance use history, illicit drug overdose was a common outcome shortly after a BMA discharge, with an estimated one overdose occurring for every 19 BMA discharges.
Implications for Healthcare Systems
The authors emphasize the urgent need for hospitals and healthcare systems to adopt these risk prediction models. They suggest that automating the assessment of higher-risk BMA discharges, along with implementing alerts and automatic enrollment in support programs, could significantly enhance patient safety.
These models provide a foundational approach to identifying patients who may benefit from additional support and intervention. By accurately predicting risk, healthcare providers can not only improve patient outcomes but also alleviate clinician moral distress associated with BMA discharges.
The research underscores a critical opportunity for overdose prevention during a vulnerable period that has largely been overlooked in healthcare planning. As the healthcare landscape continues to evolve, integrating such predictive tools could lead to more effective and compassionate patient care.
For further details, refer to the study titled “Predicting drug overdose and death after ‘before medically advised’ hospital discharge,” published in the Canadian Medical Association Journal in 2025 (DOI: 10.1503/cmaj.250492).
