Risk assessment models could improve support for patients leaving hospital against medical advice

Risk prediction tools might help identify patients at the highest risk of overdose and death after a "before medically advised" (BMA) hospital discharge according to new research in CMAJ (Canadian Medical Association Journal) https://www.cmaj.ca/lookup/doi/10.1503/cmaj.250492.

Patients who leave hospital against the advice of a physician are about twice as likely to die and about 10 times more likely to experience an illicit drug overdose in the first 30 days after leaving hospital. Such BMA discharges are initiated by about 500 000 people in the United States and 30 000 people in Canada every year.

Calculating a specific patient's risk of death and drug overdose - combined with clinical judgment and other risk scores - might help clinicians and patients have a constructive, patient-centred discussion about the decision to initiate a BMA discharge, including assessing the patient's capacity to make such a decision and discussing ways to mitigate risks after BMA discharge. By reducing uncertainty, risk estimates might also reduce clinician moral distress when faced with a BMA discharge."

Dr. Hiten Naik, University of British Columbia, Vancouver, British Columbia, with coauthors

Researchers developed 2 risk prediction models: one to estimate the risk of death from any cause during the 30-day period after a BMA discharge, and another for patients with a history of substance use to estimate the risk of illicit drug overdose in patients with a history of substance use. Using data from British Columbia, investigators examined cohort A, a group of 6440 adults from the general population who initiated a BMA discharge, and cohort B, which included 4466 people with a history of substance use who initiated a BMA discharge.

In cohort A, researchers found that death was less common than generally expected, with 1 death within 30 days for every 63 BMA discharges. Multimorbidity, heart disease, and cancer were strong predictors of death within 30 days of discharge. In cohort B, homelessness, income assistance, opioid use disorder, non-alcohol substance use disorder, drug overdose within the past year, and discharge from a surgical service were strong predictors of drug overdose after BMA discharge.

"Among patients with a history of substance use, illicit drug overdose was a relatively common outcome soon after BMA discharge (i.e., around 1 illicit drug overdose within 30 days for every 19 BMA discharges), suggesting this period is a critical but largely unexplored opportunity for overdose prevention," write the authors.

They suggest that hospitals and health systems could use risk prediction models to automate the approach to higher-risk BMA discharges, with alerts and automatic enrolment in support programs.

"These models offer a starting point for identifying patients who are high risk and may benefit from greater support."