Evaluation of a novel metric for personalized opioid prescribing after hospitalization

Background: The duration of an opioid prescribed at hospital discharge does not intrinsically account for opioid needs during a hospitalization. This discrepancy may lead to patients receiving much larger supplies of opioids on discharge than they truly require. Objective: Assess a novel discharge opioid supply metric that adjusts for opioid use during hospitalization, compared to the conventional discharge prescription signature. Design, setting, & participants: Retrospective study using electronic health record data from June 2012 to November 2018 of adults who received opioids while hospitalized and after discharge from a single academic medical center. Measures & analysis: We ascertained inpatient opioids received and milligrams of opioids supplied after discharge, then determined days of opioids supplied after discharge by the conventional prescription signature opioid-days (“conventional days”) and novel hospital-adjusted opioid-days (“adjusted days”) metrics. We calculated descriptive statistics, within-subject difference between measurements, and fold difference between measures. We used multiple linear regression to determine patient-level predictors associated with high difference in days prescribed between measures. Results: The adjusted days metric demonstrates a 2.4 day median increase in prescription duration as compared to the conventional days metric (9.4 vs. 7.0 days; P