ABSTRACT
OBJECTIVES: To characterize the effects of Michigan's controlled substance legislation on acute care prescriber behavior by specialty, in a single hospital system. DESIGN: A retrospective study of opioid and benzodiazepine prescription records from a hospital electronic medical record system between August 1, 2016, and March 31, 2019, in Detroit, Michigan. SETTING: Discharges from inpatient and emergency department visits. INTERVENTION: Evaluating the impact of implementation of state controlled substance legislation, comparing prescriptions by physicians before, upon, and after June 1, 2018, using regression discontinuity analysis. METHODS: Total daily prescriptions of opioids and total daily prescriptions of benzodiazepine by physicians in the hospital system. Prescriptions were converted to morphine and lorazepam equivalents for comparability. RESULTS: We find 38.5% (95% confidence interval [CI] : 74.1% - 2.9%) decrease of prescription in milligrams of opioid equivalents attributable to implementation of legislation. The main catalyst of the decrease was emergency medicine which experienced 63.9% (95% CI: 109.7%-18.0%) decrease in milligrams of opioid equivalent prescriptions, while surgery increased prescriptions. Though we do not find any statistically significant changes in prescriptions of milligram equivalent of benzodiazepines, we estimate 43.1% (95% CI: 82.6%-3.7%) decrease in count of these prescriptions, implying a significant increase in average dosage of prescriptions. CONCLUSIONS: The introduction of new regulatory requirements for the prescription of controlled substances led to a general decrease in morphine equivalent milligrams prescribed in most specialties, though it may have increased the dosage of benzodiazepine prescriptions. The change in prescription behavior could be motivated by regulatory hassle or by change in attitude towards opioid prescriptions and increased recognition of opioid use disorder.
