Effects of a State-Level Policy Requiring Hospitals to Provide Treatment for Opioid Use Disorder

• In Event: From Funding to Outcomes: Evaluating Opioid Policy, Spending, and Treatment Access

Abstract

The United States opioid crisis continues, with recent relative declines in opioid overdose mortality reflecting returns to pre-pandemic levels. Substance Abuse and Mental Health Services Association (SAMHSA) reports suggest approximately 1% of individuals in the United States over age 12 suffer from an opioid use disorder (OUD). Medications for opioid use disorder (MOUD) are lifesaving and life changing. However, connecting patients with OUD to MOUD may be complicated. Because many patients with OUD frequently utilize hospitals for health care, clinical evidence suggests that emergency departments and inpatient settings are crucial hubs for initiating MOUD and connecting patients to sustained outpatient treatment.This study investigates the effects of a policy that required Massachusetts hospitals with emergency departments to be capable of initiating/administering MOUD by exploring stages of MOUD treatment in the Medicare population. I examine (1) access to providers capable of prescribing MOUD, (2) hospital utilization among patients with OUD/on MOUD, (3) setting of MOUD initiation, and (4) overall treatment of OUD with MOUD following policy passage. I use SAMHSA Drug Enforcement Agency controlled substance prescription waiver data and inpatient and emergency department Medicare claims from July 2016-December 2019 (prior to the COVID-19 pandemic). First, I estimate the effects of the policy in Massachusetts on the quarterly rate of providers capable of prescribing MOUD per 10,000 state population compared to all other states utilizing synthetic-difference-in-differences estimation (Arkhangelsky, et al.), finding a but significant increase in the rate of MOUD-capable providers [ATT: 0.599/10,000 state population, p-value 0.001, SE: 0.187]. Second, I utilize a linear model with patient and quarterly two-way fixed effects to examine the impact of the policy on the logarithm of quarterly overall, emergency department (ED), and inpatient (IP) hospital visits among patients with an OUD diagnosis or on MOUD within a year prior to the hospital encounter. I find patients with OUD and on MOUD experience a small but significant approximate 1% increase in hospital visits  [OUD: 0.01, p-value 0.003, 95% CI: 0.01, 0.01 | MOUD: 0.01, p-value 0.018, 95% CI: 0.00, 0.01], which appears to be driven by IP and not ED visits. Third, I use synthetic-difference-in-differences models to estimate policy effects on the quarterly percentage of Medicare patients initiated on MOUD following an ED or IP visit. I do not find a significant difference in the percentage of patients initiated on MOUD following a hospital encounter in Massachusetts relative to other states. Fourth, I estimate the policy effect on the likelihood a patient diagnosed with OUD in a given month is treated with MOUD using logit estimation with patient and quarter fixed effects. I find a statistically significant increase in the likelihood that patients with OUD receive MOUD in a given month [Odds Ratio: 2.149, p-value < 0.001].Study results show that the policy affected treatment of OUD with MOUD in Massachusetts, although the significance of MOUD initiation during or following a hospital encounter remains unclear. Subsequent work investigates whether access to MOUD in the inpatient setting improves retention in care and reduces patient-directed discharge against medical advice.

Author

• Victoria Elizabeth Bethel, University of Georgia
  Presenting Author