(26) Overdose after Medication for Opioid Use Disorder Initiation Following Hospitalization

Background & Introduction: Opioid-related hospitalizations have increased markedly in the past several years, from 136.8 hospitalizations per 100,000 population in 2005 to 250.4 per 100,000 population in 2020 (1). Hospitalizations related to opioid use disorder (OUD) comprised about 3.4% of all hospitalizations in metro areas in 2016 (2). These hospitalizations are not always related to overdose, but rather the spectrum of morbidity that accompanies substance use, including HIV, hepatitis, endocarditis, skin and soft tissue infections, and bacteremia (3). Although hospitalization is an undesirable outcome for any individual, it represents an opportunity to engage patients in OUD treatment. Starting medication for OUD (MOUD) in the hospital should lead to improved outcomes, although it is not known if initiating MOUD during a hospitalization for patients not previously taking it is associated with decreased incidence of future overdoses. The aim of this study, therefore, is to leverage a comprehensive linked public health dataset in Oregon that includes hospital visits, prescription drug monitoring program medication dispensations, medication-related billing data, and mortality to determine if starting buprenorphine after hospitalization is associated with decreased odds of overdose at six and twelve months.

Methods: This is a retrospective observational study of the Oregon Comprehensive Opioid Risk Registry (CORR) database, based upon the Oregon All Payer Claims Database (APCD), which includes Medicaid, Medicare and commercial insurance claims for approximately 80% of Oregonians from 2013-2020. Patients in the APCD were probabilistically linked to the Oregon PDMP, vital records, emergency medical services, and hospital discharge databases (4). We identified patients with an OUD diagnosis in a hospitalization or emergency department (ED) visit in 2017-2019. We used the first visit during the time period. We then excluded those with an OUD-related hospitalization or ED visit in the previous six months, or with OUD treatment in the previous six months. Obtaining MOUD treatment (receiving an OUD formulation for buprenorphine, naltrexone, or a procedure code for methadone dispensation) in the seven days after discharge from hospitalization was the main exposure. Patients were followed 6 months and 12 months post hospital discharge and assessed for first occurrence of a fatal or non-fatal opioid-related overdose. Fatal opioid overdose events were captured using underlying cause-of-death codes. Non-fatal opioid overdose events were captured from hospital discharge records and APCD ED insurance claims. The Mass General Brigham Human Research Committee deemed this study exempt.

Results: There were 22,235 patients who had an included OUD-related hospitalization during the study period. 11,808 (53.1%) were female. 6.8% were age < 25 years, 25.0% were 25-39 years, 19.4% were 40-54 years, 37.0% were 55-75 years and 11.9% were >75 years old. The most prevalent payers were Medicaid (48.5%), followed by Medicare (28.9%), dually enrolled (11.6%) and commercial (7.7%). Most patients (58.4%) had 3 or more comorbidities coded. 1,184 (5.3%) patients received MOUD within seven days of hospitalization. Of these, 683 (57.7%) received buprenorphine, 463 (39.1%) received methadone and 46 (3.9%) received long-acting injectable naltrexone. There were 452 overdose events in 6 months and 758 overdose events in 12 months. After adjusting for age, sex, payer and number of comorbidities, patients who received MOUD within seven days had a lower adjusted odds ratio (aOR) of fatal or nonfatal overdose at six months compared with those who did not (aOR 0.63, 95% confidence interval (CI) 0.41-0.97). At twelve months, the aOR of fatal or nonfatal overdose was lower but did not reach statistical significance (aOR 0.79, 95% CI 0.58-1.08). Looking specifically at fatal overdoses, there were 46 events in six months and 76 events at twelve months. Again, the findings suggest an effect but were not specifically significant: aOR of fatal overdose at six months was 0.63 (95% CI 0.15-2.66) and aOR of fatal overdose at twelve months was 0.73 (95% CI 0.26-2.02). When evaluating specific agents received, patients had a decreased risk of fatal or nonfatal overdose at 6 months with buprenorphine (aOR 0.50, 95% CI 0.27-0.95) but decreased odds associated with methadone did not reach statistical significance (aOR 0.57, 95% CI 0.28-1.17). Outcomes related to naltrexone are not reported due to the small number of events (4 overdoses at 6 months).

Conclusion & Discussion: This study indicates that starting MOUD within seven days of hospitalization was associated with a decreased risk of fatal or non-fatal overdose at 6 months. The results suggested decreased odds at twelve months and specifically for fatal overdose as well, although these findings did not reach statistical significance. Still, the lower odds of overdose at six months is notable and will hopefully encourage commencement of MOUD for hospitalized patients. That only 5.3% of patients had evidence of MOUD initiation indicates that there is opportunity to improve. Society of Hospital Medicine guidelines (5) and others now recommend initiation of MOUD for patients hospitalized with OUD-related issues, and we hope this work will further encourage adoption. There are limitations to the study, including its retrospective observational design, that it relies on billing code data, and that it is limited to hospitalizations, medication dispensation and outcomes in a single state. Still, the findings suggest that initiation of MOUD after hospitalization is effective at reducing opioid-related overdose at six months.

References: Healthcare Cost and Utilization Project (HCUP) Fast Stats. Available at: https://datatools.ahrq.gov/hcup-fast-stats/. Accessed October 12, 2023.

Owens, P. L., Weiss, A. J., & Barrett, M. L. (2020). Hospital Burden of Opioid-Related Inpatient Stays: Metropolitan and Rural Hospitals, 2016. In Healthcare Cost and Utilization Project (HCUP) Statistical Briefs. Agency for Healthcare Research and Quality (US).

Schwetz, T. A., Calder, T., Rosenthal, E., Kattakuzhy, S., & Fauci, A. S. (2019). Opioids and Infectious Diseases: A Converging Public Health Crisis. The Journal of infectious diseases, 220(3), 346–349.

Hallvik, S. E., Dameshghi, N., El Ibrahimi, S., Hendricks, M. A., Hildebran, C., Bishop, C. J., & Weiner, S. G. (2021). Linkage of public health and all payer claims data for population-level opioid research. Pharmacoepidemiology and drug safety, 30(7), 927–933.

Calcaterra, S. L., Martin, M., Bottner, R., Englander, H., Weinstein, Z., Weimer, M. B., Lambert, E., & Herzig, S. J. (2022). Management of opioid use disorder and associated conditions among hospitalized adults: A Consensus Statement from the Society of Hospital Medicine. Journal of hospital medicine, 17(9), 744–756.