Linking data sources to elucidate non-fatal and fatal opioid-related overdose epidemiology and the role of FDA-regulated products

CERSI Collaborators: Anne Black, PhD (Yale), William Becker, MD (Yale), Gail D’Onofrio, MD, MS (Yale), David Fiellin, MD (Yale), Lauretta Grau, PhD (Yale), Kathryn Hawk, MD, MHS (Yale), Robert Heimer, PhD (Yale), Benjamin Howell, PhD, MPH (Yale) (PI)

FDA Collaborators: Yulan Ding, MS, Catherine Dormitzer, PhD, MPH, Christina Greene, PhD, Hana Lee, PhD, MS, Mark Liberatore, PharmD, RAC; Rita Noel, Yueqin Zhao, PhD, MS

CERSI Subcontractors: University of Connecticut: Linda Frisman, PhD, Hsiu-Ju Lin, PhD

Project Start: June 2017

Regulatory Science Challenge

Opioid-involved overdose, or the syndrome where opioids, with or without other drugs, cause respiratory depression and loss of consciousness/coma is the leading cause of accidental death in the United States. Medications, including diverted prescription medications and not only drugs originating in the illicit market, are frequently involved in opioid overdoses. As such, the FDA has prioritized examining the role of prescription medications, particularly opioids and benzodiazepines in the overdose crisis. One of the main challenges in this work is that it has been difficult to assess what medications individuals may have had access to or were taking prior to an overdose event. The primary reason for this difficulty is that there are many different sources of data on medication receipt and outcomes of interest, such as fatal and nonfatal overdose events, and it is challenging to link this information in the same person.

Project Description and Goals

In this project, researchers will work to overcome the challenges of studying drug overdoses by linking Connecticut-based prescription drug monitoring program, criminal justice, drug treatment, and hospital data with the Connecticut Office of the Chief Medical Examiner dataset. This will enable a more accurate assessment of medication and health services utilization that may have increased or decreased patient risk for overdose. The goals of this project include:

1. Gain access to, clean, and link the necessary datasets.
2. Determine the extent to which opioid-related overdose victims as identified through hospital and medical examiner datasets are represented in state-based Prescription Drug Monitoring Program, criminal justice, hospital, and drug treatment databases.
3. Identify specific FDA-regulated products, prescribing, and service utilization profiles associated with opioid-related overdoses.

The project will also be expanded to:

1. Determine patterns of opioid and benzodiazepine receipt via prescription drug monitoring program and addiction treatment system data in three months prior to opioid-involved deaths and concordance with post-mortem toxicology.
2. Examine association between the risk of subsequent fatal and non-fatal overdose and demographics, opioid and benzodiazepine prescription receipt, and substance use disorder treatment, including medications for opioid use disorder (MOUD) and non-MOUD-based treatment, following a non-fatal overdose.

Accessing datasets is the largest barrier to linking data sources. Without a legislative mandate for coordinated effort across state agencies, several different Memoranda of Understanding (MOU) need to be established that require a lengthy internal state agency legal review, resulting in over a year’s time to obtain approvals. In addition, other than the dataset obtained from the Office of the Chief Medical Examiner, all other datasets were considered protected personal health information, creating legal challenges related to patient privacy. Overall, during phase one of this project, Yale-Mayo CERSI linked several datasets. The creation of a unified database linking state-level administrative claims data on nonfatal overdoses, mortality data, substance use treatment data, and controlled substances prescription dispensing data in a timely and ongoing fashion served as a proof of concept for future efforts to enhance states’ capabilities to monitor the public health impact of controlled substances, identify and implement interventions, and assess their effectiveness.

Once this linked database in Connecticut was established, in the second phase of this project, the CERSI investigators determined that, between May 2016 and December 2017, 1 in 3 people with an opioid-involved overdose death in Connecticut were prescribed and dispensed an opioid in the 90 days prior to death and 1 in 4 had received a prescribed benzodiazepine. In 17% of these opioid-involved deaths, there was concordance between a recently prescribed opioid and post-mortem toxicology, and in 21% there was concordance between a recently prescribed benzodiazepine and post-mortem toxicology. Lastly, preliminary results have also suggested that people who were prescribed and received either buprenorphine or methadone following a non-fatal overdose had a lower risk of a subsequent overdose compared to people who did not receive these medications following a non-fatal overdose.

Publications

Becker WC, Heimer R, Dormitzer CM, Doernberg M, D'Onofrio G, Grau LE, Hawk K, Lin HJ, Secora AM, Fiellin DA. Merging statewide data in a public/university collaboration to address opioid use disorder and overdose. Addict Sci Clin Pract. 2021 Jan 4;16(1):1. doi: 10.1186/s13722-020-00211-9.

Howell BA, Black AC, Grau LE, Lin HJ, Greene C, Lee H, Heimer R, Hawk KE, D'Onofrio G, Fiellin DA, Becker WC. Concordance between controlled substance receipt and post-mortem toxicology in opioid-detected overdose deaths: A statewide analysis. Drug Alcohol Depend. 2023 Mar 1; 244:109788. doi: 10.1016/j.drugalcdep.2023.109788. Epub 2023 Jan 30. PMID: 36738634; PMCID: PMC9975083.