PROJECT SUMMARY Despite prevention efforts and reduced opioid prescribing, U.S. overdose rates continue to rise. Millions of patients prescribed opioids for chronic pain are co-prescribed benzodiazepines, which heighten overdose risk. However, tapering either opioids or benzodiazepines may also increase overdose and mental health risks, and clinicians lack evidence-based guidance about how to safely reduce risk in patients who are co- prescribed opioids with benzodiazepines. Rates of fatal and non-fatal overdose involving prescribed opioids or benzodiazepines combined with illicit opioids or psychostimulants are also rapidly increasing. The specific aims of this proposed study are: Aim 1) To identify risk and protective factors, prescribing patterns, and trajectories associated with progression to long-term benzodiazepine use among patients prescribed long- term opioids. This aim addresses root causes of long-term opioid-benzodiazepine co-prescription. Aim 2) To identify risk and protective factors, prescribing patterns, and trajectories associated with overdose involving prescription opioids or benzodiazepines among patients co-prescribed long-term opioids and benzodiazepines. Findings from this Aim will inform prescription overdose prevention efforts in the large population of patients who are co-prescribed opioids and benzodiazepines. Aim 3) To identify risk and protective factors, prescribing patterns, and trajectories associated with overdose involving either illicit opioids or psychostimulants among patients co-prescribed long-term opioids and benzodiazepines. This Aim will identify factors that may prevent the progression from long-term use of prescribed opioids and benzodiazepines to use of illicit opioids or stimulants. We will accomplish these aims by conducting retrospective cohort studies using two large, complementary data sources: 1) the California Prescription Drug Monitoring Program (PDMP) data linked to death certificates, and 2) the OptumLabs Data Warehouse, a national claims data source representing over 20% of the U.S. commercial and Medicare Advantage market. As the PDMP are linked to death records, the California statewide analyses will focus on fatal overdoses, while the OptumLabs claims analyses will focus on non-fatal overdoses. By capitalizing on the statistical power achieved by these large data sources, this project will elucidate risk and protective factors, drug-use patterns, and trajectories associated with initiation of benzodiazepines among patients prescribed long-term opioids and will identify factors associated with risk of overdose from prescribed opioids, benzodiazepines, or illicit opioids or psychostimulants. Findings will richly inform clinical or policy interventions and guidelines to prevent initial benzodiazepine-opioid co-prescribing and to mitigate overdose risk in patients with established co-prescriptions.