Oral opioid prescribing trends in the United States, 2002-2018

Abstract

Objective. To conduct a retrospective analysis of sequential cross-sectional data of opioid prescribing practices in patients with no prior history of opioid use. Methods. Individuals filling an oral opioid prescription who had 1 year of prior observation were identified from four different administrative claims databases for the period between January 1, 2002, and December 31, 2018: IBM MarketScanVR Commercial Database (CCAE), Multi-State Medicaid Database (MDCD), Medicare Supplemental Database (MDCR), and OptumVC De-IdentifiedClinformaticsVR Data Mart Database. Outcomes included incidence of new opioid use and characteristics of patients' first opioid prescription, including dispensed morphine milligram equivalent (MME) per day, total MME dispensed, total MME ≤300, and days' supply of prescription for ≤3 or ≤30 days. Results. There were 40,600,696 new opioid users identified. The incidence of new opioid use in the past 17 years ranged from 6% to 11% within the two commercially insured databases. Incidence decreased over time in MDCD and was consistently higher in MDCR. Total MME dispensed decreased in MDCD and increased in CCAE, with no major changes in the other databases. The proportion of patients receiving ≥30-day prescriptions decreased and the proportion of patients receiving ≤3-day prescriptions increased in MDCD, while ≤30-day prescriptions in the Optum database dramatically increased (low of 3.0% in 2003 to peak of 16.9% in 2017). Conclusions. Opioid prescribing practices varied across different populations of insured individuals during the past 17 years. The most substantial changes in opioid prescriptions over time have occurred in MDCD, with reductions in use across multiple metrics.