The Impact of Abuse-Deterrent Formulation of Extended-Release OxyContin on Non-Medical Use of Prescription Pain Relievers and Heroin Use

Background: The introduction of the abuse-deterrent formulation (ADF) of extended-release OxyContin in 2010 was a supply-side policy intended to deter and reduce non-medical use of the product. However, preliminary evidence suggests that the reformulation might have had unintended consequences, such as an increase in the initiation and utilization of heroin and heroin-related deaths. Our study explores whether there is any evidence that the introduction of ADF OxyContin triggered substitution from prescription pain relievers to heroin.

Data and Sample: Our primary data source is the National Survey on Drug Use and Health (NSDUH) from 2005-2014. Our study sample is comprised of individuals age 18 and above who engaged in misuse of any prescription pain reliever prior to the introduction of ADF OxyContin based on their reported year of first use (unadjusted N = 81,400). In addition, we have collected data on the implementation of prescription drug monitoring programs (PDMPs) in each state, as well as U.S. Census Bureau data to identify areas of residence as urban or rural.

Outcome Measures: We consider five dichotomous primary outcomes in this study: (1) any self-reported prescription pain reliever misuse; (2) prescription pain reliever use disorder (based on DSM-IV criteria); (3) any self-reported heroin use; (4) heroin use disorder (based on DSM-IV criteria); and (5) self-reported heroin initiation.

Analysis: We employ multivariate logistic regression under a difference-in-differences (DID) framework. The treatment group in our analysis is comprised of individuals who reported misusing OxyContin in any year prior to and including 2010 (unadjusted n = 17,000). The control group is comprised of individuals who reported misusing other prescription pain relievers (unadjusted n = 64,400) in the same timeframe. We also control for demographic and socioeconomic characteristics, prescription drug monitoring programs (PDMPs), and state-level fixed effects. We conduct multiple sensitivity analyses to ensure that our models are robust.

Results and Conclusions: We find a net reduction in the misuse of prescription pain relievers (DID interaction term OR: 0.791, p<0.001) and heroin initiation (DID interaction term OR: 0.422, p=0.011) after the reformulation for our treatment group relative to our control group. We find no statistically significant effects of the reformulation on our other outcome measures. These results indicate that abuse-deterrent OxyContin was effective at deterring misuse of prescription pain relievers without encouraging individuals to substitute to heroin.