Must-access PDMPs and retail opioid sales
Explainer of Shakya et al. (2022), Contemporary Economic Policy. Not the journal article.
Citation: Shakya, S., & Hodges, C. (2022). Must access prescription drug monitoring programs and retail opioid sales. Contemporary Economic Policy. Free PDF
Facts¶
Main finding¶
Must-access PDMP implementation does not uniformly reduce retail opioid sales. Certain opioids show sales declines: hydrocodone, hydromorphone, and morphine all decrease after must-access adoption. However, oxycodone - the most widely prescribed opioid - shows no statistically significant change. Methadone and buprenorphine (medications for opioid use disorder treatment) also remain unaffected. Effects are heterogeneous across drug types.
One caveat¶
Treatment effects are not uniform across opioid types, suggesting heterogeneous responses by prescriber behavior and patient demand. The analysis cannot explain why oxycodone is unaffected while other opioids decline. Policy makers should treat must-access PDMPs as one part of a wider opioid response rather than as a singular policy solution.
Policy hook¶
PDMPs are state-level electronic databases tracking controlled-substance fills, intended to curb opioid abuse and overprescribing. Most states started with voluntary access, but many transitioned to “must-access” mandates requiring prescribers to check the database before dispensing. The question is whether must-access rules actually reduce retail opioid sales, and whether effects differ across opioid types.
Data and setting¶
ARCOS database (Drug Enforcement Administration opioid sales tracking), 2006-2017, covering eight commonly prescribed opioids tracked at the state-year level. Eighteen states implemented must-access PDMP mandates during the study period; remaining states kept voluntary access throughout (except Missouri, which had no statewide PDMP). Outcome is total retail opioid sales by drug type and state-year.
Research design (plain language)¶
The paper uses difference-in-differences estimation with an interactive fixed effects model (Bai 2009) to account for unobserved time-varying confounders and non-linear state-specific trends. Controls include Good Samaritan laws, marijuana legalization, naloxone access laws, pill mill laws, Medicaid expansion, and state social and economic variables. The design exploits staggered adoption of must-access mandates across states.
PDF or DOI¶
Why it matters¶
Must-access PDMPs are a frequently implemented policy response to opioid overprescribing, yet evidence on their effectiveness at the retail sales level has been limited. This paper shows that must-access mandates reduce sales of some opioids but not others - a finding with important consequences for policy design. The fact that oxycodone sales remain unaffected despite being the most widely prescribed opioid suggests that prescriber behavior or demand is not uniformly responsive to database checking requirements. This variation across drugs raises questions about what drives differential responses and whether prescribers are substituting toward less-regulated medications. For policy makers, the heterogeneous effects suggest that must-access PDMPs are a necessary but insufficient policy tool. They should be paired with other programs (education, prescriber feedback, naloxone distribution, treatment access) to achieve meaningful reductions across all opioid types. The work also highlights the importance of disaggregating policy results by drug type rather than treating opioids as a uniform category.