How do restrictions on opioid prescribing, harm reduction, and treatment coverage policies relate to opioid overdose deaths in the United States in 2013–2020? An application of a new state opioid policy scale

Abstract

Background

Identifying the most effective state laws and provisions to reduce opioid overdose deaths remains critical.

Methods

Using expert ratings of opioid laws, we developed annual state scores for three domains: opioid prescribing restrictions, harm reduction, and Medicaid treatment coverage. We modeled associations of state opioid policy domain scores with opioid-involved overdose death counts in 3133 counties, and among racial/ethnic subgroups in 1485 counties (2013–2020). We modeled a second set of domain scores based solely on experts’ highest 20 ranked provisions to compare with the all-provisions model.

Results

From 2013 to 2020, moving from non- to full enactment of harm reduction domain laws (i.e., 0 to 1 in domain score) was associated with reduced county-level relative risk (RR) of opioid overdose death in the subsequent year (adjusted RR = 0.84, 95 % credible interval (CrI): 0.77, 0.92). Moving from non- to full enactment of opioid prescribing restrictions and Medicaid treatment coverage domains was associated with higher overdose in 2013–2016 (aRR 1.69 (1.35, 2.11) and aRR 1.20 (1.11, 1.29) respectively); both shifted to the null in 2017–2020. Effect sizes and direction were similar across racial/ethnic groups. Results for experts’ highest 20 ranked provisions did not differ from the all-provision model.

Conclusions

More robust state harm reduction policy scores were associated with reduced overdose risk, adjusting for other policy domains. Harmful associations with opioid prescribing restrictions in 2013–2016 may reflect early unintended consequences of these laws. Medicaid coverage domain findings did not align with experts’ perceptions, though data limitations precluded inclusion of several highly ranked Medicaid policies.