Primary Submission Category: Policy
Is a state-level harm reduction policy score similarly predictive of opioid overdose deaths among racialized subgroups? Extending validation of the state Opioid Environment Policy Scale.
Authors: Samantha Doonan Katherine Wheeler-Martin Corey Davis Silvia Martins Magdalena Cerdá
Presenting Author: Samantha Doonan*
Background. Racialized disparities in opioid-related overdose deaths are widening. From 2010 to 2019, the Black-white opioid mortality ratio increased from 0.50 to 0.99.1 Yet, few studies examine if opioid laws are equally effective across different racial/ethnic groups. Methods. As part of the development of an Opioid Environment Policy Scale, we used a modified Delphi process to obtain expert-ratings of 8 opioid laws on Likert scales of very harmful (0) to very helpful (4) and impact (0 to 4). Ratings were collated into policy domains and standardized. We tested whether the harm reduction domain that included naloxone access laws, overdose Good Samaritan laws, and drug-induced homicide laws (the latter reverse-coded) was differentially associated with overdose mortality across racial/ethnic groups. We used 2012-2019 legal data from the Prescription Drug Abuse Policy System and Carroll and colleagues2 and 2013-2020 mortality data from the National Vital Statistics System for 50 states and D.C. We fit Poisson GEE models with state and year fixed effects and state clustering (autoregressive correlation) with policy exposures lagged by one-year. Results. There was no association between the state harm reduction domain score overall and the relative risk of overdose mortality, nor evidence of heterogeneity by race: Black/African American (0.95, 95% CI: 0.88, 1.02), Hispanic/Latino (0.98, 95% CI: 0.79, 1.21), White (1.05, 95% CI: 0.96, 1.15). The three harm reduction laws were also not associated with overdose deaths. Discussion. States rated by experts as having enacted more helpful harm reduction laws did not experience a reduction in overdose deaths associated with these laws. Unmeasured confounding and measurement error in capturing laws and mortality outcomes is possible. We are exploring alternate model specifications and future work will also consider local heterogeneity in implementation.
Reference doi’s:
- 10.1176/appi.ajp.2021.21040381
- 10.2139/ssrn.4171058