Unintentional injury is the fourth leading cause of death among the U.S. public, primarily due to overdose deaths and followed by motor vehicle traffic deaths; together, these sources of unintentional injury resulted in over 147,000 deaths in 2021.[1] The two issues intersect with rising rates of opioid-involved fatal crashes, with the prevalence of fatally injured drivers testing positive for opioids increasing from 1% in 1995 to 7% in 2016,[2] and most recently estimated at 14% during 2020.[3, 4] Use of opioids impairs motor functioning skills that are necessary for operating a motor vehicle,[5] and drivers’ use of opioids has been shown to increase their likelihood of initiating fatal car crashes.[6–8]

Rates of opioid prescribing peaked in 2011 but remain over twice as high as prescribing rates in 1999.[9] While the ongoing opioid epidemic is presently driven by synthetic opioids, such as fentanyl, overdose deaths involving prescription opioids in 2021 were still nearly five times the number in 1999.[10] Multiple policies have been put in place to curb the high rates of opioid prescribing that drove the early phase of the opioid epidemic in the U.S., but there is limited evidence on how these policies have influenced fatal motor vehicle crashes.

Since 2010, there have been 4 primary types of state laws designed to curb potentially risky opioid prescribing:[11–15] opioid prescribing cap laws limiting the dose or duration of opioid prescriptions; mandatory prescription drug monitoring program (PDMP) enrollment laws requiring opioid prescribers to enroll in their state’s PDMP, giving them access to—but not requiring them to check—the PDMP database; PDMP query laws requiring prescribers to check the PDMP before prescribing an opioid; and pill mill laws regulating pain management clinics to prevent rogue clinics from prescribing opioids without medical indication. Research on how these laws affect opioid prescribing show mixed results due to variation in data, study design, and populations examined.[16–18] Reviews have found limited evidence to support associations between PDMP and pill mill laws and reductions in high-risk opioid prescribing practices, such as high dose prescribing.[16–18] In contrast, opioid prescribing laws have been associated with reductions in prescribing for acute,[16, 19] but not chronic,[15, 20] pain. Studies on mandatory PDMP query and pill mill laws implemented concurrently in Florida have suggested a reduction in fatal vehicle crashes attributable to these laws.[21, 22] In this study, we build upon this prior research to examine the association between state mandatory PDMP enrollment, mandatory PDMP query, pill mill, and opioid prescribing cap laws on fatal crashes.

The state-specific analytic sample sizes and balance in prelaw characteristics between the treatment state and weighted synthetic control of comparison states are shown in Appendix Section A. There was no evidence of nonparallel pre-period trends for outcomes in the augmented synthetic control analyses, shown in the fatal crash time-series graphs in Appendix Section B and the opioid-involved fatal crash time-series graphs in Appendix Section C.

Results of augmented synthetic control analyses showed small-in-magnitude, non-statistically significant changes in fatal crash outcomes attributable to state opioid prescribing laws. In each month post-law, the 13 opioid prescribing laws states were associated with a change of less than 0.50 in the rate of fatal crashes per 100,000 licensed drivers, with 95% CIs not exceeding 1 fatal crash per 100,000 licensed drivers per month in any state. (Figure 1). Changes ranged from a −0.30 (95% CI, −1.17 to 0.57) decrease in the rate of fatal crashes per 100,000 licensed drivers attributable to Mississippi’s pill mill law per month, to a 0.14 (95% CI, −0.32 to 0.60) increase in the rate of fatal crashes per 100,000 licensed drivers attributable to Ohio’s prescribing cap law per month. The difference-in-differences calculations that produced the outcomes are shown in Appendix Section D. For example, the estimated pre-law rate in Delaware and its synthetic control was 1.33 fatal crashes per 100,000 licensed drivers per month. After implementation of the prescribing cap law, the estimated monthly rate of fatal crashes per 100,000 licensed drivers in Delaware was 1.14, and 1.22 in the synthetic control ([1.14–1.33]-[1.22–1.33]=−0.08). The result of this difference-in-difference is presented in manuscript figures.

Similar small-in-magnitude and non-statistically significant changes in opioid-involved fatal crashes attributable to the laws were observed. In each month during the post-law period, the 13 opioid prescribing laws states were associated with a change of less than 0.10 in the rate of opioid-involved fatal crashes per 100,000 licensed drivers, with 95% CIs not exceeding 0.20 opioid-involved fatal crashes per 100,000 licensed drivers per month (Figure 2). Changes ranged from a −0.01 (95% CI, −0.10 to 0.08) decrease in the rate of opioid-involved fatal crashes per 100,000 licensed drivers attributable to Mississippi’s pill mill law per month, to a 0.10 (95% CI, −0.21 to 0.40) increase in the rate of opioid-involved fatal crashes per 100,000 licensed drivers attributable to Kentucky’s prescribing cap law per month. The difference-in-differences calculations that produced these outcomes are shown in Appendix Section E. Results of all sensitivity analyses were consistent with the main findings, except for one statistically significant (p<0.05) decrease of −0.21 (−0.34, −0.09) in drug-involved fatal crashes per 100,000 licensed drivers per month attributable to New York’s opioid prescribing law. Sensitivity analysis results are shown in Appendix Section F.

We did not find an association between state opioid prescribing laws and rates of fatal crashes. Across the 13 treatment states, small-in-magnitude and non-statistically significant changes attributable to the laws were observed for the rates of opioid-involved fatal crashes, drug-involved fatal crashes, single vehicle crashes, and fatally injured drivers over 65 years old. The magnitude of non-statistically significant changes in the rate of fatal crashes attributable to the laws ranged from an increase of 0.14 (−0.32, 0.60) fatal crashes per 100,000 licensed drivers attributable to Ohio’s opioid prescribing cap law, to a decrease of 0.30 (−1.17, 0.57) fatal crashes/100,000 licensed drivers attributable to Mississippi’s pill mill law. Previous studies on Florida’s mandatory PDMP query and pill mill laws suggested these policies reduced driver fatalities, with one study reporting approximately two fewer prescribed-opioid-related fatal crashes every month during the 22 months post-implementation.[21, 22] We measured opioidrelated fatal crashes, not prescription opioid-related fatal crashes, but did not find similar trends in our multi-state analysis. Florida did not meet inclusion criteria for our study as it implemented a voluntary PDMP law at the same time as its pill mill law, and an opioid prescribing cap law at the same time as its mandatory PDMP query law. Thus, it may have been the combined effects of these policies, which our study did not assess, that drove declines in opioid-involved fatal crashes observed in prior work.

Our findings suggest that state opioid prescribing laws are insufficient to help address rising rates of opioid-involved fatal crashes. Another set of laws that states have passed to target driving under the influence of drugs include DUID laws. Implementing these DUID laws is challenging due to the limitations of drug-detecting technology, the lack of an agreed-upon limit to determine impairment, poor officer training to identify drug impairment, and a lack of familiarity among judges and prosecutors on DUID laws as compared to DUI laws.[26] Further, the current evidence suggests it may never be possible to establish an impairment threshold for drugs where all (or most) drivers have their abilities impaired at concentrations above a certain limit;[42] while zero tolerance DUID laws do not need to establish a threshold for enforcement, these laws need to balance drivers that legally use prescribed opioids as they may have traceable amounts of opioids while under no impairment. As a result of these challenges, federal guidance recommends prioritizing four interrelated components: training on evidence-based drug screening, offering connections to drug treatment, using a period of intensive supervision, and utilization of combined DUI and Drug court procedures.[42] Federal guidance also recommend states test all fatally injured drivers, which would be a significant facilitator for future research involving analysis of drugged driving.

To support federal guidance to prioritize connections to drug treatment, expanded access to medications for opioid use disorder (MOUD) is needed. Treatment with FDA-approved MOUDs including buprenorphine, methadone, or naltrexone is the gold-standard of treatment for opioid use disorder.[43, 44] Recent policy changes have begun to expand access. For example, at the federal level, prescribers no longer need to obtain a waiver from the Drug Enforcement Agency (DEA) to prescribe buprenorphine for opioid use disorder; and the Substance Abuse and Mental Health Services Administration (SAMHSA) has made permanent a COVID-era rule expanding eligibility allowances for take-home doses of methadone.[45–47] At the state level, there are many ongoing policy initiatives – 82 laws with one or more MOUD expansion policy across 16 states in 2019 – aiming to expand access to MOUD via insurers, hospitals, and the criminal justice system, among others. [48] Future research is needed to understand the spillover effects of expanded MOUD access on rates of opioid-involved fatal crashes.

This study is subject to several limitations. First, there are limitations in using FARS data to identify drug-involved drivers due to incomplete state-level toxicology testing data.[24] Following previous studies, we limited our sample of opioid-involved fatal crashes to drivers who died within an hour of the crash which increased the testing rate nearly 30% across the entire sample. Additionally, we focused our results on reporting the rate of fatally injured drivers and included several sensitivity analyses using characteristics demonstrating higher risk of opioid use or opioid-involved fatal injury. Second, we included drivers who tested positive for opioids without being able to determine whether the opioids led to impairment of the driver; unlike alcohol-impaired driving, there is no consistent criteria for measuring drug impairment, so we are assuming that any presence of opioid presupposes impairment responsible for the crash. Studies have shown that prescription opioid use increases risk of fatal car crashes,[7, 8] so we assumed a positive test contributed to driver impairment. In order to isolate the effects of the opioid prescribing laws of interest on outcomes, we limited the study period to two years pre- and 1.5–2 years post- the laws of interest. While a longer time series is beneficial for establishing trends, in our case lengthening the study period introduced confounding from other types of state opioid prescribing laws. Lastly, we are unable to measure differences in implementation and enforcement across states, which may have differentially influenced the impacts of the law.

Across the 13 states in this study, there were small-in-magnitude and non-statistically significant changes in the rate of fatal crashes and opioid-involved fatal crashes attributable to the laws. Improved toxicology data collection on fatally injured drivers across states is urgently needed to support future research on strategies to address opioid-involved fatal crashes.