Spillover Effects of Opioid Prescriptions: Do Increased Prescriptions Lead to Increased Fatal Car Crashes?

Drug overdoses are now the leading killer of Americans under age 55. Widespread prescriptions for opioids may have additional negative public health consequences, the effects of which have not yet been accounted for in measuring the costs of the crisis. Car crash fatalities may be one such consequence. Although there is evidence of increasing use of prescription opioids among drivers who die in fatal car crashes, it is not yet clear whether opioid prescribing practices have caused increased traffic fatalities. Increased prevalence of opioids among fatally injured drivers may simply reflect the fact that many Americans take prescription opioids: over 17 percent of Americans filled a prescription for an opioid in 2017 (CDC 2017). Thus, higher opioid use nation-wide could lead mechanically to higher use rates among deceased drivers. This may be especially true if underlying demographic characteristics of prescription opioid users are also associated with involvement in fatal crashes.

In this paper, we address this issues by estimating the relationship between the prevalence of local opioid prescriptions and local traffic fatality outcomes. Using data on all fatal car accidents in the US between 2007 and 2016, we estimate fixed effects models that relate changes in commuting zone (CZ)-level counts of fatal crashes, deaths and driver use of alcohol and opioids to changes in CZ-level per-capita prescriptions for opioids. We estimate that a one standard deviation increase in local opioid prescriptions per 100 people leads to a 30 percent increase in the number of drivers involved in fatal crashes who test positive for opioids. While we also find that more prevalent opioid prescriptions are linked to more drug testing, we find a larger proportional increase in the likelihood of testing positive for opioids than in the likelihood of being drug tested at all. This implies that the large increase in positive opioid tests is not simply driven by more prevalent testing of drivers. The effect of local prescribing practices on drugged driving translates into additional traffic fatalities. We estimate that a one standard deviation increase in the average local opioid prescription rate is associated with 1 to 2 additional annual fatal traffic accidents per CZ. We also do not uncover any effect of local prescriptions on alcohol outcomes. There is no effect of additional local opioid prescriptions on the number of drivers tested for alcohol, or the number of drivers with alcohol in their system. This runs counter to existing evidence that suggests that opioids may act as a substitute for alcohol. At least among those who drive while intoxicated and are eventually involved in a fatal crash, additional opioids prescriptions are not associated with less alcohol use.