Funding: Insurance Institute for Highway Safety
A strong association has been established between alcohol use and motor vehicle collisions (MVCs), primarily based on emergency department (ED) studies.1 Less is known about cannabis use and MVC or the interaction of cannabis with alcohol and other drugs as a risk factor for crash. The principal psychoactive component of cannabis, delta-9-tetra-hydrocannabinol (THC), has been found to significantly impair attention, reaction time, hand-eye coordination, decision making and concentration, as well as driving performance.2 The risk for MVC with cannabis, as the most commonly used drug worldwide, and cannabis with alcohol, as the most common co-ingestion, is of particular concern in the current legal environment: 25 states and the District of Columbia (DC) have enacted legislation to decriminalize marijuana for medical use, four of those states and DC have legalized marijuana for recreational use, another four states are likely to do so by the end of the year, and initiatives are pending in others.3 The small amount of available literature suggests that cannabis increases crash risk and that cannabis and alcohol have a potentially more-than-additive effect on driving.
Most of the existing literature, however, examines either MVCs in which no injury occurred, or only fatal crashes,5 in which detailed information about quantity and context of cannabis use cannot be captured. Between 2004 and 2011, the Drug Abuse Warning Network (DAWN) data found a 100% increase in the number of ED visits related to cannabis use alone and a 62% increase in visits for cannabis use in combination with other drugs.6 Data from the U.S. Healthcare Cost and Utilization Project (HCUP) found that cannabis- related ED visits increased by 68% from 2007 to 2012 (and alcohol-related ED visits by 49%).7 Overall, there has been a steady growth in cannabis-related healthcare utilization suggesting consequences of use are worse now than in the past.8 Therefore, this proposal aims to:
1) Examine the prevalence of cannabis use, patterns and context of use among adult drivers presenting after MVC to the ED of Oregon Health & Science University (OHSU) in Portland, OR.
Hypotheses: 1a) The prevalence of cannabis alone and combined with alcohol will be high in this population and will increase over the course of the study, as cannabis markets mature and normalization increases after legalization for recreational use; 1b) Context of use will reflect increasing normalization of cannabis use (e.g., during commercial activities) and allow us to characterize risk across a wide variety of activities.
2) Using case-control and case-crossover designs, estimate the relative risk (RR) of MVC for cannabis use alone an in combination with alcohol, after controlling for context of injury and other drug use.
Hypothesis: 2a) Risk of MVC will be greater with cannabis use prior to driving compared to no cannabis use, and greater with combined use of cannabis and alcohol compared to cannabis use alone.
3) Analyze the dose-response relationship of cannabis and alcohol use for risk of MVC, quantifying the amount of use through a) serum ∆-9-THC and ethanol levels on arrival in the ED, and b) interviews regarding use of cannabis (e.g., type, mode of delivery, relative potency, estimated amount) and alcohol, both usual use and use within 6 hours before the crash, exploring the interaction of usual use on the dose-response relationship.
Hypotheses: 3a): A dose-response relationship for cannabis use will be observed among MVC patients. The dose-response interaction between alcohol and cannabis will be more pronounced than for either of the two substances alone, and will be more than additive; 3b) The dose-response relationship will be modified by usual use of cannabis and alcohol (e.g., frequency of high consumption occasions).
By recruiting participants from the ED, we focus on a population involved in clinically significant MVC’s likely to have caused injury, in quantities and in a setting that allows us to obtain detailed information about drug use, presenting a unique opportunity to examine the relationship between cannabis use and crash risk. While using a population control is likely not feasible for this research question, particularly given the requirement for obtaining biologic samples, using both case-control and case-crossover approaches together will allow us to evaluate for potential biases of either approach.