Data Analysis

We performed statistical analyses using SAS 9.4 (SAS Institute, Cary, NC). First, we examined data from the screening survey to determine the prevalence of DAPO among the general adult ED population, and calculated descriptive statistics for this sample. Second, we examined DAPO among the subset of screening participants who received the extended study survey (ie, those who reported past three-month opioid use and twice-weekly driving). We limited the analysis of factors associated with DAPO to those taking the extended study survey because many of the measures of interest (e.g., high-risk driving behavior, substance use) were only measured in that subsample. We conducted bivariate comparisons between those who did and did not endorse DAPO among respondents to the extended study survey using t-tests for continuous variables and χ² tests for categorical data. Third, adjusted comparisons between participants with and without DAPO were modeled using logistic regression. We added variables to the logistic regression model sequentially, beginning first with demographics, and then substance use, depression, and chronic pain; and, finally, driving behaviors. The determination of predictor variables in the adjusted analysis was based on both theoretical considerations and parsimony (given the relatively small study sample).