Challenges in studying statewide pedestrian injuries and drug involvement.

Background: Increasing U.S. rates of pedestrian injuries could be attributable in part to changing policies and attitudes towards drugs and associated increases in use, yet drug use has not been investigated widely as a risk factor for pedestrian injury. This study details challenges to investigating drug-involved pedestrian crashes using existing surveillance systems.

Methods: Using California police reports from 2004 to 2016, we performed simple linear regression with the proportion of data that was missing by year for drug and alcohol use as the outcome of interest. We also explored differences in the relative proportion of missing data across sex, race, and age groups through simple logistic regression. Finally, we compared missing data for alcohol and drug use indicators for pedestrians and drivers.

Results: From 2004 to 2016, 182,278 pedestrians were involved in crashes across California. Only 1.22% (n = 2219) of records indicated drug use, and 98% had missing data for drug use; the proportion of missing data did not change over time (b = - 0.040, p = 0.145, 95% CI = (- 0.095, 0.016)). The proportion of missing values for alcohol use increased each year (b = 0.49, 95% CI = (0.26, 0.72), p = 0.001). Driver drug and alcohol use indictors showed similar data missingness, and missing data did not show significant variation over time. Hispanics were more likely to have missing data for drug use compared to Whites (OR = 0.61, p < 0.001, 95% CI = (0.56, 0.67)), and Blacks were more likely to have missing data for alcohol use compared to Whites (OR = 0.87, p < 0.0001, 95% CI = (0.84, 0.91)).

Conclusions: Drug use may be a key contributing factor to pedestrian injury, but drug use remains consistently and largely unmeasured in existing surveillance systems. Without better collection of drug and alcohol data, monitoring trends in drug-involved pedestrian injury will not be feasible.