An Epidemiological Assessment of the Risk of Fatal Motor Vehicle Collisions with Benzodiazepine Use in South Korea

Abstract

Background: Motor vehicle collisions (MVCs) represent a major public health problem that has been associated with the use of benzodiazepines (BZDs) in several studies from Europe, North America, and Australia. Further epidemiological assessment is necessary to establish a similar association in Asian countries, which have different healthcare systems and patterns of medication use. In particular, South Korea has one of the highest rates of fatal MVCs and BZD prescriptions. The present epidemiological study aims to assess the risk of fatal MVCs among drivers with BZD exposure using a population-specific approach in South Korea.

Methods: The Korean Traffic Accident Analysis System database was linked to the National Health Insurance database to obtain medical information for all individuals who experienced fatal MVCs between 2010 and 2014 in South Korea. First, we evaluated the prevalence of BZDs by calculating the odds ratios of BZD exposure among the MVC cohort compared to a random 2% sample of the entire South Korean population. To control for age, we also calculated the standardized prevalence rate ratio for BZD exposure. Second, we analyzed the responsibility of the drivers in causing the fatal MVC by assigning responsibility scores for various characteristics relevant to the traffic accident. These responsibility scores were subsequently compared between drivers with BZD exposure and those without BZD exposure. Finally, we used a case-crossover design to observe the transient effects of BZD exposure on the risk of fatal MVCs by comparing BZD exposure immediately before the onset of the fatal MVC (“hazard period”) to BZD exposure at four earlier time periods (“control periods”).

Results: Prevalence of BZD exposure within the past year was significantly higher among the MVC cohort compared to the general population (OR: 2.66, 95% CI: 2.54- 2.79). With age-adjusted rates, the standardized prevalence rate ratio for BZD exposure among the MVC cohort was 1.60 (95% CI: 1.54-1.66). In the responsibility analysis, significantly more drivers with BZD exposure were classified as culpable, compared to drivers without BZD exposure (p<0.0001). In the case-crossover study, BZD exposure in a hazard period of one day before the MVC significantly increased the risk of fatal MVCs, even after adjusting for concurrent medications in the conditional logistic regression model (Adj. OR: 2.20, 95% CI: 1.88-2.58). Sensitivity analyses on various combinations of hazard and washout periods strengthened the robustness of this result.

Conclusions: The present study has established a relationship between high rates of fatal MVCs and BZD prescriptions in South Korea through three comprehensive statistical analyses. With fatal MVCs serving as one of the leading causes of death globally, the results accentuate the need for clinical and traffic policies that protect drivers’ safety and address inappropriate BZD prescriptions. This study also motivates further investigation into the risk of medication use across all types of MVCs, especially in countries that demonstrate growing trends on the use of motor vehicles and prescription drugs.