Road safety data-driven monitoring: Possibility of performing regular data integration using pre-validated vital registration data, for more accurate statistics of fatal injuries in road traffic accidents

Abstract

Background: Thailand initiated the data integration process to get more accurate records of fatal injuries in road traffic accidents (RTA). Union of data from three different datasets: death records in Vital Registration, Road Accident Victims Protection Company (insurance), and Royal Thai Police are performed on an annual basis. However, even though it is more accurate than using a single database; it takes at least 6 months to validate the causes of death in the Vital Registration at the end of every year. In order to improve the timeliness of regular data-driven planning, this study aimed to find the possibility of using pre-validated Vital Registration records, in the data integration process.
Methods: The cross-sectional study used chi-square for the proportional test to compare the differences between two integrated datasets− one using the validated Vital Registration data, and another with pre-validated data.
Results: Overall data at country level showed no statistically significant differences in proportion in both datasets, in terms of gender, age group and month of accident (p-value 0.378, 0.934 and 0.997 respectively). At the provincial level, both datasets also showed no significant differences with p-value greater than 0.05. Annually, total number of death related to road traffic accident from integrated data sets using pre-validated and validated vital registration were 17,839 and 19,480 deaths, respectively. There was a 8.42 percent difference between both methods.
Conclusions: Utilization of pre-validated data as part of data integration is possible and useful to assess road traffic deaths. This finding suggest that reporting frequency of fatal injuries in RTA could be increased to aid in the planning of road safety interventions to reduce fatalities. However, existing annual data integration process should be continued, to get more precise figures.