Catalyzing Young Driver Research via Data Linkage: Development of a Comprehensive Traffic Safety Data Warehouse

项目摘要

Project Abstract High-quality linkage of motor vehicle crash data to other existing data sources has the potential to catalyze advancements in our understanding of crashes—the leading cause of death among US adolescents. However, the field of young driver safety relies heavily on studies that analyze only publically available de-identified crash data. Analyzing crash data in isolation restricts the study period to essentially a few minutes and has important limitations that hinder the ability to utilize crash data to validly conduct critical research. Thus, the long-term goal of this R21 project is to harness the power of linked crash data to enhance knowledge of the contributors to and burden of young driver crashes. Over the last five years, the project team has developed a unique linked database—the New Jersey Traffic Safety Outcomes (NJ-TSO) database—that contains the full driver licensing, police-reported crash, and traffic citation data of all NJ drivers (10 million) from 2004 through 2014. The overall objective of the proposed project is to extend this program of research by integrating three additional statewide sources of data in NJ: birth certificate, hospital discharge, and death certificate data. The resultant NJ-TSO data warehouse, which will include individual-level longitudinal data that span the continuum from underlying contributing factors and relevant previous events to injury outcomes and death, is innovative as it will be the most comprehensive data resource focused on young drivers to date. The project has three specific aims. First, the NJ-TSO data warehouse will be created by fully integrating the three additional statewide data sources using established probabilistic linkage methods. Second, key analytic data elements will be derived for all licensed drivers and all crash-involved drivers. These will include: (1) geocoded residential addresses in order to increase the utility of the warehouse to support future spatial and neighborhood-level analyses; (2) race and ethnicity— data elements that are not included in NJ licensing or crash data but are critical to the conduct of foundational and policy-relevant young driver research—using a combination of the newly-integrated data sources and Bayesian Improved Surname Geocoding methods; and (3) crash injuries/fatalities using a combination of data sources. Third, the project will make a unique methodological contribution by quantifying the magnitude and direction of bias introduced by using crash report data (vs. hospital/death) to ascertain crash-related injuries. The proposed research is significant in that the resultant warehouse will be fully primed to directly address critical, high-priority research questions in young driver safety. As the NJ-TSO warehouse will include drivers of all ages, it will also have the potential to serve as a data resource for collaborations on a wide array of high-priority areas of traffic safety, including: impaired driving; unlicensed driver crashes; older driver (> age 65) crashes; pedestrian and pedalcyclist injuries; and child passenger safety. This will ensure that the project optimizes the ability of the NJ-TSO data warehouse to improve public health and the safety of all road users.