The availability of complete data on traffic injury cases including both the crash circumstances as well as health outcomes is a key tool to tailor appropriate motor vehicle injury prevention policies at local and regional levels. This paper explores the feasibility of the probabilistic linkage of police and hospital emergency department (ED) data regarding motor vehicle injury cases in Barcelona (Spain), including the use of methodologies to optimize the linkage process. A probabilistic linkage process was performed using local police reports (n=12,481) and hospital ED data (n=16,733) from all traffic injury cases occurring in 1997. Common variables in both files used were: age, sex, position, vehicle, hospital, date of crash, and postal code of place of residence. In all linked pairs at least one variable coincided, in almost two of every three pairs at least five of the seven variables coincided, and in 7.3% of the cases all variables coincided fully. This work has allowed the empirical development of combined methods for the linkage of police and health care sources of information at the local level, as preliminary steps towards the construction of a data base including information on the circumstances and the consequences of motor vehicle crashes. (A)
Abstract