Citation

Partyka SC. Proc. Am. Assoc. Automot. Med. Annu. Conf. 1980; 24: 156-169.

Copyright

(Copyright © 1980, Association for the Advancement of Automotive Medicine)

DOI

unavailable

PMID

unavailable

Abstract

The National Crash Severity Study (NCSS) investigated 12,050 traffic accidents in eight areas during the period January 1977 through March 1979. Detailed data on crash conditions and occupant injuries were collected and automated for statistical analysis. The measures of injury severity included the Occupant Injury Classification (OIC) and the Abbreviated Injury Scale (AIS) for up to six injuries. This paper uses the OIC and AIS codes to compute an Injury Severity Score (ISS) as developed by Susan Baker et al. Regression tools were used to model fatality rates as a function of each of the ISS and the AIS for NCSS data. The first year of data from the National Accident Sampling System (NASS), complete for 1979 with 3,417 cases, was used as a test of the models of fatality rate by injury measure. These results were then compared to the 1979 Fatal Accident Reporting System (FARS) census data. Fatalities predicted by each model, based upon the injury measures, were compared to the actual number of fatalities on the files. Both models did a good job of predicting NASS and FARS fatalities based upon the injury data.