The analysis also compared cocaine-related mortality data reported to national vital statistics with data from all published forensic case series of cocaine-related deaths that occurred during the mid-1980's. Results revealed that 75 percent more cocaine-related deaths were reported to DAWN (n=6,057) than to national vital statistics (n=3,466) from the 25 metropolitan areas studied. For individual metropolitan areas, the discrepancy between DAWN and vital statistics counts of cocaine-related deaths was as large as a sixfold difference. In six of the seven forensic case series noted in the literature search, the number of cocaine-related deaths exceeded the number of these deaths reported to vital statistics. The largest discrepancy occurred during 10 months in 1986 in New York City, where 151 deaths were reported in a case series and seven deaths were reported to vital statistics. Findings indicated the need to improve existing public health data systems both to provide accurate measurements of the impacts of drug abuse and the evaluation of the effectiveness of drug prevention programs. Figure, tables, and 42 references (Author abstract modified)