We examined the trends in hospital discharge E-coding in Maryland over a 10-year period. The overall proportion of E-coded discharges has increased from 40 percent in 1979 to 55 percent in 1988. E-coding was lower in the severely injured, the elderly, and patients with long hospital stays. Our findings demonstrate that E-code reporting varies because of the limited number of data fields available for coding of discharge diagnoses. Universal, systematic reporting of E-codes in hospital discharge data is essential if these data are to provide critically needed information about nonfatal injuries. Hospital discharge data formats should contain separate fields for E-codes and the use of these codes, we believe, should be mandated.