The authors examine the effectiveness of using hospital discharge data in assessing trends and geographic variations in the occurrence of selected chronic diseases. The Chronic Disease Surveillance System, in place from 1987 to 1991, used hospital discharge data, and Cancer Registry data to tract selected chronic diseases. The authors reviewed data on three diseases: breast cancer, cervical cancer, and lung cancer. A computerized algorithm was used to link multiple records representing a single disease occurrence. To estimate disease occurrence rates from hospital discharge data, repeat admissions for the same disease in any given calendar year were discounted. All rates were directly age-adjusted to the 1985 Maine state population. For all three diseases, the rates obtained from hospital discharge data were higher than Cancer Registry rates. Possible causes for the discrepancies and suggestions for improving the utility of hospital discharge data for chronic disease surveillance are discussed.