BACKGROUND AND PURPOSE: The age-adjusted mortality rate on the island of Quemoy was the lowest of any county in Taiwan in the years 2000-2001. However, the island's rate of mortality due to cardiovascular diseases fluctuates widely. This fluctuation suggests that Quemoy may be issuing inaccurate death certificates. METHODS: To assess the quality of death certificates from Quemoy, 123 (15.3%) medical records for 800 deaths between 1994 and 1998 were reviewed by 3 medical specialists. The underlying cause of death from the original death certificate was compared to the underlying cause of death determined by 3 reviewers based on review of all available medical records. RESULTS: The agreement index for all causes of death was 72.4%. Neoplasms had the highest sensitivity and positive predictive value for correct determination of underlying cause of death. Cardiovascular diseases had higher sensitivity, but lower positive predictive value than respiratory diseases. Neoplasms were under-reported by 25.5% but cardiovascular diseases were over-reported by 34.3%. CONCLUSIONS: Therefore, mortality statistics in Quemoy, which are based on death certificate data, may underestimate the frequency of neoplasms and overestimate cardiovascular diseases as underlying causes of death. Our findings also suggest that researchers should exercise considerable caution when using death certificate data to determine cause of death in etiologic studies, especially in neoplasms and cardiovascular diseases.
BACKGROUND AND PURPOSE: The age-adjusted mortality rate on the island of Quemoy was the lowest of any county in Taiwan in the years 2000-2001. However, the island's rate of mortality due to cardiovascular diseases fluctuates widely. This fluctuation suggests that Quemoy may be issuing inaccurate death certificates. METHODS: To assess the quality of death certificates from Quemoy, 123 (15.3%) medical records for 800 deaths between 1994 and 1998 were reviewed by 3 medical specialists. The underlying cause of death from the original death certificate was compared to the underlying cause of death determined by 3 reviewers based on review of all available medical records. RESULTS: The agreement index for all causes of death was 72.4%. Neoplasms had the highest sensitivity and positive predictive value for correct determination of underlying cause of death. Cardiovascular diseases had higher sensitivity, but lower positive predictive value than respiratory diseases. Neoplasms were under-reported by 25.5% but cardiovascular diseases were over-reported by 34.3%. CONCLUSIONS: Therefore, mortality statistics in Quemoy, which are based on death certificate data, may underestimate the frequency of neoplasms and overestimate cardiovascular diseases as underlying causes of death. Our findings also suggest that researchers should exercise considerable caution when using death certificate data to determine cause of death in etiologic studies, especially in neoplasms and cardiovascular diseases.