RESUMEN
As of January 31, 1996, 292 deaths among registered patients of Yusho were identified by three follow-up studies conducted in 1986, 1990, and 1996. In this study, we attempted to identify underlying causes of death by linkage of the registered data to the National Vital Statistics Data provided by the Management and Coordination Agency of Japan, which included 15 million deaths between 1978 and 1996. The two datasets were linked by matching for six variables; birth year/month/day, death year/month, and sex, along with a variable of death day or death place, or both. The matched cases were 203 among 235 deaths between 1978 and 1996 (matching rate was 86%). Among the 203 deaths, 58 underlying causes of death were newly identified, 146 causes of death were already grasped by the follow-up studies, and 31 deaths did not have matching pair in the National Vital Statistics data. Among the 146 deaths, 110 causes of death were concordant with each other, however, 35 causes of death were completely discord. The reason of the discordance and the unmatched deaths might be due to difference in information of the matching variables in the two datasets. In order to conduct an efficient follow-up study of Yusho patients, identification of underlying causes of death by linkage to the National Vital Statistics Date is evitable. For that, we need to substitute basic information in the Yusho database to those compatible to the National civil registration system.