The reliability of death certificate reporting for amyotrophic lateral sclerosis.

INTRODUCTION THIS STUDY was designed to determine the usefulness of mortality statistics in estimating the prevalence of amyotrophic lateral sclerosis (ALS) by analysing the death certificates of a series of patients in whom this diagnosis had been made during life. Most of the descriptive epidemiologic information of ALS has been based primarily on the analysis of death certificate data [1, 21. Analyses based on other sources of data have obvious limitations. Hospital series are biased by selection of patients except in areas like Rochester, Minnesota, where there was only one source of medical care in the community [3]. Population surveys except in Guam and the Kii Peninsula of Japan, where there are known concentrations of ALS, are impractical because of the low prevalence of the disease. Credence has been given to death certificate data because, except for Guam and the Kii Peninsula, mortality statistics are fairly uniform throughout the world [4,5]. Indeed, Kurtzke [6] has suggested that the uniformity of ALS frequency may provide a standard reference for comparison of the prevalence of other neurologic diseases. ALS present such a striking clinical picture late in its course, that it is likely that even physicians who have seen few cases of the disease would be able to recognize it.