New dimensions in cause of death statistics.

Abstract

This article discusses the limitations of traditional national cause of death statistics. These limitations derive from an attempt to conceptualize a multidimensional phenomenon and reduce down to a unidimensional framework. The article outlines the characteristics of a new multidimensional approach which involves the codification and tabulation of all causes (multiple causes) listed on death certificates. Preliminary data are presented which illustrate that multiple cause of death data do indeed represent a major new dimension to cause of death statistics. These data indicate that most major causes of death are contributory factors in many deaths in which they are not the underlying cause of death. For example, in 1976, diabetes mellitus was the underlying cause of approximately 35,000 deaths but was a contributing factor in another 100,000 deaths. This paper also demonstrates the contribution of multiple cause of death data to identifying patterns of association among diseases and the kinds of injuries resulting from various external causes. Finally, data are presented which depict the use of multiple cause data in evaluating the efficacy of the coding rules used to classify the underlying cause of death.