MORTALITY AND NUTRITIONAL STATUS--DEMOGRAPHIC ANALYSIS

点击翻译按钮获取中文摘要

The majority of deaths, prior to the last half century in the developed nations and in most third world countries today, result from infectious disease. Currently in the developed nations, these infectious diseases have been replaced by degenerative diseases as the most frequent causes of death. One hypothesis concerning the cause of this epidemiologic transition is that the infectious diseases decline and are replaced by degenerative diseases as a result of characteristic changes in nutrition. The overall goal of this project is to quantitatively investigate the effects of worldwide differences in nutritional status, as measured by food consumption and anthropometric measurements, on three demographic variables: the level of mortality, the age pattern of mortality, and the cause of death structure. First, the association of worldwide variation in nutritional status with the variation in level and age pattern of mortality will be explored, using a worldwide sample of life tables (post 1940) matched with information on nutritional status. The age pattern of mortality will be assessed using a five-parameter competing hazards model, and a comparison of the measures of mortality and nutritional status will be conducted using multiple regression analysis. Discriminant functions will be constructed to determine if characteristic nutritional patterns are associated with characteristic mortality patterns. Second, the correlations between nutritional status and the level of age pattern of 12 causes of death will be determined, using a sample of worldwide full and cause decremented life tables. The methods are similar to those described above. Third, the results of the first two aims will be used to define structural equation (causal) models of the relationship among nutrition, mortality, public health, educational, economic and social variables. These analyses will use the worldwide sample of life tables matched with nutritional status variables and a third data base consisting of environmental and social variables. The models will be fitted using standard maximum likelihood methods and the program LISREL. The resulting models will be validated by comparing them with published results of nutritional experiments, in Guatemala and India.