||Analyzing Cohort Mortality Data.
Whittemore, A. S. ;
||Stanford Univ., CA. School of Medicine.;Health Effects Research Lab., Research Triangle Park, NC.;National Institutes of Health, Bethesda, MD.;National Science Foundation, Washington, DC.
Statistical analysis ;
Poisson density functions ;
Vector analysis ;
Occupational diseases ;
Regression analysis ;
Mathematical models ;
||Most EPA libraries have a fiche copy filed under the call number shown. Check with individual libraries about paper copy.
Several methods for analyzing data from mortality studies of occupationally or environmentally exposed cohorts are shown to be special cases of a single procedure. The procedure assumes a proportional hazards model for exposure effects and represents the log-likelihood kernel for the data as that of N independent Poisson variates, where N is the total number of person-units of mortality observation time in the study. It formalizes and justifies the epidemiological techniques of classifying deaths and person-months of study time into categories defined by exposure and other covariates, and of computing standardized mortality ratios and indirectly standardized death rates. Parameters representing exposure effects can be estimated by using standard software packages. Special cases and applications are described in the context of lung cancer mortality among U.S. uranium miners. (Copyright (c) 1985 American Statistical Association.)