Mortality Studies in an Interdisciplinary Program of Occupational Disease Research.
NIOSH 1977 Aug:0 pages
Investigations made by the mortality studies program of the Division of Respiratory Diseases, NIOSH are reviewed to detect, measure, and reduce nonsampling errors in the occupational mortality studies conducted by the division. Three kinds of nonsampling errors in the standard mortality ratio (SMR) found in an 8 year mortality followup on two Appalachian probability samples of bituminous coal miners were investigated: 1) errors arising from selection of standard populations not matching the study group in time and geographic variables; 2) errors arising from confounding of effects of nonoccupational variables correlated with variation in death rates due to occupational exposures; and 3) inadequacies of death certificate disease reports. Results of the study include: 1) a modified definition of the standard population needed for computing expected deaths; 2) variations in the SMRs arising from varying the relative times and locations of study group and standard populations were found to be small compared to variations in SMRs arising from the confounding of selected mortality correlated variables; and 3) large discrepancies between autopsy results and information on death certificates.
NIOSH-Author; Sampling-methods; Biostatistics; Mortality-rates; Analytical-methods; Diseases;
Epidemiological Investigations Branch, Division of Respiratory Disease Studies, NIOSH, Morgantown, We 9 references