The "healthy worker effect"--fact or artifact?

Workers in five chemical plants with diverse characteristics were studied by attempting to locate all white male employees who worked at least 12 months over a 25 to 30 year period. Former employees were contacted directly, rather than using questionnaires or Social Security Administration files. The data were analyzed by comparing observed mortality to expected mortality based on age-sex-race specific death rates in the U.S. population. The results indicate that the mortality experience of the chemical workers studied is essentially that of the general population. It is suggested that the "healthy worker effect" may be due to methods frequently employed in searching for death claims which understate the true experience of the employee group studied, or if present, does not affect mortality rates for neoplastic or cardiovascular disease.