Power considerations in epidemiologic studies of vinyl chloride workers.
Am J Epidemiol 1981; 114(5):725-734
Nine retrospective mortality studies of vinyl-chloride (75014) workers were statistically evaluated. The statistical powers of mortality studies were calculated with respect to the liver, brain, and lung sites and plotted on power curves to relate positive and negative findings to power, and to show the variability in powers. The results of the liver and brain cancer studies were consistent with an etiologic role for vinyl-chloride. In the case of lung cancer, two studies with very high power or excess risk yielded negative results indicating that vinyl-chloride was not a lung carcinogen. The authors conclude that power considerations are a useful epidemiologic tool.
Plastics; Chemical-manufacturing-industry; Chemical-industry-workers; Epidemiology; Health-surveys; Mortality-rates
American Journal of Epidemiology