Statistical analysis of epidemiological data from a chromium chemical manufacturing plant.

A statistical analysis was performed on reported bronchogenic carcinoma incidence in a group of chromium chemical manufacturing workers over the period 1929 to 1977. The purpose was to determine if post-1950 process improvements corresponded to favorable trends in morbidity. Because the paucity of data on the population at risk prevented meaningful calculations of SMR's, a recently developed method called "probability window analysis" was used to evaluate for statistically valid changes in the "numerators" or number of reported cases. A significant downward trend in reported bronchogenic carcinoma was found which coincided with major process improvements made in 1951 and 1961. Although adjustment for all competing risk factors was not possible because of insufficient data on the population at risk, the improved working environment has to be considered a major contributing factor to the favorable trend.