Use of "time windows" to investigate lung cancer latency intervals at an Ontario steel plant.

This paper describes an application of the time windows method to an examination of the temporal pattern of lung cancer risk among steel workers. Case-control methodology was utilized. The cases were 36 men who had died of lung cancer and the controls were 289 men who had died of any other cause. The number of years of employment in the steel pouring area was used as a surrogate measure of exposure. The data were examined by contingency table analysis and by logistic regression, which permitted adjustment for exposures in multiple time windows and the use of continuous, rather than categorical, measures of exposure. It was found that lung cancer risk was associated with exposures occurring between 18 and 30 years before death. It cannot yet be determined whether this time course reflects a biological response or the temporal pattern of exposure to an as yet unidentified toxic agent.