Re: An alternate characterization of hazard in occupational epidemiology: years of life lost per years worked. Am J Ind Med 42:1-10, 2002.
Park-RM; Bailer-AJ; Stayner-LT; Halperin-W; Gilbert-SJ; Smith-RJ; Bena-JF
Am J Ind Med 2003 Mar; 43(3):334
We thank Dr. Morfeld for his interest and appreciate the careful and detailed effort that he has undertaken in analyzing our recent study on work-related years of potential life lost [Park et al., 2002]. Dr. Morfeld is concerned that our approach, while ''attractive in its simplicity,'' may suffer from several sources of bias. However, upon examination, his specific example appears to us to be inappropriate. A key element in Dr. Morfeld's argument for two of the three sources of bias is a hypothetical population in which deaths resulting from an exposure are arbitrarily postulated to have been moved forward in time by 5 years. This stipulation was not derived from empirical evidence or a failure-time model, and ignores the reference population's survival characteristics as summarized in a relevant lifetable. Alternate choices for the acceleration interval, such as 10, 20, or 30 years, would produce very different estimates of years of life lost. Therefore, a meaningful comparison of years of life lost cannot be made contrasting our approach, which assigned life-expectancy to excess deaths, and Morfeld's approach, which arbitrarily assumes a five year acceleration of death. On the third source of bias identified by Morfeld, related to censoring, we stated that our specific findings were not generalizable to other populations including, for example, this population followed over a different time period.
Epidemiology; Occupational-hazards; Statistical-analysis; Lifespan; Risk-analysis; Mortality-data; Age-groups; Survival-rate
Robert M. Park, National Institute for Occupational Safety and Health, Education and Information Division, Risk Evaluation Branch, Mail Stop C-15, 4676 Columbia Parkway, Cincinnati, OH 45226
American Journal of Industrial Medicine