The Epidemiologic Necropsy

To the Editor.— McFarlane et al,1in the July 17, 1987,JAMAarticle entitled "The `Epidemiologic Necropsy,' " address the question of the extent of the hypothetical "undetected reservoir" of lung cancer by a retrospective analysis of the autopsy findings at a Connecticut hospital for a recent ten-year period (1972 to 1981) and compare their results with the incidence rates derived from cases reported to the Connecticut Tumor Registry. The authors compute age- and sex-standardized occurrence rates that are greater by a factor of 10 (3.95 men and 14.86 women) than the rates from the Connecticut Tumor Registry. However, in computing the two sets of rates, the authors make two errors. First, there seems to be a computational inaccuracy. The authors, in obtaining the crude Connecticut Tumor Registry rates, divide by the 1970 to 1979 population of Connecticut aged 20 years and older but base the age-adjusted rates on the