Comparison of physician based reporting of tobacco attributable deaths and computer derived estimates of smoking attributable deaths, Oregon, 1989 to 1996.

BACKGROUND Tobacco use prevention programmes need accurate information about smoking related mortality. Beginning in 1989, Oregon began asking physicians to report on death certificates whether tobacco use contributed to the death. OBJECTIVE To determine the long term comparability of this method of estimating tobacco attributable mortality to estimates of smoking attributable mortality derived from a computer model. DESIGN For the period 1989 to 1996, we compared mortality resulting from tobacco use reported by Oregon physicians to estimates of smoking attributable deaths (SADs) derived by "Smoking attributable mortality, morbidity and economic costs" software version 3.0 (SAMMEC 3.0), a widely used software program that estimates SADs on the basis of smoking prevalence and relative risks of specific diseases among current and former smokers. MAIN OUTCOME MEASURES Numbers of deaths, age, sex, and category of disease. RESULTS Of 212, 448 Oregon deaths during 1989-1996, SAMMEC 3.0 estimated that 42, 778 (20.1%) were attributable to cigarette smoking. For the same 27 diagnoses, physicians reported that tobacco contributed to 42, 839 (20.2%) deaths-a cumulative difference of only 61 deaths over the eight year period. The age and sex distributions of tobacco and smoking attributable deaths reported by the two systems were also similar. By category of disease, the ratio of SAMMEC 3.0 estimates to physician reported deaths was 1.11 for neoplasms, 0.88 for heart disease, and 1.04 for respiratory disease. CONCLUSIONS Physician reporting provides comparable estimates of smoking attributable mortality and can be a valuable source of data for communicating the risks of tobacco use to the public.