Although 80–90% of lung cancer deaths occur among smokers, many nonsmokers die from lung cancer. Insights into lung cancer mortality among nonsmokers are critical because indirect approaches for estimating smoking-attributable mortality use the difference between observed lung cancer death rates and expected lung cancer death rates among nonsmokers to measure smoking burden. Our study compares lung cancer mortality among nonsmokers from the nationally-representative 1985–2015 National Health Interview Survey Linked Mortality Files (NHIS-LMF) with the non-representative Cancer Prevention Study II (CPS-II). We find that lung cancer death rates among nonsmokers are higher in the NHIS-LMF than in the CPS-II. Second, we find that smoking-attributable fractions, based on the Preston-Glei-Wilmoth indirect method, are slightly lower with NHIS-LMF rates than with CPS-II rates. Despite much lower lung cancer mortality rates among nonsmokers in CPS-II than NHIS-LMF, smoking-attributable fractions based on non-representative CPS-II data do not appear to be biased.
Presented in Session 5. Health & Mortality 1