Spousal smoking as an indicator of total secondhand smoke exposure

INTRODUCTION: This study assessed the utility of spousal smoking as a measure of secondhand smoke exposure. Methods The investigation involved secondary analysis of data from 526 female participants of a lung cancer case-control study from northeastern England. Secondhand smoke exposure was measured in the home (spousal and nonspousal), workplace, and social/other settings over the whole life course. RESULTS: Almost all women (99.1%) had at least 10 years of secondhand smoke exposure from at least one source, most commonly from parental smoking in childhood, and spousal smoking, the workplace, and social settings during adulthood. Spousal smoking was strongly correlated with overall secondhand smoke exposure in the home over the life course but was weakly correlated (Kendall's tau = -.04 to .12) with secondhand smoke exposure from other domestic sources and with secondhand smoke exposure in the workplace or social/other settings. Most women who gave no history of spousal secondhand smoke exposure recalled at least 10 years of secondhand smoke exposure in other settings: in the home through other sources (83.2% > or = 10 "smoker-years"), through workplaces (63.4% > or =10 "exposure-years"), or in social settings (82.0% > or =10 exposure-years). Almost all (96.9%) reported at least 10 years of exposure from at least one of these nonspousal sources. DISCUSSION: Using spousal smoking as a proxy of total secondhand smoke exposure would have meant that these subjects would have been misclassified as not exposed to secondhand smoke. This misclassification may bias estimates of association with health outcomes toward the null. Studies of the effects of secondhand smoke exposure on health outcomes should evaluate all potential sources of secondhand smoke exposure.