Background: Previous Canadian estimates of hospital use by smoking history have been derived by applying disease-specific "smoking-attributable fractions" to administrative data. For this analysis, health survey data were linked to hospitalization data at an individual level, permitting prospective measures of hospital use by smoking status and age.
Data and methods: Data for 28,255 respondents (outside Quebec) to the 2000/2001 Canadian Community Health Survey (CCHS) were linked to the Hospital Person-Oriented Information Database. Days in hospital over four years were quantified for each respondent and examined in relation to smoking status in 2000/2001. Multiple logistic regression was used to examine the association between smoking and hospitalization, while controlling for confounders.
Results: During the four years after their CCHS interview, current daily smokers and former daily smokers who had quit in the past five years averaged more than twice as many days in hospital as did never-daily smokers. Altogether, excess hospital days for current and former smokers aged 45 to 74 numbered 7.1 million over four years, and accounted for 32% of all hospital days used by people in this age group.