Dynamic models for the maintenance of smoking cessation: event history analysis of late relapse.

This paper presents event history analysis as an approach to examining the dynamic nature of the maintenance of smoking cessation. Data from a 1-year follow-up of 172 adult male and 209 adult female ex-smokers is used to estimate the rate at which individuals relapse and return to abstinence. Results indicate that the rate of relapse in both males and females (3.9 and 3.6% per month, respectively) is roughly half that of the rate of return to abstinence (7.5 and 6.3% per month, respectively). No evidence was found for a "safe point" during the observation interval. Individual characteristics that affect the rate at which ex-smokers relapse and return to abstinence are identified. Characteristics that influenced the rate of return to abstinence were completely different from those affecting the rate of relapse, a finding that suggests covariate asymmetry. Implications of dynamic analysis for conceptualizations of maintenance are discussed.