Do predictors of smoking relapse change as a function of duration of abstinence? Findings from the United States, Canada, United Kingdom and Australia

AIMS: To estimate predictors of time to smoking relapse and test if prediction varied by quit duration. DESIGN: Longitudinal cohort data from the International Tobacco Control Four-Country survey with annual follow up collected between 2002 and 2015. SETTING: Canada, United States, United Kingdom and Australia. PARTICIPANTS: A total of 9171 eligible adult smokers who had made at least one quit attempt during the study period. MEASUREMENTS: Time to relapse was the main outcome. Predictor variables included pre-quit baseline measures of nicotine dependence, smoking and quitting-related motivations, quitting capacity and social influence, and also two post-quit measures, use of stop-smoking medications and quit duration (1-7 days, 8-14 days, 15-31 days, 1-3 months, 3-6 months, 6-12 months, 1-2 years and 2+ years), along with socio-demographics. FINDINGS: All factors were predictive of relapse within the first 6 months of quitting but only wanting to quit, quit intentions and number of friends who smoke were still predictive of relapse in the 6-12-month period of quitting [hazard ratios (HR) = 1.20, P < 0.05; 1.13, P < 0.05; and 1.21, P < 0.001, respectively]. Number of friends smoking was the only remaining predictor of relapse in the 1-2 years quit period (HR = 1.19, P = 0.001) with none predictive beyond the 2-year quit period. Use of stop-smoking medications during quit attempts was related negatively to relapse during the first 2 weeks of quitting (HR = 0.71-0.84), but related positively to relapse in the 1-6-month quit period (HR = 1.29-1.54). Predictive effects of all factors showed significant interaction with quit duration except for perceiving smoking as an important part of life, prematurely stubbing out a cigarette and wanting to quit. CONCLUSIONS: Among adult smokers in the United States, Canada, United Kingdom and Australia, factors associated with smoking relapse differ between the early and later stages of a quit attempt, suggesting that the determinants of relapse change as a function of abstinence duration