Multivariate models for predicting abstention following intervention to stop smoking by general practitioners

Abstract

Predictors of successful smoking cessation were examined in a randomized controlled trial of 450 smokers who received an intervention by their general practitioner (GP). Pretreatment characteristics predicting outcome at 3, 6 and 12 months and for continuous abstinence to 12 months were determined using logistic regression analyses. Results showed the variables that significantly predicted abstention at 3 months were age and motivation, whereas the predictors at 6 months were socio-economic status, motivation, level of dependence and time spent with smokers. No single predictor emerged at 12 months. Predictors for continuous abstinence to 12 months were age, time spent with smokers and motivation. A model was developed which best describes the likelihood of patients achieving continuous long-term abstinence. The results show that five factors (high motivation level, older age, less time spent with smokers, low dependence level, and higher socio-economic status) together have a 76% accuracy of predicting continuous abstinence to 12 months.