Predicting smoking cessation and its relapse in HIV-infected patients: the Swiss HIV Cohort Study. HIV Medicine
Schäfer et al. assessed whether information routinely collected at 6-monthly follow-up visits within the SHCS could be used to predict (i) the smokers most likely to stop and (ii) among smokers who stop, those most likely to relapse. The authors built simple prediction models, including clinical and statistical models with patients classified into three motivational groups. Of 4’833 smokers, 1’261 (26%) stopped smoking. Among those who stopped, 48% started smoking again. More patients who stopped were in the highly motivated group. The predictive performance of the clinical and statistical models for smoking cessation and relapse was modest.
The study results suggest that smoking cessation and its relapse cannot be accurately predicted from variables typically collected in observational cohorts of HIV-infected patients. Targeting those trying to stop or those who have just stopped are the best candidates for an intervention.