Cognitive-Behavioral Therapy and Hypnotherapy for Smoking Cessation (CBT-HT)

Intensive cognitive-behaviour interventions (CBT) combined with pharmacotherapy for smoking cessation are well established and have been proved to be efficacious. Nevertheless, they yield only long-term abstinence rates about 35%. Considering the high interest of smokers in alternative medicine, the availability of a broad range of treatment methods, of which smokers choose an intervention according to their preferences, might contribute to improve treatment outcome. While hypnotherapy (HT) is an already widely promoted alternative method for aiding cessation, considerable methodological shortcomings of studies on this topic limit the interpretability of the results. In 2006, the German Academic Advisory Committee for Psychotherapy released new guidelines that included HT as an acceptable treatment for smoking cessation. The committee conceded, however, that conclusions concerning its efficacy are restricted due to the heterogeneity of findings. Hence, further well-designed studies are required to better test the efficacy of HT in comparison to accepted treatments. This randomised, controlled trial aims to compare the efficacy of CBT and HT for smoking cessation. Further, the influence of moderating variables will be investigated. It is hypothesized that 1) participants receiving CBT will evince higher abstinence rates than those receiving HT, 2) levels of nicotine dependence, self-efficacy and motivation to change will moderate the intervention effects and 3) participants with high levels of suggestibility will evince higher abstinence rates in the HT-intervention compared to participants with low levels of suggestibility. 220 adult healthy smokers will be randomized to receive either CBT or HT. Both programmes will be conducted in 6, weekly, 90-minute group sessions. Participants will be followed up at 1, 3, 6, 9 und 12 month post-treatment. Generalized estimating equation models will be conducted to analyse group differences on abstinence rates. The models will include the above mentioned moderator variables.