47 and 1.25, respectively] compared with those from other centres. The effect for the duration of the intervention see more appeared to be stronger than the calendar time effect (OR 1.19 vs. 1.04 per year, respectively). Further, middle-aged persons, IDUs, and persons with psychiatric

problems or with higher alcohol consumption were less likely to stop smoking. In contrast, cardiovascular events in the previous 2 years or high Framingham risk scores increased the probability of stopping smoking. Multivariable models allowed us to assess different levels of associations with care setting, calendar time, and an interaction term of the two. Further, we included a variable for the duration of the intervention at the Zurich centre which would capture a change of slope in the association with calendar time. The positive effect of the duration of the intervention at the Zurich centre was confirmed, and was very stable across all multivariable models: OR 1.24 [1.08–1.43 (multivariable model 1)], 1.23 check details [1.07–1.42 (multivariable model 2)] and 1.24 [1.07–1.45 (multivariable model 3)] per year. Thus, observed effects can probably not be explained by differences in population characteristics in different

cohort institutions. We found that structured training in smoking cessation counselling of all HIV care physicians at the Zurich SHCS centre led to increased smoking cessation (OR 1.23; 95% CI 1.07–1.42; P = 0.004), and fewer relapses of smoking (OR 0.75; 95% CI 0.61–0.92; P = 0.007), compared with participants at other SHCS institutions without similar training activities. The half day of training was conducted in a standardized way by specialized trainers whose theoretical background was based on the Prochaska/Di Clemente model of behavioural change [18,

19, 25], and the training was well accepted. At SHCS centres, cohort visits are part of routine care, and are carried out by the same physicians providing HIV care. Smoking cessation counselling activities at the Zurich centre were monitored at the cohort visits using a short structured checklist for physicians, which was completed in 84% of visits. Sodium butyrate Overall, physicians’ compliance with counselling was high, approaching 80%, indicating that counselling of smokers can be well integrated into routine care. Assessment of motivation in smokers at the Zurich centre showed that approximately half of them considered smoking cessation, but the intent to stop immediately was low (3.6%). The prevalence of smoking has decreased in the general population in Switzerland in recent years [29, 30]. The prevalence has also decreased among HIV-positive persons – overall from 60% (2000) to 43% (2010) – but has still remained significantly higher than in the general population. Several limitations of our study should be noted.