We included a covariable for the duration of the smoking cessation intervention at the Zurich centre. This variable was set to 0 for all settings and years except for the Zurich centre, where it was assigned values of 1, 2 and 3 for the intervention years Bleomycin in vivo 2008, 2009 and 2010, respectively. The completion of checklists was

stopped in December 2009 but the regular training was maintained. We therefore hypothesized that the positive effects would continue for some time. Because differences in patient characteristics between the different settings could potentially contribute to the effect observed, we fitted a second multivariable model with additional covariables: sex, age (grouped as <30, 30–49 and ≥50 years), HIV transmission category [with injecting drug users (IDUs) separated into former and current IDUs], occurrence of a cardiovascular event in the previous 2 years, and current psychiatric treatment or depression. Because Framingham risk scores are only defined for individuals aged 30–74 years, and collection of information on alcohol use was not started

before 2005 in the SHCS, the sensitivity analysis (model 3) could only be performed on a subset of participants aged 30–74 years with information on alcohol use. We used the upper quartiles of the 10-year risks for CVD, CHD and MI as covariables. As Framingham scores incorporate information on current smoking, we lagged these scores by 6 months to avoid reverse causality with our outcome of interest. Analyses were performed Obeticholic Acid supplier using R (version 2.10.1, 14.12.2009; The R Foundation for Statistical Computing, Vienna, Austria) [28] and Stata software (version 11.2; StataCorp, College Station, TX). A total of 11 056 SHCS participants with available smoking information had 121 238 follow-up

visits and 64 118 person-years of follow-up between April 2000 and December 2010, and contributed to the smoking prevalence analyses (Fig. 1). During the intervention at the Zurich centre from November 2007 to December 2009, 1689 participants were seen at this centre. The effect of the intervention was assessed in a smoking cessation analysis among 5805 smokers with at least three follow-up visits, and in a relapse analysis among 1953 participants who had stopped smoking over at least two Phospholipase D1 consecutive semi-annual visits. Participants at the Zurich centre were around 6 years older than those in other settings (Table 1), and were less likely to be alcohol abstinent (36% vs. 55% in other centres, and 50% in private care). Private physicians tended to care for more men who have sex with men (50% vs. 42% at the Zurich centre, and 26% in other centres), and for those with less advanced HIV disease [20% in Centers for Disease Control and Prevention (CDC) stage C vs. 24% at the Zurich centre, and 28% in other centres].