Statistical analyses 17-DMAG Phase 2 Data were analyzed using SPSS Version 14.0. Demographic and psychosocial variables were characterized by descriptive statistics. Participants who did not report their smoking status were categorized as ��smokers.�� One-way analysis of variance (ANOVA) and chi-square analysis were used to assess potential differences between study conditions (intervention and control) on demographic and psychosocial variables. We conducted 2 (Study Conditions) �� 5 (Timepoints) repeated measures of ANOVAs to evaluate changes in psychosocial variables over time by study group. Significant interactions were examined using a simple main effects analysis with pairwise comparisons and Bonferroni correction. The kappa (��) statistic was used to measure agreement between the smoking status of self-reported and the biomarker CO.

Finally, univariate and multivariate logistic regression analyses were performed to examine predictors of smoking cessation at 6-month follow-up. In the multivariable analysis, the treatment group was forced to enter into the regression models, whereas all other factors were selected into the models by stepwise forward method. Results Comparison of baseline measures for treatment and control conditions No significant differences in gender (89.1% male vs. 86.2% male), age (43.9 vs. 45.0 years), annual income (large majority 72%�C77% earning <$20,000 per year), education (majority 33%�C35% with less than high school education), marital status (majority 78%�C84% married), smoking status (majority 46%�C56% smoke regularly), or influence of physician's advice in motivating to quit (majority 41%�C45% considered themselves not at all influenced) were found between intervention and control groups at baseline (Table 1).

In terms of baseline psychosocial measures, no significant differences were found between intervention and control groups. Overall, most participants were favorable toward the NRT. About 32% of the control had ceased smoking by the end of the study period, suggesting that NRT affected the cessation rate. However, NRT combined with MI produced a much higher cessation rate of 67% for the intervention group. Effect of treatment conditions on psychosocial variables over time All participants over time showed increased risk perceptions, Huynh�CFeldt, F(2.7, 305) = 29.756, p < .001; self-efficacy, F(2.5, 282.4) = 83.

837, p < .001; and cons of smoking, F(3.3, 347.5) = 19.538, p < .001. Participants also showed decreasing pros of smoking over time, F(3.3, 352) = 6.451, p < .001(Table 2). Table 2. Changes in psychosocial variables over time by study condition M (SD) A main effect of study condition was found Drug_discovery for risk perceptions, F(1, 115) = 9.891, p < .01; self-efficacy, F(1, 113) = 5.318, p < .05; and cons of smoking, F(1, 107) = 10.047, p < .01.