Data Analysis
Data entry was performed at the University of California at Los Angeles by one of the authors (Brook). All analyses used SAS 9.2 (SAS Institute, Inc., Cary, NC, USA). Descriptive statistics were used to characterize study variables. The primary outcome used to evaluate the program was the change in the nurses’ self -reported frequency of cessation interventions and referral of patients to the quitline. Differences between responses about clinical practice before and after the educational intervention were examined using nonparametric sign tests. We used the McNemar test to examine the increase, from baseline to 3 months, in the proportion of nurses who consistently (“always” or “usually”) intervened using the five A’s and referral to the quitline. Additionally, we compared the frequency of those who consistently (“always” or “usually”) intervened with smokers using the five A’s, and referred smokers to a quitline by the nurses’ smoking status (dichotomized as current vs. former/never smoker). Baseline and 3-month data were analyzed separately for smokers versus nonsmokers using chi-square tests. Subsequently, proportional differences of pre-post change in frequency of use of the various elements of the intervention by smoking status were examined with nested analysis of participants within hos
pitals using generalized estimating equation modeling for dichotomous outcomes. We calculated the OR for the difference in performance between the smokers and nonsmokers at baseline and 3 months