69 (95% CI = 0.66–0.72; Supplementary Fig. E2).2 We used data from a large English household survey to BIBW2992 mw assess the validity of a single-item rating of motivation to quit smoking: the Motivation To Stop Smoking (MTSS) scale. The scale effectively combines both current desire and intention to stop smoking – two key components of motivation (Smit et al., 2011) – into one single response scale, ranging from 1 (lowest) to 7 (highest level of motivation to stop smoking). Scores on the MTSS predicted quit attempts in the following 6 months in a linear fashion. The degree of association was good, with those at the top of the scale having 6.8 times the odds of trying
to stop than those at the bottom, as was the degree of accuracy. The accuracy of our measure this website of motivation in discriminating between smokers who quit and who did not quit during follow up was 0.67, which is considered to be broadly acceptable (Hosmer and Lemeshow, 2000). In the tobacco research literature, the reporting of psychometric indicators (sensitivity, specificity, ROCAUCs) for predictors of behavioral change from prospective research is scarce. A study conducted in the 1990s compared the validity of the Stage of Change Model with
a prediction equation that combined four smoking- and quitting-related variables in predicting long-term cessation and reported ROCAUCs of 0.55 and 0.69, respectively ( Farkas et al., 1996). An internet survey conducted in the 2000s assessed the validity of two measures of dependence in predicting short-term cessation and reported ROCAUCs that were either not significant or very marginal (0.55; Etter, 2005). In a similar but more recent study, the same research group reported ROCAUCs between 0.67 and 0.76 for the same two measures of dependence in predicting abstinence at 8-day follow-up but again marginal ROCAUCs for the 31-day follow-up (0.51–0.58; Courvoisier and Etter, 2010). We could not find literature on ROCAUCs for predictors of quit attempts. It should be noted that we conducted our analysis on all respondents who were smokers at the time of our survey, but that these respondents
from comprise a heterogeneous group in terms of personal and smoking characteristics. For example, it has been shown that low level smokers are more motivated to quit than moderate-to-heavy smokers (Kotz et al., 2012). Other factors have been shown to be associated with motivation to quit as well, including age, nicotine dependence and previous quit attempts (Marques-Vidal et al., 2011). However, our aim was to evaluate the predictive validity of the MTSS across all subgroups of smokers to maximize generalizability and usability of the scale. An additional point of interest is the significant minority of smokers who made a quit attempt soon after reporting no intention, desire or belief that one should stop smoking (i.e.