Examining Antepartum Quit Attempts Using A Latent Factor Model Of A Hypothetical Cigarette Purchase Task

Previous research has validated the use of hypothetical purchase tasks to measure smoking demand among pregnant women. This study extends that research by (1) examining the factor loading pattern of a hypothetical cigarette purchase task (CPT) in a sample of pregnant women who smoke, and (2) comparing the ability of the latent factor solution to predict antepartum quit attempts relative to more conventional predictors. Participants were 665 pregnant women seeking enrollment in a smartphone-based smoking-cessation trial. Data were taken from an intake assessment that included the CPT, the Kirby delayed discounting task, sociodemographic and smoking-history questionnaires, and assessment of antepartum quit attempts. Bivariate analyses compared sociodemographic and smoking characteristics between women who reported zero versus \u3e one antepartum quit attempt. Confirmatory Factor Analysis using a Principal Component Analysis method was used to assess whether the five CPT indices (Intensity, Omax, Pmax, Breakpoint, Alpha) loaded onto two latent factors (Amplitude & Persistence). Finally, stepwise regression modeling was conducted to examine associations between CPT latent factors and antepartum quit attempts adjusting for other variables that were significant at the bivariate level. All associations with p \u3c 0.05 were retained in final models. Factor analysis confirmed a two-factor solution to the CPT whereby Intensity of demand loaded exclusively on one factor (Amplitude), and all other indices onto another factor (Persistence). Significant predictors of antepartum quit attempts in the final adjusted regression model included CPD antepartum, time to first cigarette (TFC) antepartum, Persistence, and menthol use. As CPD and Amplitude were highly correlated, a second regression was conducted excluding CPD. The variables retained in that model were TFC antepartum, Amplitude, Persistence, and race/ethnicity. These results extend the two-factor solution of CPT indices to pregnant women. The latent factors Amplitude and Persistence are significantly and independently associated with antepartum quit attempts, although the significance of the association of Amplitude with quit attempts is conditional on the presence of CPD as a predictor. These results lend further support to the potential utility of the CPT for examining individual differences in attempting to quit smoking upon learning of pregnancy