Abstract
Objective: Smoking-related illnesses are the leading cause of death among Latinos, and within this ethnic group, Mexican Americans are the largest subgroup in the U.S. Understanding the factors associated with successful smoking cessation could inform interventions for this population. Although socioeconomic status (SES) is a powerful predictor of cessation outcomes in the general U.S. population, it has generally been a poor predictor of quitting smoking among Latinos. Within a sample of Spanish-speaking Mexican Americans (n = 199), this study examined a broad array of objective and subjective indicators of SES (i.e. income, education, employment, subjective social status, financial strain, insurance status) as predictors of smoking cessation. Design: Data for the current study came from a longitudinal cohort study examining the pathways linking the social determinants of health with smoking cessation. Generalized estimating equation modeling examined the association of each predictor variable with smoking abstinence across quit day, and the 3 and 26-weeks post-quit time points. Results: Results indicated that both low financial strain and insurance status predicted an increased likelihood of abstinence when controlling for covariates in the intention-to-treat analyses (p = .02 and p = .01, respectively). However, these models only approached significance in the multiple imputation analyses (all ps > .05). Other indicators of SES (i.e. income, education, employment) that have been predictive of cessation in other populations were not predictive of abstinence in this sample. Conclusions: These findings suggest that SES may indeed influence smoking cessation among Spanish-speaking Mexican Americans similarly to its influence in other populations, but that capturing the construct of SES may require assessing a broader range of SES indicators. Specifically, low financial strain and having insurance predicted a greater likelihood of achieving smoking abstinence, whereas other indicators of SES (i.e. income, education) were not predictive.