Patterns and predictors of smoking by race and medical diagnosis during hospital admission: A latent class analysis

Abstract

Hospital-based tobacco treatment programs provide tobacco cessation for a diverse array of admitted patients. Person-centered approaches to classifying subgroups of individuals within large datasets are useful for evaluating the characteristics of the sample. This study categorized patients who received tobacco treatment while hospitalized and determined whether demographics and smoking-related health conditions were associated with group membership. Chart review data was obtained from 4854 patients admitted to a large hospital in South Carolina, USA, from July 2014 through December 2019 who completed a tobacco treatment visit. Smoking characteristics obtained from the visit interview were dichotomized, and then latent class analysis (LCA) was conducted to categorize patients based on smoking history and interest in stopping smoking. Finally, logistic regressions were used to evaluate demographics and smoking-related health conditions as predictors of class membership. LCA generated 5 classes of patients, differentiated by heaviness of smoking and motivation to quit. Patients who were black/African American were more likely to be lighter smokers compared to white patients. Hospitalized patients with a history of hypertension, diabetes, and congestive heart failure were more likely to be motivated to quit and also were more likely to be lighter smokers at the time of hospitalization. Hospitalized patients who smoke and receive tobacco treatment are heterogeneous in terms of their smoking histories and motivation to quit. Understanding latent categories of patients provides insight for tailoring interventions and potentially improving tobacco treatment outcomes.

Keywords

• tobacco
• smoking
• hospital
• latent class analysis
• health outcomes