Clustering of Modifiable Behavioral Risk Factors and Their Association with All-Cause Mortality in Taiwan’s Adult Population: a Latent Class Analysis

Abstract

BackgroundModifiable risk behaviors, such as smoking, diet, alcohol consumption, physical activity, and sleep, are known to impact health. This study aims toward identifying latent classes of unhealthy lifestyle behavior, exploring the correlations between sociodemographic factors, identifying classes, and further assessing the associations between identified latent classes and all-cause mortality.MethodsFor this study, the data were obtained from a prospective cohort study in Taiwan. The participants’ self-reported demographic and behavioral characteristics (smoking, physical activity, alcohol consumption, fruit and vegetable intake, and sleep) were used. Latent class analysis was used to identify health-behavior patterns, and Cox proportional hazard regression analysis was used to find the association between the latent class of health-behavior and all-cause mortality.ResultsA complete dataset was obtained from 290,279 participants with a mean age of 40 (12.4). Seven latent classes were identified, characterized as having a 100% likelihood of at least one unhealthy behavior coupled with the probability of having the other four unhealthy risk behaviors. This study also shows that latent health-behavior classes are associated with mortality, suggesting that they are representative of a healthy lifestyle. Finally, it appeared that multiple risk behaviors were more prevalent in younger men and individuals with low socioeconomic status.ConclusionsThere was a clear clustering pattern of modifiable risk behaviors among the adults under consideration, where the risk of mortality increased with increases in unhealthy behavior. Our findings can be used to design customized disease prevention programs targeting specific populations and corresponding profiles identified in the latent class analysis.