Lifestyle behavior patterns and mortality among adults in the NHANES 1988-1994 population: A latent profile analysis

Evidence suggests interdependent associations of individual modifiable behaviors with health outcomes. However, such interrelations have not been accounted for in previous behavior-outcome associations. We conducted latent profile analysis (LPA) on self-reported levels of alcohol consumption, restaurant dining, vitamin/mineral supplement use, physical activity (PA) and smoke exposure (first- and second-hand smoke) separately for smokers (N = 4530) and non-smokers (N = 13,421) using data from the third National Health and Nutrition Examination Survey (NHANES III) to identify subgroups with similar levels within and across behaviors. Cox-proportional hazards models were used to compare mortality rates between subgroups from cancer, cardiovascular disease (CVD) and all-causes at an average of 16.4 (±6.1) years follow-up. Five behavioral typologies were identified in non-smokers ("Moderates", "Low Risk Factors", "Restaurant Diners", "Moderate Passive Smokers" and "Heavy Passive Smokers"), and four in smokers ("Moderates", "Low Risk Factors", "Heavy Smokers" and "Physically Active"). As a group, "Moderates" had levels of each behavior that were not significantly different from at least one other group. Compared to "Moderates", in non-smokers "Restaurant Diners" had lower hazard from all-cause (hazard ratio (HR):0.84, 95% CI:0.74-0.97) and CVD (HR:0.59, 0.43-0.82) mortality, while "Low Risk Factors" had higher cancer mortality (HR:1.38,1.03-1.84). In smokers, compared to "Moderates", higher hazards for mortality were found for "Heavy Smokers" (all cause: HR:1.34, 1.12-1.60; CVD: HR:1.52, 1.04-2.23; cancer: HR:1.41 1.02-1.96) and "Low Risk Factors" (all-cause: HR:1.58, 1.14-2.17). Taken together, when restaurant dining, PA and smoking exposures are grouped together, novel predictions for mortality occur, suggesting data on multiple behaviors may be informative for risk stratification.