Identifying subgroups of preventive health behaviors and associated factors among older adults in Hainan, China: a latent class analysis

Preventive health behaviors (PBs) are critical for reducing the burden of non-communicable diseases in aging populations, yet PBs among the elderly in China are notably insufficient. Patterns of multiple PBs among older adults in China remain poorly characterized, and the potential influencing factors are unclear. This study aimed to identify subgroups of preventive health behaviors and examine their associated factors among older adults in Hainan, China. Data were drawn from the 2023 wave of the National Health Service Survey in Hainan Province, comprising 4,112 adults aged 60 years and older. Six self-reported PBs—smoking avoidance, alcohol abstinence, physical activity, body mass index maintenance, oral hygiene, and physical examination participation—were included in a latent class analysis (LCA) to identify behavioral subgroups. A two-level multilevel generalized linear model with random intercepts at the household level was used to examine sociodemographic, economic, and resource-related factors associated with class membership. LCA identified two distinct latent classes: a More favorable PB group (76.4%) and a Less favorable PB group (23.6%). The Less favorable PB group was characterized by high prevalences of current smoking (92.6%) and alcohol consumption (56.6%). In the fully adjusted multilevel model, factors independently associated with membership in the More favorable PB group included older age (OR = 1.07, 95% CI: 1.05–1.08), female sex (OR = 17.62, 95% CI: 1.46–37.61), being married (OR = 1.54, 95% CI: 1.18–2.03), college education or above (OR = 2.75, 95% CI: 1.36–5.56), and greater social support within the family (OR = 1.02, 95% CI: 1.01–1.07). Older adults residing in the central region of Hainan were less likely to be in the More favorable PB group compared with those in the eastern region (OR = 0.81, 95% CI: 0.70–0.93). Preventive health behaviors among older adults in Hainan cluster into two distinct patterns, primarily differentiated by smoking and alcohol consumption. The identified subgroups and their sociodemographic correlates highlight the need for targeted, behavior-specific interventions that address clustered risk behaviors, particularly among younger, male, and less educated older adults.