Background: The Community Healthy Activities Model Program for Seniors (CHAMPS) survey, summarized into weekly caloric expenditures, is a common physical activity (PA) assessment tool among older adults. Specific types of PA reported in the CHAMPS have not been systematically analyzed. We applied latent class analysis to identify the patterns of PA among sedentary older adults with diabetes reported in the CHAMPS survey.
Methods: Latent class models of PA were identified using the CHAMPS survey data reported by 115 individuals aged ≥60 years with type 2 diabetes whom volunteered for a clinical study of PA. Multinomial logistic regression was used to assess independent predictors of a specific latent class, including age, sex, and performance in physical function tests.
Results: Ninety-three percent of the participants were classified into 3 latent classes. Participants in latent class 1 (60.9%) primarily reported domestic-focused activities. Participants in latent class 2 and 3 (19.5% and 19.6%, respectively) reported domestic-focused activities, in addition to leisure-time physical activities and structured exercise activities. Latent class 1, with more women than men (73% vs.27%), had the lowest caloric expenditure, whereas class 3, with fewer women than men (28% vs. 72%), had the highest caloric expenditure (all p < 0.001). Latent class 2 had the fastest Timed-Up- and Go (7.65 ± 1.28 s; p = 0.03).
Conclusions: Individual PA response in CHAMPS can be categorized using latent class models into meaningful patterns which can inform PA interventions. Customized PA programs should consider the heterogeneity of the activities among sedentary older adults.
Trial registration: ClinicalTrials.gov Identifier NCT00344240; retrospectively registered 23 June 2006.
Keywords: CHAMPS; Diabetes; Physical activity; Physical function.