Development of a weighted scoring model for social activities to predict disability incidents among older Japanese adults

Abstract

Background: Social activities contribute to health improvements in older adults, but methods for evaluating these activities are not yet established. We developed a scoring model for social activity, weighted by specific activities, to assess the association between disability incidence in older adults and social activities.

Methods: Data were obtained from Japan's National Center for Geriatrics and Gerontology Study of Geriatric Syndromes (NCGG-SGS). Social activity was evaluated across 16 domains. Disability was determined using data extracted from Japan's long-term care insurance system.

Results: Data from 4998 older adults were analyzed; among them, 422 (8.4 %) developed a disability within 35 months (Interquartile range: 32-39). The Cox proportional hazards model was used to assess 16 domains of social activity. The results yielded risk factors for disability incidence in six social activity domains: work, travel, hobbies, babysitting, family caregiving, and events. The coefficients for these activities were assigned weights of 3, 3, 2, 1, 1, and 1, respectively. The weighted social activity scoring model significantly improved the ability to predict disability incidence when the number of social activities in which individuals participated was considered (social activity score: area under the curve [AUC] 0.691, 95 % confidence interval [CI] 0.664-0.717; number of social activities: AUC 0.681, 95 % CI 0.654-0.707, P = 0.042).

Conclusions: The composite score derived from the weighted social activity scoring model serves as a valuable tool due to its enhanced predictability, which complements established background factors associated with the incidence of disability in older adults.

Keywords: Disability incidence; Older adults; Social activity; Weighted scoring model.