In Japan's super-aged society, frailty prevention is a critical public health issue. While social media use is increasing among older adults, its impact on well-being is complex and not fully understood beyond a simple active-passive dichotomy. This study aims to explore the relationship between specific social media usage patterns, users' psychosocial backgrounds, and frailty risk.

We conducted a cross-sectional online survey of 950 community-dwelling older adults in Japan. Frailty was assessed using the Kihon Checklist. A two-stage analysis was performed: 1) K-means cluster analysis was used to classify participants based on eight SOCIAL MEDIA use behavior variables, followed by a Kruskal-Wallis test to compare psychosocial characteristics across clusters; 2) Ordinal logistic regression analysis was used to identify specific behaviors associated with frailty after adjusting for covariates.

Three distinct user clusters were identified: “Balanced, reality-oriented communicators” (n = 333), “Low-engagement, passive users” (n = 419), and “Active, multi-purpose users” (n = 198). The “Low-engagement” group reported the highest levels of loneliness and unhappiness, despite high real-world social participation. Regression analysis revealed that the specific behavior of “using social media to keep a record of my own activities” (self-documentation) was independently associated with an increased risk of frailty (β = 0.105, p < .05). Conversely, high exercise self-efficacy (p < .001), sleep satisfaction (p < .01), and satisfaction with face-to-face communication (p < .001) were protective factors.

Social media usage patterns reflect underlying psychosocial states, but are not direct predictors of frailty. The motivation-driven behavior of self-documentation, rather than general activity, is a risk factor. These findings suggest that interventions should focus on the quality and function of social media use, promoting digital literacy that enhances well-being in older adults.