Internet use patterns and their relationship with frailty in older Japanese adults

Internet use is increasing among older adults worldwide, raising interest in its potential empowering effects on healthy aging. However, the relationship between internet use and frailty among older adults remains underexplored.

We conducted a postal survey between February and March 2021 in Osaka, Japan. The survey included 1,288 respondents aged ≥ 65 years, yielding a response rate of 71.6%. Internet use patterns were divided based on the frequencies of 8 internet use activities using a k-means cluster analysis. Frailty was assessed using the Japanese Kihon Checklist with a cut-off score of ≥ 8 defining a frail status. Covariates included age, sex, living alone, economic status, work, multimorbidity, smoking, and physical activity. We employed logistic regression models to investigate the associations. Stratified analyses were also conducted by sex and age (65-74 years, ≥ 75 years).

After excluding individuals with incomplete data on internet use or long-term care users or living a nursing home, we analyzed 908 participants (45.42% female, average age 73.74 years, 25.37% frail). The K-means cluster analysis identified three internet use patterns: "less use" (n = 478), "social use" (n = 261), and "functional use" (n = 169). Logistic regressions with less use as a reference showed a negative relationship between social use and frailty (adjusted OR, 0.54; 95% CI, 0.35-0.84). The stratified analysis revealed significant relationships between social use and frailty only in males aged 65-74 years and females ≥ 75 years. A relationship was not observed between functional use and frailty.

We confirmed the segmentation of internet use patterns and its associations with frailty in older populations, noting age-sex differences. The heterogeneity in the association between internet use and frailty provides evidence for the incorporation of digital technology into health care for older adults, highlighting its role in enhancing social interaction. These findings are cross-sectional, which limits causal inference. Further longitudinal study is needed.