Relationship between mobile phone use and multidimensional frailty: evidence from UK Biobank

To investigate the longitudinal temporal relationships between weekly mobile phone call time and four dimensions of frailty.

Data were derived from three instances (T1, T2, T3) of the UK Biobank. Frailty encompassed four dimensions: psychological, physical, social, and cognitive. Separate Random Intercept Cross-Lagged Panel Models (RI-CLPM) were constructed for each frailty dimension with weekly mobile phone call time, measured by self-reported weekly mobile phone call time standardized to baseline. Chi-square difference tests and the BIC were adopted for model comparisons.

This study included 5,533 participants aged 40-70 years. Unconstrained models outperformed constrained models for all dimensions (P < 0.001). For psychological frailty, after false discovery rate (FDR) correction, only the path from psychological frailty T1 to weekly mobile phone call time T2 was marginally significant (β = 0.086, SE = 0.031, P = 0.008, P_FDR = 0.048), while the other cross-lagged paths (psychological frailty T2 to weekly mobile phone call time T3: β = 0.053, SE = 0.023, P = 0.025, P_FDR = 0.101; weekly mobile phone call time T2 to psychological frailty T3: β = 0.067, SE = 0.027, P = 0.010, P_FDR = 0.069) were not significant, thus robust bidirectional associations were not supported. Physical frailty was not significantly associated with subsequent weekly mobile phone call time at either interval (T1 to T2: β = 0.035, SE = 0.048, P = 0.465; T2 to T3: β = 0.052, SE = 0.029, P = 0.075; both P_FDR > 0.05). Social frailty was significantly associated with subsequent weekly mobile phone call time at T2 to T3 (β = 0.078, SE = 0.025, P = 0.002, P_FDR = 0.033). No cross-lagged associations were observed for cognitive frailty (all P > 0.05). The random intercepts of weekly mobile phone call time were all significant (P < 0.001), suggesting stable inter-individual differences rather than within-individual fluctuations. In summary, after FDR correction, only two cross-lagged associations reached statistical or marginal significance: one from psychological frailty T1 to weekly mobile phone call time T2 (β = 0.086, P_FDR = 0.048), and the other from social frailty T2 to weekly mobile phone call time T3 (β = 0.078, P_FDR = 0.033), each observed in only one time interval, with no significant associations in other directions or intervals. These results indicate weak and inconsistent temporal associations rather than consistent or robust bidirectional effects.

This study examined prospective within-person associations between weekly mobile phone call time and frailty dimensions in a middle-aged and older adult sample (baseline age 40-70 years). After FDR correction for multiple testing, most cross-lagged associations were not statistically significant. Only two paths reached significance or marginal significance: social frailty T2 → weekly mobile phone call time T3 (β = 0.078, P_FDR = 0.033) and psychological frailty T1 → weekly mobile phone call time T2 (β = 0.086, P_FDR = 0.048), each occurring in only one time interval. No consistent or robust bidirectional patterns were observed across dimensions or intervals. The significant random intercepts suggest stable inter-individual differences in phone use over time. Overall, these findings do not support clear temporal relationships between weekly mobile phone call time and most frailty dimensions, and the few observed associations are weak.