Digital-Use Discontinuation as a Follow-Up Marker of Cognitive and Depressive Vulnerability in Later Life: A Multicohort Longitudinal Study

Background Digital inclusion in later life is often measured as static user/non-user status, although older adults may enter, remain in, remain outside, or leave digital environments. Objective To examine whether digital-use transitions, particularly discontinuation, mark cognitive and depressive vulnerability. Methods We analyzed harmonized longitudinal data from CHARLS, ELSA, HRS, and SHARE. Adjacent-wave internet-use transitions were classified as persistent non-use, adoption, persistent use, or discontinuation. Cohort-specific forward, reverse, pre-transition, and marker analyses were pooled with random-effects methods where appropriate. Results Across 250,379 transition intervals, discontinuation was preceded by lower cognition and more depressive symptoms, and was associated with lower subsequent cognition (β = −0.145) and higher depressive symptoms (β = 0.094) than sustained use. Adoption was associated with higher cognition and, on average, fewer depressive symptoms than persistent non-use. Baseline vulnerability predicted lower adoption and higher discontinuation. Adding discontinuation to baseline risk models produced only small gains in discrimination.