Sometimes, what feels most alien isn’t the stars, it’s what we find in each other. He landed on a Tuesday, though no one noticed. The sky di...

Amidst the hum of enforced connection, a silent rebellion of human touch and shared air begins to stir, hinting at a world yearning for tangible links.

Background: The rapid advancement of digital technologies has profoundly transformed communication practices. However, this technological revolution has also led to "digital isolation," a form of social disconnection caused by limited or absent engagement with digital communication tools, including smartphones, computers, email, and the internet. This issue is particularly concerning for older adults, as it may increase their likelihood of developing mental health disorders, with depression being a primary concern. Although digital isolation has been studied less frequently than traditional social isolation, it may be a significant contributor to both the initiation and progression of depression in this population. Objective: This investigation seeks to assess longitudinal relationships between multidimensional digital disengagement (encompassing four dimensions: mobile device utilization, computer interaction, electronic correspondence, and web-based engagement) and incident depression among older adults, utilizing longitudinal data from the nationally representative National Health and Aging Trends Study (NHATS). Methods: The analysis was conducted based on the NHATS dataset, a nationally representative longitudinal survey employing multistage sampling to represent community-dwelling Medicare beneficiaries aged 65 and older in the US. We analyzed data from 2011 (Round 1) to 2018 (Round 8), including 8,199 participants in the discovery and validation cohorts. Digital isolation was measured using a 4-item index based on self-reported non-use of mobile phones, computers, email, and the internet. Participants were categorized into high (aggregate score ≥3) or low (aggregate score ≤2) digital isolation groups. Weighted Cox regression models with proportional hazards assumptions were employed to quantify longitudinal associations between the digital isolation index (and its individual components) and incident depression, incorporating multivariable adjustment for sociodemographic characteristics (age, sex, race or ethnicity), socioeconomic indicators (education level, family income, marital status), and clinical profiles (tobacco use history, multimorbidity burden). Time-to-event analyses were visualized through Kaplan-Meier estimators, complemented by prespecified subgroup analyses evaluating effect modification patterns through interaction term testing. Results: A high level of digital isolation, as measured by the composite index, was associated with a significantly greater risk of incident depression (fully adjusted model: HR=1.35, 95%CI 1.18-1.55; P