Dong Liu, Na Ta, Mengyang Yu, Min Liu, Yuhan Fu
Public guidance often treats screen time as a uniform exposure, yet associations with health may vary by outcome and use definition. We conducted a prospectively registered umbrella review of 59 meta-analyses published from July 1, 2015 to May 15, 2026; post-registration amendments were reported transparently. Methodological quality was assessed with AMSTAR-2, and representative effects were summarized using native-metric contrasts. General screen exposure showed small-to-moderate associations with myopia and overweight/obesity, whereas problematic or high-intensity indicators did not consistently amplify these outcomes. Sleep showed an intermediate contrast (descriptive OR ratio = 1.95; approximate interval [1.45, 2.64]). Mental health showed the largest selected within-metric contrast: time spent using social networking sites was weakly associated with depression symptoms (r = 0.11), whereas problematic use showed a larger association (r = 0.29); no correlation ratio was calculated. Neck pain was numerically higher under smartphone-overuse definitions (descriptive OR ratio = 1.72; approximate interval [0.97, 3.06]), but the interval included 1 under the independence assumption. Primary-study overlap was generally low, heterogeneity was high, and 56 of 59 reviews were rated critically low by AMSTAR-2. Publication-bias-restricted paired contrasts were available only for myopia, mental health, and sleep, precluding claims of complete cross-domain robustness. These findings motivate the Outcome-Differentiated Amplification Framework (ODAF), a hypothesis-generating account in which cumulative exposure, behavioral-postural disruption, and dysregulated engagement may matter differently across outcomes; ODAF is not a confirmed causal mechanism.