Early-Life Screen Exposure and Language Development Delay in Children: A Systematic Review and Meta-Analysis

Abstract

Screen media are now near-universal in the early childhood environment, and language developmental delay is among the most frequently identified conditions of the preschool period. Whether quantified early-life screen exposure is associated with language delay across heterogeneous populations, outcome domains and measurement approaches remains incompletely resolved. We aimed to systematically review and quantitatively synthesise observational evidence on the association between early-life screen exposure and language development delay in children. MEDLINE, Embase, Scopus, Web of Science, PsycINFO and CENTRAL were searched from inception to 6 September 2026. Observational studies reporting a quantified screen-media exposure measured in early childhood together with a validated language or communication outcome were eligible. Two reviewers extracted data in duplicate. Risk of bias was appraised with the design-specific Newcastle-Ottawa Scale. Random-effects models (DerSimonian-Laird) pooled the most fully adjusted odds ratios for dichotomous delay and Fisher z-transformed coefficients for continuous language ability. Thirty-two reports describing 33 study-cohorts and 58,740 children were included; 18 study-cohorts were at low and 15 at moderate risk of bias, and none at high risk. For the highest versus the lowest category of early-life screen exposure, the pooled odds ratio for any language or communication delay was 2.47 (95% CI 1.59 to 3.84; 7 study-cohorts; I² = 58.3%; τ² = 0.185; 95% prediction interval 0.75 to 8.10). For expressive language delay the pooled odds ratio was 1.53 (95% CI 1.20 to 1.95; 3 study-cohorts; I² = 0.0%). The pooled correlation between early-life screen exposure and continuous language ability was -0.22 (95% CI -0.28 to -0.15; 5 study-cohorts; I² = 45.4%). The Egger regression intercept was 0.62 (P = 0.709); trim-and-fill imputed two study-cohorts and returned an adjusted odds ratio of 2.05 (95% CI 1.33 to 3.16). Across a large and geographically diverse observational evidence base of low-to-moderate risk of bias, higher early-life screen exposure was consistently associated with poorer language outcomes, with a magnitude that was small on the continuous scale and moderate on the categorical scale. Heterogeneity in exposure operationalisation and the predominance of caregiver-reported measurement constrain causal interpretation.