Wearable AI for continuous pediatric patient monitoring: a systematic review and meta-analysis
TL;DR
Wearable continuous monitoring devices integrating analytical algorithms show promising diagnostic accuracy and potential clinical benefits in pediatric populations, however, the evidence base remains limited by moderate heterogeneity and a predominance of observational designs.
Abstract
Continuous physiological monitoring in pediatric patients is crucial for the early identification of critical events such as cardiac arrhythmias, seizures, and sepsis. Wearable devices embedded with analytical algorithms offer a novel approach for non-invasive, continuous surveillance. Despite rapid technological advances, a comprehensive evaluation of their diagnostic accuracy, clinical effectiveness, safety, and acceptability in pediatric cohorts remains lacking. To systematically evaluate the diagnostic performance, clinical outcomes, safety, and usability of wearable continuous monitoring devices with integrated analytical algorithms in pediatric patients (neonates to adolescents), compared with standard monitoring methods or wearable devices without analytical integration. A systematic search of PubMed, Embase, Cochrane Library, Web of Science, and Scopus was conducted from inception through March 31, 2026. Eligible studies included randomized controlled trials, prospective and retrospective cohorts, and diagnostic accuracy investigations involving wearable devices with embedded analytical algorithms for continuous monitoring in pediatric populations. Independent dual reviewers performed study selection, data extraction, and risk of bias assessment. Diagnostic accuracy metrics were synthesized using bivariate random-effects models. Clinical outcomes were pooled using random-effects meta-analyses. Certainty of evidence was appraised using the GRADE framework. From 3,842 screened records, 24 studies enrolling 4,376 pediatric patients were included. Wearable analytical devices demonstrated a pooled sensitivity of 87.4% (95% CI 82.1–91.5) and specificity of 89.2% (95% CI 85.0–92.4) for detecting critical events. Device use was associated with reduced hospitalization rates (RR 0.68; 95% CI 0.52–0.89) and shorter time to clinical intervention (MD − 1.8 h; 95% CI − 2.7 to − 0.9). Subgroup analyses showed consistent diagnostic accuracy across device modalities and clinical conditions. No significant publication bias was detected. Wearable continuous monitoring devices integrating analytical algorithms show promising diagnostic accuracy and potential clinical benefits in pediatric populations. However, the evidence base remains limited by moderate heterogeneity and a predominance of observational designs. Further large-scale, high-quality randomized controlled trials are warranted to validate long-term efficacy, safety, and cost-effectiveness before widespread clinical adoption.