Validation of a national actigraph in children and adolescents

Abstract

Introduction: Sleep is a neurophysiological process essential to child and adolescent development, playing a central role in cognitive, behavioral, and emotional functions. Early identification of sleep disorders is key to prevent health impairments, which highlights the need for reliable evaluation methods. While polysomnography (PSG) offers detailed analysis of sleep architecture, actigraphy has emerged as a less invasive, more accessible alternative capable of prolonged sleep monitoring. However, validation of actigraphy in children and adolescents is still lacking. Objectives: This study aimed to validate the ActTrust2 actigraph and to compare the performance of three algorithms (Actiware, Cole-Kripke, and Sadeh) in estimating sleep parameters in children and adolescents, using PSG as the reference. Methods: Thirty-three participants, aged 2 to 16 years, underwent simultaneous PSG and actigraphy recording. Evaluated parameters included total sleep time, sleep latency, wake after sleep onset (WASO) and sleep efficiency. The three algorithms were compared using performance metrics such as accuracy, sensitivity, specificity, positive predictive value and negative predictive value. Data normality was assessed using the Shapiro–Wilk test, followed by nonparametric analyses including Friedman, Wilcoxon, and Spearman correlation tests with Bonferroni correction. Results: The Actiware algorithm showed the best overall performance, with higher accuracy (82.4%), sensitivity (86.6%), lower bias, and strong correlations with PSG for total sleep time, WASO, and sleep efficiency. However, its correlation for sleep latency was weak non-significant. The Cole-Kripke algorithm demonstrated intermediate performance, with balanced sensitivity (76.0%) and specificity (78.6%), and moderate bias. Despite higher bias, lower accuracy (69.5%), and sensitivity (65.6%), the Sadeh algorithm was the only one that did not show a significant difference in sleep latency compared to PSG suggesting better agreement. Conclusions: The results demonstrate that the ActTrust2 actigraph, when combined with appropriate algorithms, is a valid and reliable tool for assessing sleep in children and adolescents.