Comparing the reliability of sleep metrics across two wrist actigraphs: Philips Actiwatch Spectrum Plus and ActiGraph CentrePoint Insight watch

Abstract Background Actigraphy is widely used in research settings to objectively assess sleep–wake patterns; however, inconsistencies in device algorithms, hardware capabilities, and scoring procedures pose challenges for comparing data across devices. This study investigated the agreement between two commonly used wrist actigraphs (i.e., Philips Actiwatch Spectrum Plus & ActiGraph CentrePoint Insight) and offers a systematic procedure to improve comparability in sleep indicators. Methods Thirteen participants (Mage = 19 years, SD = 3 years; 85% female; 62% Asian, 15% Latinx, 15% multiracial, and 8% White) wore both devices concurrently on the same wrist for 4 to 7 nights. Key sleep metrics were compared between the two devices, including in-bed time, out-of-bed time, sleep onset/offset, rest duration, sleep duration, wake after sleep onset (WASO), sleep efficiency, latency, and wake-to-rise interval. Agreement between devices was assessed using intraclass correlation coefficients (ICCs), Standard Error of Measurement (SEm), and Bland–Altman plots with limits of agreement. Results Excellent agreements were achieved for in-bed time, sleep onset, sleep offset, and sleep duration, whereas out-of-bed time and rest duration showed moderate agreement. Indicators dependent on fine-grained sleep–wake transitions (WASO, sleep efficiency, latency, and wake-to-rise interval) demonstrated poor agreements, possibly reflecting hardware and algorithmic differences between devices. Conclusions Findings illustrate challenges in cross-device comparability and highlight the importance of understanding device-specific scoring characteristics before integrating or transitioning between actigraphy systems. This study provides a practical coding framework to refine ActiGraph Centrepoint metrics and offers recommendations for improving comparability across actigraphy devices.