Evaluation of the Agreement Between Research-Grade Actigraphy Sleep, Consumer-Grade Smartwatches and Self-Reported Sleep Diaries in Masters Endurance Athletes

Abstract

Sleep monitoring is a tool widely used to support recovery and performance in endurance athletes. This study aimed to assess agreement between research-grade actigraphy (ActiGraph GT9X), consumer-grade smartwatches (Garmin), and self-reported sleep diaries in masters endurance athletes. Seventy athletes (43 males, 46.3 +/- 7.3 years; 27 females, 49.3 +/- 8.3 years) wore ActiGraph and smartwatch devices on their non-dominant wrist while maintaining a self-reported sleep diary for seven consecutive nights. ActiGraph recorded the shortest total sleep time (332 +/- 87 min), whereas the diary and smartwatch recorded longer sleep durations by 109 and 126 min, respectively (p < 0.001). Sleep efficiency (%) was also higher in the sleep diary and smartwatch compared to ActiGraph, with mean biases of -5.9% and -4.1%, respectively. Sleep diary values closely agreed with smartwatch values (ICC = 0.880, 95% CI: 0.624 to 0.946), while poor agreement was found between ActiGraph and the sleep diary (ICC = 0.190, 95% CI: -0.149 to 0.459). Proportional bias was evident in both the sleep diary and smartwatch, with greater differences in total sleep time and efficiency observed in athletes with shorter durations and lower sleep efficiency, respectively. Sex differences emerged, with stronger agreement between smartwatch and ActiGraph in sleep efficiency in females (ICC = 0.690, 95% CI: 0.336 to 0.857) than males (ICC = 0.481, 95% CI: -0.020 to 0.723). Findings suggest that both consumer-grade devices and self-reported sleep diaries report longer sleep durations and higher sleep efficiency relative to actigraphy. Sleep metrics from these methods should be interpreted with caution, particularly in athletes with shorter or more fragmented sleep.