Table 5.

Intraclass correlation coefficients between each device and PSG across characteristic of sleep.

	Oura n = 35	Fitbit n = 33	Apple Watch n = 29
	ICC (95% CI)	ICC (95% CI)	ICC (95% CI)
Total Sleep Time (min)	0.74 (0.54–0.86)	0.56 (0.28–0.76)	0.85 (0.70–0.93)
Light Sleep (min)	0.40 (0.08–0.64)	0.52 (0.22–0.73)	0.37 (0.00–0.64)
Deep Sleep (min)	0.32 (−0.01–0.59)	0.36 (0.02–0.62)	0.13 (−0.24–0.47)
REM (min)	0.27 (−0.06–0.55)	0.13 (−0.22–0.45)	0.37 (0.01–0.64)
Sleep Latency (min)	0.95 (0.90–0.97)	0.95 (0.89–0.97)	0.94 (0.87–0.97)
WASO (min)	0.63 (0.38–0.80)	0.41 (0.08–0.66)	0.72 (0.48–0.86)
Sleep Efficiency (%)	0.74 (0.55–0.86)	0.56 (0.28–0.76)	0.85 (0.71–0.93)

Notes. ICCs represent comparisons between PSG and each device (Oura, Fitbit, and Apple Watch). Sample size for sleep latency is n = 34 for PSG vs. Oura, n = 32 for PSG vs. Fitbit, and n = 29 for PSG vs. Apple Watch. Reliability is deemed ‘poor’ if the ICC is below 0.4, ‘fair’ if between 0.4 and 0.59, ‘good’ if between 0.60 and 0.74, and ‘excellent’ if above 0.75.