Table 3. The sensitivity to sleep, specificity for sleep, agreement, and Cohen κ coefficient for multistate categorization of Fitbit Charge 4, Garmin Vivosmart 4, and WHOOP versus polysomnography.
| Wearable and paper | Sensitivity to sleep (%), mean (SD) | Specificity for sleep (%), mean (SD) | Sensitivity to light sleep (%), mean (SD) | Sensitivity to deep sleep (%), mean (SD) | Sensitivity to REMa sleep (%), mean (SD) | Agreement (%), mean (SD) | Cohen κ for multistate categorization, mean (SD) | |
|---|---|---|---|---|---|---|---|---|
| Fitbit Charge 4 | ||||||||
| Doheny et al [33] | 93.6 (2.6) | 48.8 (17.7) | N/Ab | 96 (N/A) | 97 (N/A) | N/A | N/A | |
| Renerts et al [34] | 90 (N/A) | 73 (N/A) | N/A | 54 (N/A) | 76 (N/A) | N/A | N/A | |
| Dong et al [18] | 89.9 (4.0) | 62.2 (26.2) | N/A | N/A | N/A | N/A | N/A | |
| Mean | 91.2 (N/A) | 61.3 (N/A) | N/A | 75 (N/A) | 86.5 (N/A) | NA | N/A | |
| Garmin Vivosmart 4 | ||||||||
| Mouritzen et al [29] | 98 (3) | 30 (17) | 60 (17) | 45 (26) | 34 (26) | 48 (10) | 0.20 (0.11) | |
| Stone et al [30] | N/A | N/A | N/A | N/A | N/A | N/A | N/A | |
| Mean | 98 (3) | 30 (17) | 60 (17) | 45 (26) | 34 (26) | 48 (10) | N/A | |
| WHOOP | ||||||||
| Stone et al [30] | N/A | N/A | N/A | N/A | N/A | N/A | N/A | |
| Miller et al [31] | 95 (N/A) | 51 (N/A) | 62 (N/A) | 68 (N/A) | 70 (N/A) | 64 (N/A) | 0.47 (N/A) | |
| Miller et al [32] | 90 (N/A) | 60 (N/A) | 61 (N/A) | 64 (N/A) | 66 (N/A) | 63 (N/A) | 0.47 (N/A) | |
| Miller et al [22] | 90 (N/A) | 56 (N/A) | 58 (N/A) | 62 (N/A) | 66 (N/A) | 60 (N/A) | 0.44 (N/A) | |
| Mean | 91.7 (N/A) | 56 (N/A) | 60 (N/A) | 65 (N/A) | 67 (N/A) | 62 (N/A) | 0.46 (N/A) | |
a
REM: rapid eye movement.
b
N/A: not applicable.