Table 2 Sensitivity and specificity: NWT algorithms >60 to >180 minutes using accelerometry versus diary.

From: 24 h-accelerometry in epidemiological studies: automated detection of non-wear time in comparison to diary information

NWT algorithma

ActivE study (N = 47)

KORA FF4 study (N = 559)

number of NWT periods

sensitivityb

specificityb

number of NWT periods

sensitivityb

specificityb

diary

accelerometry

diary

accelerometry

total time of assessment c

>60 min

50

378

0.72

0.00

476

1,535

0.64

0.08

>90 min

25

81

0.80

0.16

269

420

0.70

0.57

>120 min

18

34

0.83

0.54

199

284

0.76

0.74

>150 min

11

17

0.82

0.78

156

223

0.72

0.79

>180 min

8

10

0.88

0.92

134

192

0.72

0.82

waking d

>60 min

47

75

0.70

0.93

369

531

0.65

0.91

>90 min

22

34

0.76

0.97

214

257

0.70

0.97

>120 min

14

17

0.71

0.99

145

178

0.76

0.98

>150 min

7

8

0.57

0.99

112

132

0.68

0.98

>180 min

3

3

0.67

1.00

91

108

0.68

0.99

sleeping d

>60 min

3

299

1.00

0.59

22

868

0.90

0.77

>90 min

3

44

1.00

0.94

21

130

0.86

0.97

>120 min

3

16

1.00

0.98

21

81

0.86

0.98

>150 min

3

9

1.00

0.99

21

75

0.86

0.98

>180 min

3

6

1.00

1.00

20

66

0.85

0.99

  1. NWT, non-wear time.
  2. aAlgorithms were applied to diary and accelerometry data to check for self-reported NWT periods >60, >90, >120, >150, and >180 minutes or NWT periods based on consecutive acceleration zero-counts >60, >90, >120, >150, and >180 minutes, respectively.
  3. bPeriods with an overlap between NWT according to diary and accelerometry of ≥50% were included in this analysis; number of epochs with an overlap of <50% was: in ActivE, waking, 90-min algorithm: 1; in KORA FF4, total time, 60- to 180-min algorithm: 30, 7, 8, 5, and 4, respectively; in KORA FF4, waking, 60- to 180-min algorithm: 14, 4, 5, 2, and 1, respectively; in KORA FF4, sleeping, 60-min algorithm: 1.
  4. cAnalyses were conducted from the first to the last recorded time point of assessment in the diary.
  5. dWaking and sleeping phases were derived from participants’ diary entries.
  6. bold: algorithms showing high sensitivity and specificity in ActivE and KORA FF4, respectively.
Back to article page