Skip to main content

Table 2 Validity indices (accuracy, bias, and precision) averaged across wear location for each speed bout

From: A catalog of validity indices for step counting wearable technologies during treadmill walking: the CADENCE-Kids study

 

Treadmill speed, km/h (mph)

  

Slow

Normal

Fast

0.8 (0.5)

1.6 (1.0)

2.4 (1.5)

3.2 (2.0)

4.0 (2.5)

4.8 (3.0)

5.6 (3.5)

6.4 (4.0)

7.2 (4.5)

8.0 (5.0)

Waist

 MAPE ± SD

92.4 ± 11.2

74.1 ± 25.2

40.7 ± 26.8

16.6 ± 18.3

5.9 ± 12.1

4.1 ± 11.6

2.7 ± 8.9

3.1 ± 9.1

2.6 ± 8.6

2.6 ± 6.1

 MPE ± SD

− 92.4 ± 11.2

−74.1 ± 25.3

− 39.7 ± 28.2

− 15.3 ± 19.4

−5.0 ± 12.5

−2.6 ± 12.0

−1.7 ± 9.2

−1.7 ± 9.4

−0.9 ± 8.9

1.2 ± 6.6

Wrist/Arm

 MAPE ± SD

93.8 ± 9.4

83.6 ± 19.0

68.5 ± 28.1

56.2 ± 34.3

52.4 ± 37.1

50.1 ± 36.8

51.2 ± 36.1

52.2 ± 33.8

54.9 ± 32.0

53.2 ± 28.7

 MPE ± SD

− 93.8 ± 9.4

− 83.6 ± 19.0

− 68.4 ± 28.3

− 56.1 ± 34.5

−52.3 ± 37.2

−49.7 ± 37.3

−51.1 ± 36.2

−52.0 ± 34.1

− 54.6 ± 32.5

− 51.6 ± 31.5

Thigh

 MAPE ± SD

60.2 ± 28.2

10.9 ± 14.7

3.2 ± 8.7

3.5 ± 9.0

3.2 ± 8.9

3.3 ± 9.2

3.7 ± 9.9

6.0 ± 14.7

9.8 ± 23.1

11.1 ± 24.7

 MPE ± SD

− 59.5 ± 29.7

−8.9 ± 16.0

−1.6 ± 9.1

−1.2 ± 9.5

−1.8 ± 9.3

−1.7 ± 9.7

− 2.1 ± 10.4

−5.0 ± 15.1

−8.5 ± 23.6

−7.2 ± 26.2

Ankle

 MAPE ± SD

14.9 ± 19.8

4.4 ± 13.1

3.4 ± 13.2

3.6 ± 13.8

3.1 ± 13.4

2.9 ± 10.4

4.1 ± 12.0

8.7 ± 15.7

20.7 ± 18.0

23.5 ± 7.9

 MPE ± SD

− 14.6 ± 20.0

−0.6 ± 13.8

−1.5 ± 13.5

−1.3 ± 14.2

−1.8 ± 13.6

−1.1 ± 10.8

−3.0 ± 12.4

−8.0 ± 16.1

−20.4 ± 18.4

−23.5 ± 7.9

  1. All mean absolute percentage error (MAPE) and mean percentage error (MPE) values are presented as mean percentage ± standard deviation (SD). Waist devices: Actical, ActiGraph GT3X+, NL-1000, Digi-Walker SW-200. Non-dominant wrist device: ActiGraph GT3X+; Arm device: SenseWear. Thigh device: activPAL. Ankle device: StepWatch. MAPE and SD values closer to 0% indicate higher accuracy and precision, respectively. MPE values closer to 0% indicate lower bias and negative values mean undercounting