Table 5 Data quality results for the outpatient cohort data, presented as the aggregated scores for all participants of the respective cohort.

From: Data quality evaluation in wearable monitoring

KCL

N patients = 14*

Sum h:m:s = 11,209:23:59

 

Completeness (%)

On-body (%)

EDA† (%)

BVP† (%)

TEMP† (%)

Duration h:m:s

Mean

30.1

93.9

88.0

62.3

99.3

800:40:17

Median

26.0

98.4

88.1

63.2

99.9

683:50:17

Std

20.0

15.7

7.8

11.4

1.8

799:28:26

Min

6.3

40.2

72.9

44.1

93.3

57:07:04

Max

65.8

100.0

99.0

77.9

99.9

2977:36:48

MCR

N patients = 14

Sum h:m:s = 65,394:52:36

 

Completeness (%)

On-body (%)

EDA† (%)

BVP† (%)

TEMP† (%)

Duration h:m:s

Mean

76.9

97.8

76.0

53.9

97.6

4671:03:45

Median

85.0

99.0

77.6

54.1

99.3

4933:49:38

Std

22.7

3.8

13.0

17.6

5.1

2663:24:57

Min

18.5

85.2

52.9

4.0

80.3

270:09:58

Max

97.4

100.0

95.6

71.8

100.0

8097:26:03

UKF

N patients = 11*

Sum h:m:s = 14,172:02:42

 

Completeness (%)

On-body (%)

EDA† (%)

BVP† (%)

TEMP† (%)

Duration h:m:s

Mean

34.1

96.5

78.7

55.1

99.3

1288:22:03

Median

27.1

98.6

83.7

53.1

99.7

1137:18:34

Std

22.3

4.6

18.2

22.2

0.7

939:58:26

Min

7.3

85.3

48.3

6.0

98.1

312:01:18

Max

68.3

100.0

95.6

80.5

100.0

3000:34:58

  1. We report the amount of data recorded between start and end of the recording as completeness. From this we test how long the device is worn on the body, and score the signal quality of the modalities EDA, BVP, and TEMP. Additionally we report the duration of the recorded data in h:m:s = hours:minutes:seconds.
  2. *Includes data from dropped-out participants, and those with deliberately shortened follow-up.
  3. †Signal quality scores are in relation to data periods estimated as on-body in the previous step.
Back to article page