Skip to main content

Table 4 Step 1 experimental results

From: V-Model: a new perspective for EHR-based phenotyping

Evaluation item

n

LifeLines (N = 40)

V-Model (N = 40)

Statistical analysis

n_c (accuracy)

RT (sec.) median, IQR(25–75)

n_c (accuracy)

RT (sec.) median, IQR (25–75)

chi-squared test (p-value)

MWU test (p-value)

Representation

       

[P1] P-A connection

140

71 (50.71)

43.32 (33.08-61.91)

115 (82.14)

35.31 (24.99-50.24)

<0.000

<0.000

[P2] non-explicitness

40

-

-

36 (90.00)

-

-

-

[P3] proximity

80

-

-

74 (92.50)

-

-

-

Reasoning

       

[R1] P precedes A

40

36 (90.00)

35.48 (25.23-47.77)

40 (100.00)

14.9 (12.15-18.74)

0.116

<0.000

[R2] qualitative relation

80

76 (95.00)

7.6 (5.16-11.97)

69 (86.25)

9.95 (6.23-13.84)

0.058

0.036

[R3] implicit distance

40

-

-

40 (100.00)

-

-

-

Visualization

       

[V1] intuitive P-A relation

140

104 (74.29)

32.17 (21.46-54.07)

125 (89.29)

19.48 (15-

29.58)

<0.000

<0.000

 

40

25 (62.50)

56.82 (44.92-80.35)

21 (52.50)

45.72 (33.93-62.79)

0.366

0.006

[V5] long history

40

22 (55.00)

92.7 (69.64-134.48)

38 (95.00)

35.23 (25.84-40.46)

<0.000

<0.000

Evaluation item

Group

LifeLines (%)

V-Model (%)

[P04] uneven granularity

V-Model participants

3.75

96.25

(selection)

LifeLines participants

7.5

92.5

 

mean

5.63

94.38

[V04] dynamic scale

V-Model participants

5

95

(preference)

LifeLines participants

25

75

 

mean

15

85

  1. N, number of data; N, number of participants; n_c, number of correct answers.