Table 1 Participant characteristics according to data set.

From: Comparisons of the prediction models for undiagnosed diabetes between machine learning versus traditional statistical methods

 

Training & internal validation set

External validation set

n = 23,369 (71.2%)

n = 9,458 (28.8%)

Non-diabetes

Undiagnosed diabetes

p-value

Non-diabetes

Undiagnosed diabetes

p-value

n = 22,361

n = 1,008

n = 8,972

n = 486

Age, yr

48.0 (15.8)

57.2 (12.3)*

 < 0.001

47.9 (15.9)

57.2 (12.7)*

 < 0.001

Height, cm

163.4 (9.1)

163.0 (9.5)

0.187

164.46 (9.2)

163.75 (9.3)

0.095

Weight, kg

63.5 (12.2)

70.0 (13.6)*

 < 0.001

64.8 (13.0)

71.8 (15.3)*

 < 0.001

BMI, kg/m2

23.7 (3.5)

26.2 (3.9)*

 < 0.001

23.8 (3.6)

26.6 (4.2)*

 < 0.001

WC, cm

80.7 (9.9)

89.0 (9.6)*

 < 0.001

83.1 (10.3)

92.0 (10.2)*

 < 0.001

WHtR

0.5 (0.1)

0.5 (0.1)*

 < 0.001

0.5 (0.1)

0. 6 (0.1)*

 < 0.001

RHR, bpm

69.4 (9.4)

72.1 (10.8)*

 < 0.001

69.5 (9.6)

72.1 (10. 8)*

 < 0.001

SBP, mmHg

116.8 (16.1)

126.3 (16.6)*

 < 0.001

117.5 (15.8)

126.3 (15.6)*

 < 0.001

DBP, mmHg

75.4 (10.0)

79.3 (10.9)*

 < 0.001

76.0 (9.7)

79.7 (9.9)*

 < 0.001

Sleep time, (hour/day)

7.1 (1.3)

7.0 (1.4)*

 < 0.001

7.0 (1.3)

6.8 (1.3)*

 < 0.001

Physical activity (MET-min/week)

      

 Work physical activity

55.8 (275.3)

62.73 (369.3)

0.441

96.6 (666.8)

64.16 (519.0)

0.291

 Leisure physical activity

336.5 (818.2)

260.2 (738.8)*

0.004

326.9 (744.2)

239.9 (680.2)*

0.012

 Walking

832.3 (1184.6)

844.3 (1366.2)

0.754

777.9 (938.6)

802.8 (998.1)

0.570

 Total Physical activity

1714.3 (1985.7)

1643.3 (2317.2)

0.271

1614.3 (1803.3)

1603.1 (1861.0)

0.894

Sex

  

 < 0.001

  

 < 0.001

 Men, n (%)

9,328 (41.7)

550 (54.6)

 

3,904 (43.5)

266 (54.7)

 

 Women, n (%)

13,033 (58.3)

458 (45.4)

 

5,068 (56.5)

220 (45.3)

 

 Family history of diabetes, n (%)

4,719 (21.1)

330 (32.7)

 < 0.001

1978 (22.0)

179 (36.8)

 < 0.001

Alcohol consumption (drinks/day),

  

 < 0.001

 

 < 0.001

 < 1

17,996 (80.5)

738 (73.2)

 

7,318 (81.6)

366 (75.3)

 

 1–4.9

3,606 (16.1)

200 (19.8)

 

1,371 (15.3)

96 (19.8)

 

 ≥ 5

759 (3.4)

70 (6.9)

 

283 (3.2)

24 (4.9)

 

Smoking, n (%)

8,260 (19.5)

425 (23.3)

 < 0.001

1,639 (17.4)

119 (23.2)

0.003

Hypertension, n (%)

14,720 (34.7)

1,025 (56.2)

 < 0.001

2,492 (26.5)

275 (53.5)

 < 0.001

  1. Data are presented as mean (standard deviation) or number (%), All variables were tested by the T-test or chi-square test. Significant differences were found between non-diabetes, undiagnosed diabetes (p < 0.05), *Significantly different from non-diabetes. BMI = Body mass index, WC = Waist circumference, WHtR = Waist to Height Ratio, RHR = Resting heart rate, SBP = Systolic blood pressure, DBP = diastolic blood pressure, Total physical activity = Work physical activity + Leisure physical activity + Walking.
Back to article page