Skip to main content

Table 1 Descriptive statistics of the input and output variables used in neural network-based initial screening model

From: Using neural network as a screening and educational tool for abnormal glucose tolerance in the community

Items

2 hPG&

P value

(two-sided)

 

Abnormal(n = 230)

Normal(n = 3043)

 

Input variables

   

Family history (yes: no, yes %)

21:209, 10%

156:2887, 5.4%

< 0.05*

Age (years, Mean(SD), Min to Max)

49.1(9.0), 125.0 to 73.0

40.6 (10.2), 19.0 to 71.0

< 0.01

Height (cm, Mean(SD), Min to Max)

161.1(8.0), 143.0 to 184.0

164.9 (7.4), 135.0 to 191.0

< 0.01

Weight (kg, Mean(SD), Min to Max)

67.1(11.3), 43.0 to 101.0

63.7 (9.7), 37.5 to 112.0

< 0.01

Waist circumference (cm, Mean(SD),

Min to Max)

85.4(11.2), 57.8 to 115.0

77.6(10.0), 51.0 to 119.3

< 0.01

Hip circumference (cm, Mean(SD),

Min to Max)

92.7(6.9), 75.5 to 111.5

89.4(6.2), 64.5 to 124.8

< 0.01

Output variable

   

2 hPG BG(mmol/L, Mean (SD), Min to Max)

11.9 (5.0), 7.8 to 36.0

5.1 (1.2), 0.6 to 7.7

< 0.01#

  1. & 2-hour plasma glucose after 75-g oral glucose tolerance test, referring to the standard of WHO 1998 to diagnose normal and abnormal glucose tolerance
  2. * Chi-square test
  3. # Wilcoxon rank sum test
  4. all other variables were near-normal distribution and compared by using two-sample t-test.