Skip to main content

Table 5 Statistical analysis of characteristics selected by FNN

From: Combinational risk factors of metabolic syndrome identified by fuzzy neural network analysis of health-check data

Study

Characteristic

model 1

model 2b

model 3c

model 4d

  

OR (95%CI)

P value

OR (95%CI)

P value

OR (95%CI)

P value

OR (95%CI)

P value

Original Study

γ-GTP (doubling)a

4.71 (2.63–8.41)

1.69 × 10−7

5.98 (3.12–11.5)

7.25 × 10−8

4.06 (1.33–12.4)

0.014

  
 

WBC (1000 cells/μL)

1.83 (1.39–2.41)

1.62 × 10−5

1.94 (1.41–2.65)

3.86 × 10−5

2.69 (1.44–5.02)

0.002

  

Replication Study

γ-GTP (doubling)a

2.64 (2.45–2.85)

<1.0 × 10−99

2.84 (2.62–3.08)

<1.0 × 10−99

1.32 (1.17–1.51)

1.71 × 10−5

1.33 (1.17–1.51)

1.69 × 10−5

 

WBC (1000 cells/μL)

1.31 (1.26–1.35)

2.73 × 10−51

1.30 (1.25–1.35)

5.06 × 10−42

1.05 (0.99–1.12)

0.107

1.06 (0.99–1.12)

0.094

Combined Studye

γ-GTP (doubling)a

2.65 (2.46–2.86)

<1.0 × 10−99

2.86 (2.64–3.10)

<1.0 × 10−99

1.35 (1.19–1.53)

3.11 × 10−6

  
 

WBC (1000 cells/μL)

1.32 (1.27–1.36)

1.28 × 10−55

1.31 (1.26–1.36)

1.62 × 10−45

1.07 (1.01–1.14)

0.031

  
  1. Odds ratio (OR) with 95 % confidence interval (CI) indicates the proportional change in risk associated with each increase by the amount indicated in parentheses.
  2. ageometric mean (2.5–97.5 %) was used because of skewed distriburions.
  3. Differences in characteristics between case and healthy control subjects were evaluated by logistic regression analysis.
  4. bDifferences were evaluated by logistic regression analysis with adjustment for age, drinking habit and smoking habit.
  5. cDifferences were evaluated by logistic regression analysis with adjustment for age, drinking habit, smoking habit and the components of MetS.
  6. dDifferences were evaluated by logistic regression analysis with adjustment for age, drinking habit, smoking habit, the components of MetS and exercise habit.
  7. eIn combined study, the difference of study was added in adjustment.