Skip to main content

Table 3 Reduced model: two-part model effects between fasting insulin and CRP among adults without diabetes

From: Association between fasting insulin and C-reactive protein among adults without diabetes using a two-part model: NHANES 2005–2010

Covariate

Logit model

GLM model

Combine two-part model

Predictive margins

OR

p-value

β

p-value

Predicted CRP mg/L (CI)

Age, years

1.05

<.001

0.02

.013

 

 20–39

    

3.56 (3.26, 3.85)

 40–59

    

4.07 (3.72, 4.42)

 60–85

    

4.62 (4.03, 5.22)

Sex

 Male

Reference

   

3.78 (3.31, 4.25)

 Female

0.83

.373

0.73

.001

4.22 (3.90, 4.53)

Race

     

 Non-Hispanic White

Reference

   

3.92 (3.57, 4.27)

 Mexican/other Hispanic

2.29

.013

0.06

.851

4.41 (3.80, 5.03)

 Non-Hispanic Black

0.75

.204

0.41

.149

4.72 (4.23, 5.21)

 Other race/multi

0.51

.037

− 0.81

<.001

2.96 (2.51, 3.41)

Poverty-income ratio

1.02

.673

− 0.08

.168

 

 < 130% FPR

    

4.78 (4.35, 5.22)

130–349% FPR

    

4.18 (3.83, 4.53)

 >=350% FPR

    

3.51 (3.17, 3.86)

Smoking

 Non/former smoker

Reference

   

3.83 (3.46, 4.21)

 Current smoker

1.26

.280

1.26

<.001

4.62 (4.08, 5.17)

Physical activity

 No

Reference

   

5.19 (4.66, 5.72)

 Yes

1.08

.757

− 0.86

.008

3.55 (3.21, 3.89)

Fasting Insulin, pmol/L

1.02

<.001

0.03

<.001

 

 < 60

    

2.86 (2.66, 3.06)

 60–119

    

4.39 (3.98, 4.80)

 ≥ 120

    

7.58 (6.37, 8.79)

  1. Mexican/other Hispanic: Mexican or other Hispanic ethnicity; Other race/multi: other race or Multiracial; CRP: C-reactive protein; OR: odds ratio; CI: confidence interval; FPR: federal poverty ratio
  2. Reduced Model: controlling for age, sex, race, socioeconomic status (poverty-income ratio), smoking status, physical activity, and fasting insulin
  3. Two-part regression model used; reference group is the reference group used for each categorical variable
  4. Each level of categorical variable used to provide a predicted CRP based on the model
  5. National estimates based on complex survey design