Table 4 Multilevel logistic regression analysis for overweight &/or obesity (referent: normal weight) (N = 12,073).

From: Application of single-level and multi-level modeling approach to examine geographic and socioeconomic variation in underweight, overweight and obesity in Nepal: findings from NDHS 2016

Variables

model 1

model 2

model 3

model 4

 

OR

LCL

UCL

OR

LCL

UCL

OR

LCL

UCL

Intercept

−1.52(0.18)

−2.612(0.20)

  

−2.93(0.21)

  

−3.525(0.19)

  

Age(years)

          

15–25

 

ref

  

ref

  

ref

  

25–35

 

4.29

3.67

5.01

4.147

3.5

4.86

2.81

2.35

3.35

35–45

 

5.83

4.97

6.83

5.872

5

6.9

4.05

3.35

4.90

45–55

 

5.05

4.27

5.98

5.136

4.3

6.1

3.85

3.13

4.74

55–65

 

3.74

3.10

4.50

3.847

3.2

4.65

3.11

2.47

3.92

>65

 

2.44

1.97

3.02

2.571

2.1

3.2

2.32

1.77

3.04

Sex

Male

 

0.70

0.64

0.77

0.681

0.6

0.75

0.62

0.55

0.69

Female

          

Education

No education, preschool

    

ref

  

ref

  

Primary

       

1.45

1.25

1.68

Secondary

       

1.55

1.33

1.79

Higher

       

1.56

1.30

1.88

Marital status

Formerly/ever married

       

0.82

0.67

1.00

Never married

       

0.34

0.27

0.42

Currently married

       

ref

  

Residency

Urban

       

1.66

1.45

1.91

Rural

       

ref

  

Wealth quintile

Poorest

    

ref

  

ref

  

2

    

2.478

2.1

2.98

2.37

1.97

2.86

3

    

1.614

1.3

1.94

1.61

1.34

1.94

4

    

3.904

3.2

4.74

3.67

3.02

4.46

Richest

    

6.263

5.2

7.6

6.08

4.97

7.45

Error variance

Level 2 intercept*

0.24(0.15)

0.23(0.14)

  

0.16(0.10)

   

0.14(0.084)

 

Level 3 intercept*

0.26(0.06)

0.38(0.069)

  

0.19(0.04)

   

0.10(0.029)

 

ICC-level 2

6.80

6.50

  

4.60

   

4.10

 

ICC-level 3

7.30

10.40

  

5.50

   

2.90

 

PCV-level 2

 

4.4%

  

32.4%

   

39.7%

 

PCV-level 3

 

42.5%

  

24.7%

   

60.3%

 

AIC

12306.2

11491.7

  

11060.6

   

10868.3

 

SBIC

12306.0

11473.7

  

11034.6

   

10830.3

 
  1. Model 1: empty model, model 2: adjusted for age and sex, model 3: model 2 plus wealth quintile, and model 4: model 3 plus education, marital status, and residency. The level two intercepts are for provinces and level 3 are for the districts. Abbreviation: AIC: akaike information criterion, LCL: lower conflidence limit, SBIC: schwarz bayesian information criterion, UCL: upper confidence limit.
Back to article page