Skip to main content

Table 3 Multilevel regression results of individual/household and community-level factors associated with health facility delivery

From: What predicts health facility delivery among women? analysis from the 2021 Madagascar Demographic and Health Survey

Variables

Model 0

Model 1

aOR (95% CI)

Model 2

aOR (95% CI)

Model 3

aOR (95% CI)

Fixed effect

 Individual/household-level variables

  Age

    

   15–19

 

Ref

 

Ref

   20–24

 

0.89 (0.71–1.12)

 

0.88 (0.70–1.10)

   25–29

 

1.21 (0.94–1.55)

 

1.19 (0.92–1.52)

   30–34

 

1.58** (1.18–2.10)

 

1.53** (1.14–2.04)

   35–39

 

1.80** (1.29–2.52)

 

1.73** (1.24–2.42)

   40–44

 

1.51* (1.03–2.20)

 

1.44 (0.98–2.11)

   45–49

 

2.13** (1.33–3.42)

 

2.14** (1.34–3.43)

  Education

   No education

 

Ref

 

Ref

   Primary

 

1.19 (0.98–1.44)

 

1.15 (0.95–1.39)

   Secondary/higher

 

1.72*** (1.38–2.13)

 

1.62*** (1.30–2.01)

 Marital Status

   Never in union

 

Ref

 

Ref

   Married

 

1.33** (1.07–1.65)

 

1.32** (1.06–1.64)

   Cohabitation

 

1.28 (0.96–1.71)

 

1.28 (0.96–1.70)

   Widowed

 

2.34*** (1.49–3.67)

 

2.25*** (1.43–3.55)

   Divorced

 

1.37* (1.05–1.80)

 

1.34* (1.02–1.76)

  Employment status

   Not working

 

Ref

 

Ref

   Working

 

0.84 (0.69–1.04)

 

0.86 (0.70–1.06)

  Parity

   One birth

 

Ref

 

Ref

   Two births

 

0.61** (0.50–0.75)

 

0.61** (0.50–0.75)

   Three births

 

0.48*** (0.37–0.62)

 

0.49*** (0.38–0.63)

   Four or more births

 

0.37*** (0.28–0.49)

 

0.38*** (0.29–0.51)

  Mass media

   No

 

Ref

 

Ref

   Yes

 

1.21* (1.03–1.42)

 

1.18* (1.00-1.39)

  ANC visit

   No

 

Ref

 

Ref

   1 or more

 

7.07*** (5.02–9.96)

 

6.95*** (4.95–9.77)

  Health insurance coverage

   No

 

Ref

 

Ref

   Yes

 

1.67 (0.98–2.84)

 

1.55 (0.91–2.63)

  Religion

   Christianity

 

Ref

 

Ref

   Islam

 

0.94 (0.53–1.68)

 

0.92 (0.52–1.64)

   Traditional

 

0.54** (0.34–0.85)

 

0.56* (0.35–0.89)

   No/other

 

0.68*** (0.55–0.83)

 

0.73** (0.59–0.90)

  Terminated Pregnancy

   No

 

Ref

 

Ref

   Yes

 

0.97 (0.81–1.15)

 

0.98 (0.82–1.16)

  Wealth index

   Poorest

 

Ref

 

Ref

   Poorer

 

1.62*** (1.32–1.99)

 

1.53*** (1.25–1.87)

   Middle

 

1.72*** (1.37–2.15)

 

1.53*** (1.22–1.92)

   Richer

 

2.50*** (1.95–3.20)

 

2.01*** (1.55–2.60)

   Richest

 

3.97*** (3.00-5.26)

 

2.74*** (1.99–3.77)

Community-level variables

 Problem with permission to go to the health facility

  Big problem

  

Ref

Ref

  Not a big problem

  

0.94 (0.78–1.13)

0.94 (0.78–1.13)

 Problem with money needed for treatment

  Big problem

  

Ref

Ref

  Not a big problem

  

1.27*** (1.11–1.45)

1.05 (0.91–1.21)

 Problem with distance to health facility

  Big problem

  

Ref

Ref

  Not a big problem

  

1.22* (1.04–1.43)

1.28** (1.09–1.50)

 Problem with not wanting to go to the health facility alone

  Big problem

  

Ref

Ref

  Not a big problem

  

1.11 (0.94–1.31)

1.10 (0.92–1.32)

 Place of residence

    

  Urban

  

Ref

Ref

  Rural

  

0.86 (0.66–1.14)

1.17 (0.88–1.56)

 Community literacy level

  Low

  

Ref

Ref

  Medium

  

1.86** (1.44–2.41)

1.16 (0.89–1.50)

  High

  

3.18*** (2.38–4.24)

1.54** (1.15–2.08)

 Community socioeconomic status

  Low

  

Ref

Ref

  Moderate

  

1.26 (0.87–1.82)

0.98 (0.68–1.40)

  High

  

3.13*** (2.35–4.16)

1.72*** (1.28–2.31)

 Random effect results

  PSU variance (95% CI)

2.14 (1.83–2.50)

0.97 (0.80–1.18)

1.01 (0.83–1.22)

0.89 (0.73–1.09)

  ICC

39.4

22.8

23.5

21.3

  Wald chi-square

Ref

721.85***

418.31***

920.90***

 Model fitness

  AIC

10,769.68

9,760.38

10,333.45

9,709.52

  BIC

10,783.96

9,967.43

10,411.98

9,980.82

N

9,315

9,315

9,315

9,315

Number of groups

649

649

649

649

  1. Source: 2021 MDHS
  2. Ref-Reference category, p < 0.05, **p < 0.01, ***p < 0.001, aOR adjusted odds ratio
  3. PSU Primary Sampling Unit, ICC Intra-Class Correlation, aOR adjusted odds ratio, AIC Akaike’s Information Criterion, BIC Bayesian Information Criterion
  4. *p < 0.05, **p < 0.01, ***p < 0.001