Skip to main content

Table 3 Effects of child characteristics on death outcome upon fitting reduced logit model to MDHS data

From: Impact of ignoring sampling design in the prediction of binary health outcomes through logistic regression: evidence from Malawi demographic and health survey under-five mortality data; 2000-2016

 

Unweighted Reduced Logit Model

Weighted Reduced Logit Model

 

2004 DHS

2010 DHS

2015 DHS

2004 DHS

2010 DHS

2015 DHS

Variable

aOR (p-value)

aOR (p-value)

aOR (p-value)

aOR (p-value)

aOR (p-value)

aOR (p-value)

Child’s Sex

      

      Male\(*\)

      

      Female

0.74 (\(<0.0001\))

0.84 (0.004)

0.81 (0.021)

0.80 (0.014)

0.86 (0.022)

0.80 (0.017)

Birth order

      

      1\(*\)

      

      2-5

-

-

-

-

-

-

      6+

-

-

-

-

-

-

Caesarean birth

      

      No\(*\)

      

      Yes

1.30 (0.282)

1.47 (0.009)

1.56 (0.008)

1.18 (0.631)

1.35 (0.076)

1.56 (0.022)

Delivery type

      

      Singleton \(*\)

      

      Multiple

5.51 (\(<0.0001\))

4.58 (\(<0.0001\))

4.83 (\(<0.0001\))

5.82 (\(<0.0001\))

4.43 (\(<0.0001\))

4.73 (\(<0.0001\))

Birth place

      

      Home or other \(*\)

      

      Health facility

-

-

-

-

-

-

Maternal age

      

      <20 \(*\)

      

      20-34

1.03 (0.932)

1.52 (0.258)

1.20 (0.725)

0.98 (0.953)

1.38 (0.390)

1.23 (0.704)

      35-49

1.21 (0.593)

2.22 (0.033)

1.66 (0.334)

1.18 (0.667)

1.93 (0.076)

1.63 (0.380)

Preceding birth-interval

      

      <24 months \(*\)

      

      24 to 36 months

0.54 (\(<0.0001\))

0.46 (\(<0.0001\))

0.55 (\(<0.0001\))

0.57 (\(<0.0001\))

0.47 (\(<0.0001\))

0.54 (\(<0.0001\))

      > 36 months

0.45 (\(<0.0001\))

0.42 (\(<0.0001\))

0.39 (\(<0.0001\))

0.45 (\(<0.0001\))

0.46 (\(<0.0001\))

0.38 (\(<0.0001\))

Contraceptive use

      

      Not using \(*\)

      

      Using

0.64 (\(<0.0001\))

0.72 (\(<0.0001\))

0.60 (\(<0.0001\))

0.63 (\(<0.0001\))

0.72 (\(<0.0001\))

0.58 (\(<0.0001\))

Maternal educ

      

      No education \(*\)

      

      Primary

0.94 (0.441)

1.11 (0.177)

1.12 (0.369)

0.92 (0.391)

1.08 (0.419)

1.14 (0.337)

      Secondary or above

0.58 (0.007)

0.95 (0.671)

1.43 (\(<0.0001\))

0.69 (0.139)

0.85 (0.250)

0.98 (0.922)

Maternal job status

      

      Not working \(*\)

      

      Working

1.02 (0.773)

1.08 (0.199)

1.43 (\(<0.0001\))

1.03 (0.754)

1.08 (0.279)

1.41 (0.003)

Residence

      

      Urban \(*\)

      

      Rural

-

-

-

-

-

-

Region

      

      Northern \(*\)

      

      Central

1.12 (0.420)

1.25 (0.023)

1.19 (0.190)

1.14 (0.419)

1.37 (0.003)

1.16 (0.271)

      Southern

1.13 (0.356)

1.47 (\(<0.0001\))

1.08 (0.552)

1.24 (0.172)

1.59 (\(<0.0001\))

1.11 (0.438)

AIC

4870.06

8268.38

4285.64

-

-

-

  1. aOR = adjusted odds ratio; \(``*''\) = reference level; \(``-''\) in AIC row implied weighted model output did not include AIC nor log-likelihood value; “Reduced” Logit Model implied a model with some covariates that formed part of the initial model dropped in this second model