Skip to main content
. 2021 Mar 30;16(3):e0248678. doi: 10.1371/journal.pone.0248678

Table 4. Multilevel logistic regression analysis of neonatal illness among health facility deliveries in North Gondar zone, Northwest Ethiopia, 2019.

Variables Model I a Model II b Model III c Model IV d
Individual (Neonatal and Maternal) level factors
Number of live children
    ≤ 3 children - 2.03 (1.1, 3.7)* - 1.96 (1.1 3.6)*
    ≥ 4 children 1 1
ANC attendance at current pregnancy
    No - 1.90 (1.1, 3.5)* - 1.83 (1.1, 3.3)*
    Yes 1 1
Type of pregnancy
    Single - 1 - 1
    Twins or triplets 2.75 (1.1, 6.9)* 2.43 (1.1, 6.1)*
Birth length
    Low - 8.86 (4.0, 19.5)** - 7.93 (3.6, 17.3)***
    High 1 1
Breast-feeding quality
    Poor - 2.70 (1.6, 4.5)** - 2.37 (1.4, 4.0)**
    Good 1 1
Health facility level factors
Delivery process quality
    Poor - - 2.34 (1.4, 4.0)* 1.38 (0.8, 2.3)
    Good 1 1
Cord care quality
    Poor - - 1.87 (1.1, 3.2)* 1.42 (0.9, 2.4)
    Good 1 1
Random effects and model comparison
ICC 0.29 0.21 0.27 0.20
Log likelihood -421.69487 -389.99733 -413.1064 -388.29483
AIC 847.3897 796.9947 834.2128 794.5897

a An empty or null model estimating the contextual variation in neonatal morbidity.

b Model simultaneously adjusting for individual-level factors.

c Model simultaneously adjusting for facility-level factors.

d Model simultaneously adjusting for all variables in Model II and Model III.

***p < 0.001, **p < 0.01, and *p < 0.05.