Skip to main content
. 2020 May 23;9(5):1587. doi: 10.3390/jcm9051587

Table A1.

Sensitivity analysis of the model to predict low birthweight: Comparison of the regression coefficients, standard errors (SE), and p-values for complete case analysis (CCA) and multiple imputation (MI) for generalized linear mixed model fit by maximum likelihood. (Number of cluster = 10, Standard deviation = 0.06308).

Predictor Variable * Complete Case Analysis Multiple Imputation
Β SE P-Value β SE P-Value
Age (<20) 1.610 0.3770 <0.001 1.593 0.3700 <0.001
Formal education (no) 0.478 0.4469 0.2863 0.479 0.3246 0.2841
BMI (<18.5) 1.533 0.3100 <0.001 1.516 0.3076 <0.001
Height (<155cm) 1.235 0.3110 <0.001 1.225 0.3038 <0.001
Hemoglobin (<12.0 mg/dl) 1.206 0.3024 <0.001 1.213 0.2998 <0.001
Gravidity (prim-gravida) 0.577 0.3168 0.0600 0.606 0.3068 0.0490
Previous ANC (no) 0.011 0.4644 0.9814 0.011 0.3818 0.9801
Comorbidity § (yes) 1.471 0.6221 0.0248 1.475 0.6043 0.0246

* Variables retained in the reduced model using likelihood ratio test are: age, BMI, hemoglobin, education, gravidity, comorbidity, unplanned pregnancy, and alcohol use. § Comorbidity include pulmonary: history of asthma or COPD (chronic obstructive pulmonary disease); cardiac: history of heart failure or ischemic heart disease; renal diseases. BMI: Body Mass Index; ANC: Antenatal care.