Factors associated with low birth weight in Nepal using multiple imputation.

Area of Science:

• Maternal and child health
• Biostatistics
• Demography

Background:

• Birth weight data in low-income countries often suffer from missing values, impacting analysis and potentially leading to underestimation of low birth weight (LBW) prevalence.
• Existing studies acknowledge but often exclude missing birth weight data and fail to address missing determinants, hindering a comprehensive understanding of LBW.

Purpose of the Study:

• To identify determinants associated with low birth weight (LBW) in Nepal.
• To address the challenge of missing data in birth weight and its determinants using multiple imputation techniques.

Main Methods:

• Utilized the Nepal Demographic and Health Survey (NDHS) 2011 child dataset (n=5,240).
• Applied a transform-then impute method to handle missing data (87% of records had missing variables, 21% had no recorded birth weight).
• Employed survey logistic regression on imputed datasets, accounting for survey design and sampling methods.

Main Results:

• The adjusted prevalence of LBW was 15.4% after imputation.
• Increased maternal autonomy in health decisions was associated with a lower likelihood of LBW (OR 1.87 for highest autonomy vs. husband/others; OR 1.57 for joint decisions vs. husband/others).
• Use of highly polluting cooking fuels was associated with a higher likelihood of LBW (OR 1.49).

Conclusions:

• Ignoring missing data in birth weight analyses leads to underestimation of LBW prevalence.
• Maternal autonomy and cooking fuel type are significant determinants of LBW, highlighting potential intervention points.