Pooled logistic regression to describe characteristics associated with wasting and recovery

  • Empirical Model

A pooled logistic regression approach is being used to estimate risk ratios (or incidence rate ratios), adjusting for all potential confounders, within each study. Pooled estimates across individual studies will be via random effects meta-analysis.


Incident wasting and recovery from wasting among children who are wasted.

Predictors (other than time)

Sex, gestational age at birth, birthweight, WHZ at birth, LAZ at birth, birth order, calendar month of birth, delivery location (hospital), height, BMI, age, education, marital status, breastfeeding practices, asset-based wealth index, number of rooms, total number of persons, number of children < 5 y, animals, flooring (improved vs. unimproved), sanitation facility (JMP definitions), safe water source (JMP definitions), treats drinking water, soap present in the home.