Methodological approaches to reducing misclassification bias in exclusive breastfeeding measurement.

Exclusive breastfeeding is a key pillar of healthy nutrition in infancy. While the importance of breastfeeding is undisputed, measuring early breastfeeding behavior remains challenging. This study evaluated the impact of addressing missingness in feeding variables, incorporating prelacteal feeding, reducing sampling bias through age weighting, and adjusting measurement periods on estimated exclusive breastfeeding prevalence. We show that imputing missing values reduces estimated prevalence from 42.9% (95% CI, 40.3-45.6) to 41.2% (95% CI, 38.5-44.0). Including prelacteal feeding in the definition further decreases estimates to 32.2% (95% CI, 29.5-34.8), while applying age weights slightly increased the prevalence to 33.6% (95% CI, 31.0-36.3). Measuring exclusive breastfeeding compliance at 6 months, rather than looking at average feeding behavior in the first six months of an infant's life, yields an estimated exclusive breastfeeding prevalence of 5.9% (95% CI, 4.6-7.6). Given the major differences in exclusive breastfeeding prevalence estimates from different methodological adjustments, a standardized exclusive breastfeeding measurement is urgently needed to accurately assess early-life nutrition and guide policy based on reliable and comparable estimates.