Estimation of the Distribution of Duration of Breastfeeding from Cross-Sectional Data: Some Methodological Issue

Background: Duration of breastfeeding is an important health indicator of mother and child. There are various indirect epidemiological methods available to estimate the duration of breastfeeding from cross sectional data. Objective: To estimate the distribution of duration of breastfeeding at national level cross sectional data and compare various available technique. The impact of the sampling frame (ascertain of the individual understudy) is also evaluated. Method: National Family Health Survey (NFHS-IV) data is used. Duration of breastfeeding of only those children who were born before 60 months from survey date were included in the study. The technique of Current Status Data, Life Table Analysis, and Kaplan Meier (KM) estimator is applied to assess the distribution of duration of breastfeeding. Result: The mean estimate is 32.84, 33.14 and 33.64 months by Kaplan Maier Estimator, Current Status Data and Life Table Analysis respectively. The Current Status and Life Table method are better than Kaplan Meier Estimator as it is doesn’t based on recall data and heaping present in the data. Conclusion: One must be very cautions while estimating the various epidemiological parameters from cross section data set. The assumptions of the methodology as per data available should be evaluate. If such data is not available, the available methodology may be modified. Regression analysis based on Current Status data technique may be used to assess the impact of various clinical and epidemiological factors (such as nutrition of mother, health status of mother etc.) on duration of breastfeeding.