Generalized linear models to forecast infant anemia with factors relates

Abstract

Anemia is a worldwide problem. In Peru, the 2013, about 34% of children of 6-59 months old have anemia. The purpose of this work was to evaluate generalized linear models to forecast infant anemia with factors relates. The study employ the results of ENDES 2013 conducted by the INEI, includes 8983 children with ages in the range of study. INEI only included surveys with complete information. The program MINITAB 17 was used for statistical analysis. The response variables were the prevalence and severity of anemia. In the analysis of binary logistic regression, the prevalence of anemia was associated with the factors: area of residence, natural region, age, sex, birth order, birth period and education level of the motherFor every factor was determined the categories that increase or decreased the prevalence of the infant anemia, and in all cases the estimated model was adequated. The ordinal regression analysis for anemia severity showed association with the factors in study, and except the children age and the mother education level, the estimated models were adequate. The study shows that the techniques of generalized linear models may be employed to forecast adequally the prevalence and the severity infantil anemia.