ESTIMACIÓN DE LA PREVALENCIA DE ANEMIA EN ÁREAS PEQUEÑAS: UN ENFOQUE MULTINIVEL

Abstract

Small area estimation seeks to remedy the problem of many surveys designed at the regional level requiring information about more local areas. Unfortunately, the sample size of such regional surveys is usually too small for direct estimation on sub -regional level. This is just the case of a health survey in the eastern region of Cuba to determine the prevalence of childhood anemia. Given this scenario, this paper proposes a strategy to generate estimates at the municipal level, using an approach based on multilevel models. In the study, the influence of individual and contextual characteristics on the prevalence of anemia is evaluated and the risk of acquiring the disease in different strata of the population under two years old is estimated. The spatial variance among different risk groups demonstrates the importance of location on individual characteristics. A graphic diagnosis of the model residuals allows to identifying vulnerable areas where implementing priority programs of food and health must be achieved by optimizing the use of limited resources.