MODELO DE COX MULTINIVEL ESTRATIFICADO CON INTERACCIÓN POR ÍNDICE DE RIQUEZA PARA ANÁLISIS DE RIESGO DE MUERTE DE NIÑOS MENORES DE CINCO AÑOS - ENDSA 2008
Date
Journal Title
Journal ISSN
Volume Title
Publisher
Inv. y Des.
Abstract
Este estudio busca determinar en qué medida el estatus socioeconómico familiar, vinculado a un indicador de riqueza, afecta la probabilidad que los niños menores de cinco años tienen de sobrevivir, utilizando la información de la Encuesta Nacional de Demografía y Salud de Bolivia 2008 (ENDSAB-2008). La población objeto de estudio fueron los niños nacidos vivos, de mujeres entre 15 a 49 años, en los cinco años anteriores a la ENDSAB 2008. La variable dependiente en este estudio es la mortalidad de menores de cinco años (5q0), definida como el riesgo de morir entre el nacimiento y el quinto año de vida. Se estimó un modelo multinivel de Cox estratificado con interacción por las categorías de una variable de estrato denominada como "Índice de Riqueza". Se estimaron 3 modelos multinivel, uno por cada estrato (pobre, medio y rico). El modelo estimado para el estrato pobre del Índice de Riqueza logra explicar una proporción mayor de la varianza entre clústeres, más del 60%, determinado por la zona de residencia, características a nivel niño-madre, hogar y contextual-comunitario. Lo contrario ocurre en el modelo estimado para el estrato rico del Índice de riqueza ya que solo logra explicar el 31% de la varianza entre clústeres.
This study seeks to determine to what extent family socioeconomic status, linked to an indicator of wealth, affects the probability that children under five years old have to survive, using the information from the National Survey of Demography and Health of Bolivia 2008 (ENDSAB 2008). The population under study was children born alive, from women between 15 and 49 years old, in the five years prior to ENDSAB 2008. The dependent variable in this study is the mortality of children under five years old (5q0), defined as the risk of dying between birth and the fifth year of life. A stratified multilevel Cox model was estimated with interaction by the categories of a stratum variable called "Wealth Index". Three multilevel models were estimated, one for each stratum (poor, medium and rich). The estimated model for the poor stratum of the Wealth Index manages to explain a greater proportion of the variance between clusters, more than 60%, determined by the area of residence, characteristics at the child-mother, household and community-contextual level. The opposite occurs in the model estimated for the rich stratum of the wealth index since it only manages to explain 31% of the variance between clusters.
This study seeks to determine to what extent family socioeconomic status, linked to an indicator of wealth, affects the probability that children under five years old have to survive, using the information from the National Survey of Demography and Health of Bolivia 2008 (ENDSAB 2008). The population under study was children born alive, from women between 15 and 49 years old, in the five years prior to ENDSAB 2008. The dependent variable in this study is the mortality of children under five years old (5q0), defined as the risk of dying between birth and the fifth year of life. A stratified multilevel Cox model was estimated with interaction by the categories of a stratum variable called "Wealth Index". Three multilevel models were estimated, one for each stratum (poor, medium and rich). The estimated model for the poor stratum of the Wealth Index manages to explain a greater proportion of the variance between clusters, more than 60%, determined by the area of residence, characteristics at the child-mother, household and community-contextual level. The opposite occurs in the model estimated for the rich stratum of the wealth index since it only manages to explain 31% of the variance between clusters.
Description
Vol. 19, No. 2