Regresar

Application of control charts to detect anomalies in child mortality in Ecuador

Abstract:

In this work, control charts have been designed in order to detect anomalies in the infant mortality data of the urban area in Ecuador. In order to design these charts, negative binomial regression models were constructed for the annual mortality rate of boys and girls aged between one and less than five years. With the constructed models, two types of control charts were prepared: the first one, an EWMA (Exponentially Weighted Moving Average) version of the deviance residues of the constructed models and the second, a chart that monitors the average number of deaths through the estimate obtained by the regression model. For both charts, an approximate 95 per cent confidence level has been considered. The EWMA chart was more efficient in detecting the sub-records that usually occur in the deaths of urban children in Ecuador.