Utilidad de procalcitonina para el diagnóstico de neumonía adquirida en la comunidad bacteriana en pacientes de 1 mes a 5 años ingresados en el Hospital Roberto Gilbert Elizalde en el periodo enero 2015 - julio 2016.

Abstract

Pneumonia is a frequent disease in children <5 years of age that has a high impact and morbidity and mortality, considered a health problem associated with this. We have the difficulty that the doctor faces in order to differentiate between bacterial and viral pneumonias. Objectives.- Design a problem tree based on procalcitonin to guide a diagnosis of bacterial pneumonia Methods.-Predictive level study that used supervised data mining and sorting tree using WEKA software, performed with 316 patients 1 month to 5 years admitted to "Hospital Roberto Gilbert Elizalde" during January 2015 to July 2016. Results.-60% of patients were male, the average age was 2.2years, 56% with leukocyte values higher than 12500 and 59.2% with neutrophilia higher than 65%, the average of procalcitonin 3.44 ug / dland the cutoff point for prediction of bacterial pneumonia was 1.03 ug / ml,the chest radiograph is the most useful test to differentiate pneumonia etiology, the decision tree for diagnosis of bacterial pneumonia is given by the procalcitonin value. Conclusion.- The application of the classification tree elaborated in this study allows us to have an accuracy of 91.77% for the diagnosis of bacterial pneumonia and to make the decision of the use or not of antibiotics.