Fuzzy model to estimate the number of hospitalizations for asthma and pneumonia under the effects of air pollution

Abstract

OBJECTIVE: Predict the number of hospitalizations for asthma and pneumonia associated with exposure to air pollutants in the city of Sao Jose dos Campos, Sao Paulo State. METHODS: This is a computational model using fuzzy logic based on Mamdani's inference method. For the fuzzification of the input variables of particulate matter, ozone, sulfur dioxide and apparent temperature, we considered two relevancy functions for each variable with the linguistic approach: good and bad. For the output variable number of hospitalizations for asthma and pneumonia, we considered five relevancy functions: very low, low, medium, high and very high. DATASUS was our source for the number of hospitalizations in the year 2007 and the result provided by the model was correlated with the actual data of hospitalization with lag from zero to two days. The accuracy of the model was estimated by the ROC curve for each pollutant and in those lags. RESULTS: In the year of 2007, 1,710 hospitalizations by pneumonia and asthma were recorded in Sao Jose dos Campos, State of Sao Paulo, with a daily average of 4.9 hospitalizations (SD = 2.9). The model output data showed positive and significant correlation (r = 0.38) with the actual data; the accuracies evaluated for the model were higher for sulfur dioxide in lag 0 and 2 and for particulate matter in lag 1. CONCLUSIONS: Fuzzy modeling proved accurate for the pollutant exposure effects and hospitalization for pneumonia and asthma approach.