879. Artificial Neural Networks to Predict Infection in the Surgical Site in Patients over 70 Years Old

Abstract

BackgroundBetween July 2016 and June 2018, a survey was carried out in five hospitals on surgical site infection (SSI) in patients over 70 years old, who underwent surgery procedures, in the city of Belo Horizonte, a city with more of 3,000,000 inhabitants. The general objective is to statistically evaluate such incidences and enable an analysis of the predictive power of SSI, through MLP (Multilayer Perceptron) pattern recognition algorithms.MethodsThrough the Hospital Infection Control Committees (CCIH) of the hospitals involved in the research, data collection on SSI was carried out. Such data is used in the analysis during your routine SSI surveillance procedures. Thus, three procedures were performed: a treatment of the database collected for use of intact samples; a statistical analysis on the profile of the collected hospitals and; an evaluation of the predictive power of five types of MLPs (Backpropagation Standard, Momentum, Resilient Propagation, Weight Decay and Quick Propagation) for SSI prediction. The MLPs were tested with 3, 5, 7 and 10 neurons in the hidden layer and with a division of the database for the resampling process (65% or 75% for testing, 35% or 25% for validation). They were compared by measuring the AUC (Area Under the Curve - ranging from 0 to 1) for each of the configurations.ResultsFrom 11277 records, 3350 were complete for analysis. It was found that: the average age is 79 years (from 74 to 84 years); the average surgery time is 123 minutes; the average hospital stay is 58 days (with a maximum of 114 days), the death rate reached 7.1% and that of SSI 2.59%. A maximum prediction power of 0.642 was found.ConclusionThere was a loss of almost 70% of the database samples due to the presence of noise, however it was possible to evaluate the hospitals profile. The predictive process, presented configurations with results that reached 0.642, what promises the use of the structure for the monitoring of automated SSI for patients over 70 years submitted to surgeries. To optimize data collection, enable other hospitals to use the prediction tool and minimize noise from the database, two mobile application were developed: one for monitoring the patient in the hospital and another for monitoring after hospital discharge. The SSI prediction analysis tool is available at www.nois.org.br.Disclosures All Authors: No reported disclosures