Chief complaint classification with recurrent neural networks

Syndromic surveillance detects and monitors individual and population health indicators through sources such as emergency department records. Automated classification of these records can improve outbreak detection speed and diagnosis accuracy. Current syndromic systems rely on hand-coded keyword-based methods to parse written fields and may benefit from the use of modern supervised-learning classifier models. In this paper, we implement two recurrent neural network models based on long short-term memory (LSTM) and gated recurrent unit (GRU) cells and compare them to two traditional bag-of-words classifiers: multinomial naïve Bayes (MNB) and a support vector machine (SVM). The MNB classifier is one of only two machine learning algorithms currently being used for syndromic surveillance. All four models are trained to predict diagnostic code groups as defined by Clinical Classification Software, first to predict from discharge diagnosis, and then from chief complaint fields. The classifiers are trained on 3.6 million de-identified emergency department records from a single United States jurisdiction. We compare performance of these models primarily using the F₁ score, and we measure absolute model performance to determine which conditions are the most amenable to surveillance based on chief complaint alone. Using discharge diagnoses, the LSTM classifier performs best, though all models exhibit an F₁ score above 96.00. Using chief complaints, the GRU performs best (F₁ = 47.38), and MNB with bigrams performs worst (F₁ = 39.40). We also note that certain syndrome types are easier to detect than others. For example, chief complaints using the GRU model predicts alcohol-related disorders well (F₁ = 78.91) but predicts influenza poorly (F₁ = 14.80). In all instances, the RNN models outperformed the bag-of-words classifiers suggesting deep learning models could substantially improve the automatic classification of unstructured text for syndromic surveillance.