Design and Development of a Triage System in Predicting Patient Disposition using Artificial Neural Networks

Medical Decision is based on acknowledging complex patterns of the patients signs and symptoms. Neural Networks have proven effective in this type of pattern recognition. In this study, a neural network is used to foretell which of the patients seen in an emergency room need to be admitted, transferred to a specialty care or discharged. A multilayer feedforward network model maps input datasets to a corresponding output. The complexity of multilayer feedforward can be altered by changing the number of layers and the number of nodes in each layer. It has been shown that the multilayered neural network can estimate virtually any function to any desired accuracy with the given hidden nodes and enough data. Feedforward describes how the neural network processes the pattern and remembers the patterns. Backpropagation describes how this type of neural network is trained. The algorithms backpropagation and feedforward have been used together here to compare the accuracy of patient disposition in triage. The current study investigates the performance of the back propagation algorithm to train Feed Forward Neural Network to produce better results in the area of patient triage disposition.