This document summarizes a study that used artificial neural networks and genetic algorithms to predict mortality in critically ill patients. Researchers analyzed data from over 4,000 patients, identifying characteristics like demographics, physiology, and organ failure to train a neural network. An initial set of 20 randomly generated neural networks evolved over generations, with structure optimized by a genetic algorithm. The best-performing 7th generation network used 18 variables to predict mortality with comparable accuracy to logistic regression models. Receiver operating characteristic curves demonstrated the neural network and logistic regression tests' ability to identify patients at risk of death.