Improving health care organizational management through neural network learning.

摘要: In order to provide more ethical and objective measures of the likelihood Intensive Care Unit (ICU) recovery, hospitals have turned increasingly decision support system software packages, such as APACHE. However, these packages derive estimates from parametric techniques, Binary Logit Regression (BLR) in APACHE case, require developer specify advance functional relationships among variables model. Recent rapid advancements computer hardware technology encouraged researchers use computationally intensive, non-parametric techniques Neural Networks (NNs), which are purported be better than models terms prediction capabilities. The present study applies both methodologies a sample ICU patients shows that NN technique predicts mortality rates correctly BLR, offers promising alternative hospital settings.