Predicting ambulance diversion.

摘要

This research effort is undertaken with the objective of determining the probability of a hospital going on diversion. This is achieved by developing and evaluating various causal models, using methods such as logistic regression and the Markov property. These models are developed using real data from 29 hospitals and the EMS system of Kansas City Missouri, for a period of two years. This dissertation presents results obtained with the data, and outlines the methods used and scope for future research.; A program was developed using 'R', a Statistical software, to analyze the unique form of the data obtained. Logistic regression is used for analysis. Once the correlation between 911 calls and diversion was established, other influential factors were taken into consideration and a model was developed whereby a region could predict the likelihood for diversion to occur in hospitals of that region.; It was found that 911 calls can be used as a predictor to determine the likelihood of a hospital going on diversion. Also, the current state of diversion that a hospital is in, is a major factor in determining the status of diversion in the future. Other influential factors were also discovered.