Abnormal diagnosis of Emergency Department triage explored with data mining technology: An Emergency Department at a Medical Center in Taiwan taken as an example

Triage helps to classify patients at emergency departments to make the most effective use of resources distributed. What is more important is that accuracy in carrying out triage matters greatly in terms of medical quality, patient satisfaction and life security. As the numbers of patients in emergency departments increase, learning from the examples of abnormal diagnosis of triage in order to make modifications, constitutes a significant issue. The researcher worked with the Emergency Department of a Taiwan Medical Center to build a model to view abnormal diagnoses in the database from the establishment of a flow path and the selection of parameters for sampling. Data on patients were derived from the database. Two-stage cluster analysis (Ward’s method and K-means) and decision tree analysis were made on 501 abnormal diagnoses in an emergency department. It was found that nursing personnel make more frequent triage diagnoses than physicians do. Most of abnormal diagnoses stems from patients rather than the diagnosis on the day. Pulse and temperature have greater distinction. The researcher proposes seven correlation laws based on confidence and support proportions, derived from sample point conforming to correlation law that abnormal diagnosis is most likely in diseases of pneumonia and cirrhosis, etc. Through data mining technology, the researcher’s triage expert system is written in simulation. After periodic updates, it can improve the system and education training without the influence of the subjective factor.