טכניון מכון טכנולוגי לישראל
הטכניון מכון טכנולוגי לישראל - בית הספר ללימודי מוסמכים  
M.Sc Thesis
M.Sc StudentSharon Wissmark
SubjectImproving the Quality of Treatment in the Emergency
Department
DepartmentDepartment of Industrial Engineering and Management
Supervisors Full Professor Shtub Avraham
Mr. Sinreich David (Deceased)
Full Thesis text - in Hebrew Full thesis text - Hebrew Version


Abstract

Improvement and efficiency of the Emergency Department (ED) has been a high profile issue in the past few years. Various studies were performed on the subject. Some of these studies used IE techniques such as simulation and data mining.

Simulation is well known as a strong analytic tool that permits "What if….?" questions and demonstrates the consequences without affecting the current system. Simulation also helps in establishing general time estimations, which cannot be obtained by observations, especially in complex systems such as the ED.

Data mining is an expanding and developing area. This tool was used in medical research for finding patient classification and high-risk profile populations.

One of the subjects yet to be researched is predicting load and capacity planning in the short term. This subject could lead to improved quality of treatment in the ED - which is the goal of this current study.

This research addresses the subject of prediction and capacity planning in 3 parts:

  • Patient classification - this stage will use data mining. A large database was collected. The database includes a large number of fields such as: sex, age, tests, releasing doctor, complaints, and hospitalization department. Many iterations were performed in order to achieve the most accurate and reliable patient classification.
  • Analysis of the treatment process of each type of patient, and feeding the data into the simulation model. This process was performed in order to obtain general treatment times and create data that will be compared to the model.
  • Building a capacity planning model. Building the model was based on IE capacity planning models.

The relevant data was placed in the models and the model results were compared to the actual load that was needed at that time (obtained from the simulation results). The models were tested in different resolutions, to check their accuracy and try and improve it. The model was tested on one resource - the nurses. A statistical analysis found no difference between the predicted load (the model) and the current load (the simulation).

The accuracy of prediction of the load for an 8-hour resolution stands at 14%, which is below the goal of 15% accuracy required for a shift.

Validation of the tool was not conducted due to the complications of acquiring data from other analyses and reanalyzing the data classification and simulation model. Validation will be possible once a uniform patient classification is generated.