Date of Award
8-2013
Culminating Project Type
Thesis
Degree Name
Applied Statistics: M.S.
Department
Department of Mathematics and Statistics
College
College of Science and Engineering
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.
Abstract
The aim of this research is to forecast patient volumes in the Emergency Department of a regional hospital in Minnesota, which eventually will aid in addressing the issue of registered nurse staffing fluctuation, more specifically, productivity and capacity planning in the ED. Several methods are applied to forecast arrival patient volume, and cumulative patient volume to evaluate each model’s performance. The methods considered are linear regression, time series models and dynamic latent factor method. Long term forecast for as long as six months ahead is the goal here due union regulations that only allows for significant changes in registered nurse staffing schedule be put in place six months in advance. This long term forecast will enable administrators implement effective and timely changes to enhance productivity.
The patient arrival count, where each patient is counted once in the system, is analyzed to see how many patients the department encounters hourly. Also, cumulative patient count which gives us an idea of how many patients are in the department at any given time was also considered, here patients are counted for every hour they are in the emergency department (ED). Patient who come to the ED are categorized by their acuity level. Of all the patients that came to the ED, 52% need urgent care; this group is also analyzed to predict their arrival volume.
Lastly data was simulated with different patterns and the forecasting results from the different methods were compared and estimated. The forecast accuracy and performance for these models is then evaluated using out-of-sample forecasts for up to six months ahead. Mean square error (MSE), Root mean square error (RMSE) and mean absolute error (MAE) were utilized tosee which method is most reliable and also consistent.
Recommended Citation
Nwoke, Uchechukwu A., "Forecasting Emergency Department Volumes Using Time Series and Other Techniques" (2013). Culminating Projects in Applied Statistics. 7.
https://repository.stcloudstate.edu/stat_etds/7
