Forecasting emergency department presentations

Abstract

Objective: To forecast the number of patients who will present each month at the emergency department of a hospital in regional Victoria. Methods:The data on which the forecasts are based are the number of presentations in the emergency department for each month from 2000 to 2005. The statistical forecasting methods used are exponential smoothing and Box-Jenkins methods as implemented in the software package SPSS version 14.0 (SPSS Inc, Chicago, Ill, USA). Results:For the particular time series, of the available models, a simple seasonal exponential smoothing model provides optimal forecasting performance. Forecasts for the first five months in 2006 compare well with the observed attendance data. Conclusions:Time series analysis is shown to provide a useful, readily available tool for predicting emergency department demand. The approach and lessons from this experience may assist other hospitals and emergency departments to conduct THE ABILITY TO PREDICT the demand for attendance at the emergency department (ED) of a hospital is valuable -at a micro level for planning rosters for staff in ED, and at a macro level for financial and strategic planning for the hospital. Also, ED attendances are indicators for many aspects of health care in the community, and hence, such predictions may be useful for regional health care planning.There are many factors that will influence the attendance rate at an ED. Indeed, in England, Milner et al found "startling" variation in ED presentation patterns in 190 health districts. 1 Farmer and Emami 2 also noted regional variation in demand for health services in England and the What does this study add?This study develops a model for forecasting the number of presentations each month at the emergency department of a hospital in regional Victoria. The forecasts provided by the model are shown to compare very well with the actual outcomes. The study demonstrates how time series analysis can be used for forecasting, at least in the short term, the demand for emergency services in a hospital. What are the implications for practice?Every hospital is unique. However, the forecasting methods and approach in this paper could be applied in other hospitals to improve resource allocation and strategic planning.

From Model to Forecasting: A Multicenter Study in Emergency Departments

“…This is also important because new models of emergency crowding have been developed using a mathematical model of the incoming flow (patient visits). 7 Numerous authors have already become interested in the prediction of patient visits in EDs, [8][9][10][11][12][13][14][15] as well as in walk-in clinics, 6,[16][17][18] mainly using calendar variables in multiple linear regression models. 19 These methods examine correlations between patient visits and a number of independent determinants, mostly calendar variables.…”

Early detection of abnormal patient arrivals at hospital emergency department

“…They discussed about strength and weakness of each method. [12] The results presented in the literature show that ARMA model [15], [16] and its variants [17] provides an analytical tool of time series very accessible both in terms of methodological constraints and the level of mathematical models used, which are less complex linear stochastic equations. In this study, ARMA model was used to describe the variation in demand for emergency care in an ED system.…”

Estimating Hospital Admissions with a Randomized Regression Approach

“…They used linear regression and found that calendar features are more useful than weather features. Champion et al [5] propose using ARMIA models to forecast monthly ED presentations. Jones et al [6] investigated techniques, such as linear regression, SARIMA, exponential smoothing, time series regression and artificial neural network to forecast daily ED presentations.…”