Experts have evolved a software that will help hospital emergency staff to predict demand on their services and thus enable administrators to introduce necessary adjustments in their operations.
The Australian e-Health Research Centre has shown that it is possible to accurately predict how many patients will present at hospital emergency departments, their expected medical needs and the number of hospital admissions.
"Accurate forecasting will assist many areas of health management from basic bed management and staff resourcing to scheduling elective surgery - not to mention reducing stress for staff and improving patient outcomes," said Dr David Green, Director of Emergency Medicine at Gold Coast Hospital.
The Patient Admission Prediction Tool (PAPT) will allow on-the-ground staff to see what the patient load will be like in the next hour, the rest of the day, into next week, or even on holidays with varying dates, such as Easter.
"We've shown PAPT vastly improves successful prediction of patient presentation and admission in two hospitals with very different populations," said Dr David Hansen, Research Director of The Australian e-Health Research Centre.
"Emergency departments already know there's a pattern to presentations and admissions, but existing models are very simplistic. PAPT uses historical data to provide an accurate prediction of the expected load on any day," he added.
The prototype PAPT package has a simple interface designed in consultation with those who will ultimately use it every day.
"Over the next year we plan to assess and quantify the impact of using the forecasts", Dr Hansen said.
The aim is to turn the prototype software package into a product that can be used through Queensland.
This work was presented at the 2008 Health Informatics Conference held in Melbourne on Monday 1 September.
A joint venture between CSIRO and the Queensland Government, The Australian e-Health Research Centre is a leading national research facility for health care innovations in information and communication technologies.