How technology is helping hospitals to predict patient admissions
What if you knew how many patients would walk into your hospital today—and when and why?
Australian hospitals are constantly managing changing demands on their services. Emergency departments have to rapidly respond to increases in visitation caused by a long list of factors, such as an influx of patients during the flu season, or unexpected events caused by natural disasters or outbreaks of illness.
Having too much or too little capacity creates a cost burden on hospitals, and a burden to patients. An accurate view of future admission patterns can enable hospitals to improve patient access to care, the financial position of the hospital, and improve compliance with national targets. Until recently, there has been no reliable way of predicting future admission volumes but thanks to a digital solution developed by CSIRO and Queensland Health, and delivered by Telstra Health, technology is helping to address this challenge.
Why is it so challenging to predict demand?
Hospitals have access to rich sources of patient and operational data, which could, in turn, offer rich insights. However, because these data often sit across several different systems, it can be difficult to get an accurate picture of the demand a hospital can expect in its emergency departments. The disparate systems and limited visibility means that hospitals have little ability to identify the root cause of fluctuations in demand to leverage past trends intelligently.
What does that mean for hospitals?
While hospitals are good at predicting demand based on experience and consideration of known external factors, demand predictions are often only possible the day before they are expected to play out, leaving little time to plan for fluctuations in demand. With little time to plan, fluctuations in demand can have broad implications on service delivery. One activity that tends to suffer is elective surgery, which may see increased cancellations as a tactic to create more capacity. This is obviously a negative experience for patients and it can also compromise a hospital’s ability to meet its elective surgery targets. Understanding demand can help minimise elective surgery cancellations and the need for rescheduling. In addition, there are a number of resources that hospital leaders manage, such as budgets, staff and beds. Without long-term demand forecasts to assist with planning, the opportunity to proactively optimise these resources is lost. If we consider the end-of-year holiday season or public holidays, when many hospital staff go on leave and demand in hospitals generally reduces, there is a need and opportunity to close beds. However, it can be difficult to know exactly how many beds to close each year. Even with past years’ data and allowing for standard population growth, it can still be a guessing game. If too many beds are closed, agency staff may need to be hired at the last minute. Improved demand visibility makes planning more accurate and easier to manage.
Using technology to help predict a more accurate future
CSIRO and Queensland Health have collaborated on a project to test the predictability of hospital admissions, and have proved that emergency department volumes are not random and can, in fact, be predicted using data analytics. The outcome of this project has been the development of Telstra Health Predict—a digital tool that is available now to help hospitals predict demand in admissions with increased accuracy.
Predict uses complex algorithms to analyse historical emergency and inpatient admissions data to predict the number of patients admitted and discharged in the future. Austin Hospital, which has been using Predict for approximately three years, has been able to predict the number of expected presentations with specific injuries or illnesses with 90% accuracy. This has helped the hospital facilitate efficient planning of staff, beds and other resources. Information is presented in an easy-to-use dashboard for management to easily review and analyse. Hospitals like Austin Health are using Predict to leverage the benefits of forecasting to assist with winter planning, hospital staffing and longer term capacity planning. Specifically, it has helped with the improvement of bed management, staff resourcing, and scheduling of elective surgery—exactly the outcomes many hospitals are looking for. From a patient standpoint, Predict has helped to enable more timely delivery of emergency care and less time spent in the hospital by patients.
This article first appeared in the June edition of The Health Advocate.