黑料吃瓜群网

What if you knew who would听walk into your hospital today- when and why?

Dave Piggott, Executive Director of Health IQ, investigates how the partnership between Austin Health, CSIRO and Health IQ works to provide Austin Health these answers.
Without the ability to accurately predict patient admissions, hospitals are limited in their ability to effectively plan for changing demand. As a result, they face a range of complex issues that negatively impact patient access and care, the financial position of the hospital, and their ability to comply with national targets.
To better understand of how the challenges of planning ahead plays out for hospital decision makers, I spoke to Fiona Webster, Executive Director, Acute Operations at Austin 黑料吃瓜群网.
It鈥檚 been a long time since Fiona began looking for a way of predicting hospital demand more accurately - but still, like for most Australian hospitals, the Austin 黑料吃瓜群网 is limited in its ability to accurately forecast what the demand for even just the next day will be.
鈥淲e look to what happened the day before, at the average number ED presentations, the number of booked elective surgeries to try and estimate the demand for the next day. We also look at the number of patients waiting in the hospital and the number of estimated discharges for the day to gain visibility of the hospitals鈥� capacity. If it looks like there鈥檚 going to be significant gap between the expected demand and available capacity, the issue is escalated 鈥� so that actions can be identified to reduce the imbalance.鈥�
Why can鈥檛 hospitals effectively predict demand?
Many disparate systems
Because of the large number of disparate systems in use, it becomes quite difficult to consolidate all the information required to reach a real picture of the demand they can expect. 鈥淭he real swings in bed requirements happen not because of a single factor, but due to a coalescence of four or five factors. For example, we may reduce beds because a large number of surgeons are away on leave, but if there is suddenly an influx of patients coming in through ED, then we鈥檙e left with too few beds. Currently, effectively tracking such disparate factors that may impact demand in the future is extremely challenging.鈥� said Ms Webster.
Limited understanding of what causes demand fluctuations
The disparate systems and limited visibility also means that the hospital has very little ability to identify the root cause of the various fluctuations 鈥� and by the time a cause is identified, the situation may have already changed significantly. There is no opportunity to identify and leverage past trends intelligently. 鈥淎t the end of the day, we鈥檙e simply making an informed guess.鈥�
What does that mean for Austin Health?
The limited ability to accurately predict demand is the cause of some fundamental challenges.
Last-minute decisions
Currently, demand predictions are often only possible the day before they are expected to play out. Due to this, decision makers only have a limited time to plan against these predictions 鈥� meaning that they are constrained in the actions they can take to address the expected fluctuations in demand.
Unfortunately though, elective surgery is one of these factors, and it鈥檚 often the one that gets cut in order create more capacity. 鈥淥ur ability to optimally flex beds and other resources is limited, so the default is cancelling elective surgery. Nobody wants to cancel surgeries: they take a lot of effort to set up, both on the patient鈥檚 side and the hospital鈥檚. Cancelling them is very inefficient鈥�. In addition, cancelling surgeries can compromise a hospital鈥檚 ability to meet their NEST.
Inability to effectively plan ahead
As a key decision maker, there are a number of levers Fiona has to manage as part of her planning 鈥� such as budgets, staff, beds, and other resources. However without long term demand forecasts to plan against, the opportunity to proactively optimise these levers is lost. For instance, take the holiday season. With many hospital staff going on leave, and demand generally reducing, a number of beds are usually closed. However each year, it鈥檚 difficult to know exactly how many beds to close.
Even with past years鈥� data and allowing for standard population growth, it鈥檚 still a guessing game. If too many beds are closed, agency staff may need to be hired at the last minute. 鈥淚f we did have the visibility of a large surge in demand in May for instance, then I would be able to plan ahead, hire the right number of staff now without relying on agency 鈥� but currently that鈥檚 not possible.鈥�
Inefficiencies that cannot be resolved
As you know, hospitals are full of highly valuable resources, human and otherwise, and everyone does their job because they care about delivering the best outcome for their patients. In such an environment, not being able to address daily inefficiencies due to lack of demand prediction is very frustrating.
However, more often than not, there鈥檒l either be elective surgeries cancelled due to too much demand or surplus beds that are not utilised鈥� because demand can鈥檛 be accurately predicted!

I thought, if you can forecast the weather, surely you can forecast the number of patients who we can expect to come through the door. When I heard that patient admissions could be predicted, and with accuracy, I was immediately interested.

Towards a more accurate future
Fiona, who had long sought a better solution for patient admission prediction, was aware of the work CSIRO were doing on Patient Admission Prediction in Queensland. 鈥淚 thought, if you can forecast the weather, surely you can forecast the number of patients who we can expect to come through the door. When I heard that patient admissions could be predicted, and with accuracy, I was immediately interested.鈥�
What resulted was a partnership between Austin Health, CSIRO and Health IQ, who are working together to implement the Patient Admission Prediction Tool (PAPT) at Austin Health. Launched in October 2014, the tool is expected to provide Fiona and her colleagues the ability to better predict bed demand, optimise resource allocation, and maximise patient access in their hospitals. Fiona is looking to achieve accurate demand forecasts as much as one year in advance.
How it works
PAPT is a software tool that utilises complex algorithms by applying them to historical data in order to predict the number of patients admitted and discharged in the future. The tool runs unsupervised and updates data regularly, allowing new information to improve efficiency without draining staff resources. With 90% accuracy, PAPT can predict the number of expected presentations with specific injuries or illnesses, facilitating efficient planning of staff, beds and other resources.
Expected outcomes
Fiona looks forward to realising the trial鈥檚 goals: 鈥淭he ability to anticipate emergency department attendances and inpatient beds is an important aid not only to the daily challenges of bed management but it will forecast a year ahead which will assist with winter planning, hospital staffing and longer term capacity planning鈥� (Austin Health, 2014).
While we鈥檙e waiting to see exactly what outcomes Austin will experience at the conclusion of the trial in mid-2015, we can already see the benefits that have been experienced by Queensland hospitals who already leverage a version of this tool.
Specifically, it has helped these facilities improve their bed management, staff resourcing, and scheduling of elective surgery (CSIRO, 2014) 鈥� exactly the outcomes Fiona is looking for. From a patient standpoint, PAPT has enabled the delivery of improved healthcare outcomes, such as the timely delivery of emergency care, improved quality of care, and less time spent in the hospital (CSIRO, 2014).
For example, the tool played a central role at Gold Coast Health in managing the influx of patients during Schoolies. Dr James Lind, Director of Access and Patient Flow at Gold Coast Health says that of with PAPT鈥檚 prediction technology, they are able to expect 鈥渁round 2,700 presentations to our emergency department in total and around 20% of these will be school leavers in just the first week of the celebrations鈥� (as quoted in CSIRO, 2013). This knowledge enables Dr Lind and his team to better plan the staff, medical supplies, and beds for this increased demand while also catering to the needs of other non-schoolies patients (CSIRO, 2013).
听
Further Information
If you would like to find out more about how your hospital can take advantage of such improved demand prediction ability, please call Dave on (03) 9425 8012 or email dave.piggott@healthiq.com.au.
听

---

Dave Piggott
Executive Director Health IQ
Dave Piggott is the Executive Director of Health IQ, and is focused on helping Australian health services achieve better visibility and communication within and across their hospitals. Dave has over 20 years鈥� experience in Health IT. A graduate of the Australian Institute of Company Directors (AICD) and with a Masters in Open Systems (IT), Dave has worked extensively in the Patient Flow area, and helped over 30 Australian hospitals to improve their flow of patients.
听

---

听
References

1. Austin Health. (2014, October 31). Newsroom: New technology will boost Austin Health service delivery.听Retrieved from Austin Health: http://www.austin.org.au/newsroom


2. CSIRO. (2013, November 15). CSIRO. Retrieved from The one Toolie that鈥檚 welcome at Schoolies:听http://www.csiro.au/Portals/Media/Technology-predicts-surge-of-sick-schoolies.aspx


3. CSIRO. (2014, February 26). CSIRO. Retrieved from Cutting hospital waiting time: http://www.csiro.au/Organisation-Structure/Flagships/Digital-Productivity-and-Services-Flagship/Health-services/PAPT鈥揷ase-study.aspx