Data modelling of patient flow for NEAT purposes
The CSIRO has released a whitepaper outlining how evidence-based analytical and decision-support tools can help hospitals understand barriers to reducing emergency department waiting times and help them reach the four-hour target by 2015.
The whitepaper, “Evidence driven strategies for meeting hospital performance targets”, covers a number of tools and techniques that CSIRO has developed that allow hospitals to use data modelling to help improve bed management, patient flow, identify 'frequent flyers' and make evidence-based decisions on system reconfiguration.
With the introduction of the National Emergency Access Target (NEAT) – also known as the four-hour rule – in 2011, hospitals are now under pressure to reach the target of either admitting or discharging emergency department presentations within four hours by 2015.
Figures from the Australian Institute of Health and Welfare on 2012 performance, released yesterday, show that while some jurisdictions are almost reaching their targets, there is still some work to be done.
Debate has raged for many years about what the solutions are to access block and ED overcrowding, and many clinicians fervently believe the answer is quite simple – more beds and more staff. With budget limitations and cuts to health budgets now being instituted in NSW, Victoria and Queensland in particular, that is not likely to happen any time soon.
Sarah Dods, health services research theme leader at CSIRO's Digital Productivity and Services Flagship and lead author of the whitepaper, said understanding the reasons for access block and ED overcrowding – and the potential solutions to these problems – should always be based on evidence, which was now starting to appear as electronic data collection makes it easier to model and predict hospital performance.
On the question of whether or not more beds and more staff will ease access block, Dr Dods said data modelling and analysis can now provide evidence for those kinds of discussions.
"Perhaps you do need more beds, but there should be clear evidence,” she said. “There is also evidence that we can make better use of what we've got. As hospitals move to electronic data systems and we are able to do more patient flow analysis, the discussion can move on from beliefs and assumptions, and becomes 'here is what is currently happening, and here is the impact if we make this change'.”
The whitepaper covers many aspects of how CSIRO's research and development of data modelling tools can provide the evidence for those discussions. It explains how data analysis can help understand and improve patient flow; how linking ambulance, ED and admissions data through a data integration tool can improve patient flow research; and how evaluating length of stay performance in EDs provides evidence that a high proportion of discharged ED patients fail the NEAT four-hour target during the early hours of the morning.
Data modelling has also found evidence that instituting strategies to make discharges earlier in the day does in fact help with patient flow and hospital overcrowding if the discharge peak occurs earlier than the peak in admissions.
CSIRO has also developed tools for better bed demand prediction, patient flow visualisation, bed configuration and to identify 'frequent flyers'. According to CSIRO, this model can be used to identify inpatients pending discharge who have high risk of readmission to hospital.
CSIRO is also looking to roll out its patient admission prediction tool (PAPT), developed in association with the Queensland government through the Australian e-Health Research Centre and now in use throughout the state.
The whitepaper states that contrary to conventional wisdom that emergency patient volumes are unpredictable, the number of admissions per day can be predicted with remarkable accuracy.
PAPT uses historical data to provide an accurate prediction of not only the expected patient load but their medical urgency and specialty, and how many will be admitted and discharged.
“Queensland Health now runs the system on their servers, and the software is available in 27 hospitals in Queensland, which are at different stages of maturity in using it,” Dr Dods said.
“It's a really good example of how a research partnership working closely with clinicians can help to understand the right problems to solve and provide solutions that work for that market.
“Overcrowding in hospitals is an international crisis and CSIRO hopes to commercialise the tool with an international partner. I think what we are also seeing is that the modelling patient admissions is just one part of modelling a very complex system.”
In a practical sense, hospitals can now start to look at whole-of-hospital or whole-of-system solutions by intelligently using the data that they already collect, she said.
“The value is in working with various people in the hospitals and discussing how they make their decisions, and the workflows they use to make those decisions. It even comes down to redesigning forms – this is a piece of information that you didn't have before that you now need to know, so add an extra box to the form.
“We do actually need to take it down to that level – the technology is not enough; it needs to sit within the clinical workflow.”
CSIRO is keen to encourage hospital managers and state health departments to work with it and discuss their needs.
“The latest NEAT performance results came out yesterday and they are showing that we've still got quite a long way to go to reach those targets,” Dr Dods said. “Clearly there are some new tools that are needed – people are still struggling to find the right way to meet those targets.
“If you have a look at the MyHospitals website, there are clear differences in NEAT performance between different groups of hospitals. Large regional hospitals do the best, and the major city tertiary hospitals do the worst.
“Using this kind of modelling helps health services and health departments understand what is needed and how big the challenge is.”
Posted in Australian eHealth