Big Data, Better Hospitals

[Music plays and the Maths inside, UTS, AAMT and CSIRO logos and text appears: Investigating the maths inside, Big Data, Better Hospitals]

 

[Images move through of a rear and then facing view of Justin Boyle walking outside of a hospital]

 

Justin Boyle: I grew up in Brisbane, Darwin and Sydney and as a kid growing up in Australia, I enjoyed the outdoors, swimming, bike-riding, skateboards.

 

[Image changes to show a rear view of Justin entering the Royal Brisbane and Women’s Hospital and then the image changes to show a roster chart displaying data]

 

At school I didn’t actually want to follow a career in maths, but I wanted to fix things.

 

[Image changes to show Justin working on a computer]

 

I wanted to have solutions to things and make sense of the world.

 

[Image changes to show Justin smiling at the camera and text appears: Justin Boyle, Research Scientist, CSIRO]

 

 Hi, I’m Justin Boyle. I’m a Research Scientist at CSIRO and we solve problems.

 

[Music plays and a graph background with equations appears faintly in the background and text appears across it: #Mathematics]

 

[Images changes to show Justin talking to the camera and then the image changes to show an outside view of the hospital]

 

One of the problems that we’re looking at is improving the flow of patients through hospitals.

 

[Image changes to show an ambulance moving along and then the image changes to show people in a queue and then the image changes to show Justin talking to the camera]

 

When you, or your parents, or your grandparents get sick and you need to go to hospital and you require a bed, you expect there to be a bed available and sometimes there’s people queueing up. So, in health care, we might say, “Let’s build a bigger hospital, more beds” but that costs a lot of money.

So, we’re using maths to improve the efficiency of the flow of patients through hospital.

 

[Image changes to show people moving in and out of a hospital door]

 

Dr James Lind: The patient admission prediction tool is a tool to look at exactly what it says. It predicts to about 95% accuracy, which patients are coming in and when.

 

[Image changes to show ambulances outside of the hospital and then the camera zooms in on the side of an ambulance and then the image changes to show a patient being wheeled into the hospital]

 

We know today that there are 12 people coming in with broken arms and legs.

 

[Image changes to show Dr James Lind and colleagues looking at computers]

 

Only one of them has come in up to date but we know there’s another 11 out there. So, what we’ve been able to do is set aside emergency theatre time for these people already.

 

[Image changes to show an orderly wheeling a patient in a bed towards the camera]

 

We know that they’re coming, and we know we can treat them.

 

[Image changes to show James talking to the camera and text appears: Dr James Lind, Director of Access and Patient Flow, Gold Coast Hospital]

 

It was difficult at first because many people didn’t believe the tool could do what we said it could do.

 

[Image changes to show James and his colleagues working and in conversation and then the image changes to show a patient’s face]

 

Up to recently a fallacy existed that all hospitals had to be at 85% occupancy for optimal patient flow.

 

[Image changes to show James talking to the camera]

 

Using the mathematics of CSIRO, we’ve actually dispelled that rumour and we can actually show categorically that that’s not true and we’ve actually worked out optimal occupancies for not just our hospital but other hospitals.

 

[Image changes to show outside views of the Robina Hospital]

 

The proof of the pudding really of this tool is we’re in the middle of winter.

 

[Image changes to show the Emergency doors opening to show people at reception]

 

It’s the worst point for Emergency Department because of the winter surge that occurs.

 

[Image changes to show James talking to the camera]

 

Up to recently you would have seen pictures of ambulances queueing outside to get into Emergency and all the beds being full.

 

[Images move through of people working in the Emergency Department, a hospital bed, a male wheeling a patient in a bed and an ambulance outside the hospital]

 

If you look today on one of our busiest winter’s day, you can see there are still three beds in the Emergency Department and there is only one ambulance outside which has managed to offload its stretcher.

 

[Image changes to show James talking to the camera]

 

What we’re able to do with this tool is show people that actually what happens in health care is very predictable on a day by day basis.

 

[Image changes to show a female working in an ambulance and then the image changes to show Kim Sullivan talking to the camera and text appears: Kim Sullivan, Austin Hospital, Melbourne]

 

Kim Sullivan: Running a hospital is like running a 1,000-bed hotel and what I do on a daily basis is try and smooth out the bottlenecks.

 

[Camera zooms in on Kim’s face as she talks and then the image changes to show data on a screen]

 

The patient prediction tool allows us to forecast tomorrow’s activity which means I can look and see how many patients are arriving to the Emergency Department, how many are coming in via ambulance and how many are presenting via the elective surgery stream.

 

[Image changes to show Kim talking to the camera and then the image changes to show a rear view of a female looking at data on a screen]

 

So, this tool allows us to match the right bed for the right patient who’s presenting on that particular day.

 

[Image changes to show hands waving in the air and then the image changes to show young adults dancing on the beach and then the image changes to show Justin talking to the camera]

 

Justin Boyle: There’s an event held at the Gold Coast every year called Schoolies Week and we see a fast, sudden increase in the numbers of young adults presenting at a hospital with particular injuries.

 

[Image changes to show the sign on the outside of the hospital and then the image changes to show Justin talking to the camera]

 

And so, whenever we have data that goes up and down often we use statistics to say whether a change is actually important. So, whether it’s statistically significantly different.

 

[Images move through of a sign on the outside of the Emergency Department, ambulances outside the hospital and a patient in a bed being wheeled through the door and down the corridor]

 

In our modelling we use data about the patients that come to hospital.

 

[Image changes to show Justin talking to the camera and text appears: 6000 x 365 = 2,190,000]

 

At a busy hospital there might be 200 patients coming every day and if we have maybe 30 really big hospitals in a state, then that’s 6,000 patients coming every single day. And then across a year, that might be over 2,000,000 patients coming every day.

 

[Image changes to show a line graph on a screen]

 

So, there’s a lot of Big Data.

 

[Image changes to show line, bar and pie graphs on a screen]

 

We’re talking about certain times that the patient arrives, their age, how sick they were and when they leave.

 

[Camera zooms in on an Inpatient Breakdown pie graph and then the image changes to show Justin talking to the camera]

 

We also have to make sure the models can take into account a sudden departure or an outbreak.

 

[Camera zooms out on Justin sitting at a desk and talking to the camera]

 

So, for example, a few years ago there was a big outbreak of swine flu and so, our models broke. So, we had to work out how to adjust to take into account changes in the patterns of patients that are coming in.

 

[Camera zooms in on Justin’s face as he talks]

 

The first part of modelling is usually visualising the data and describing that data in terms of some descriptive statistics.

 

[Text appears to the left of Justin: Mean, Median, Standard Deviation]

 

So, we use terms like the mean of the data, or average, the median. We look at the spread of that data with terms like standard deviation.

 

[Image changes to show Predictions Summary data on a screen]

 

Usually this data is unique to a particular site.

 

[Image changes to show a line of data on a screen and then the image changes to show Justin talking to the camera]

 

Every hospital will be different and so understanding that data using maths and statistics to provide some better insight into useful ways of how we’re going to model that data.

 

[Image changes to show people working in the Emergency Department and then the image changes to show a hospital bed]

 

Sankalp Khanna: Everything about your hospital visit and your health gets recorded in one hospital information system or another.

 

[Image changes to show an ambulance outside the hospital and then the image changes to show Sankalp Khanna talking to the camera and text appears: Sankalp Khanna, Research Scientist, CSIRO]

 

That’s millions and millions of patient records and that’s what we call Big Data.

 

[Image changes to show three females in conversation in an office]

 

Now using some really cool mathematics we can use it to work out how many people are going to come in and out of hospital.

 

[Camera zooms in on the females in the office and then the image changes to show people working on a patient in a bed and then the image changes to show two females wheeling a patient in a bed]

 

We can build models and then even use that to determine which of these people that are leaving hospital are going to get sick and come back to visit us within a few weeks.

 

[Image changes to show Sankalp talking to the camera]

 

We can use data modelling and machine learning techniques to figure out what days are going to get really busy.

 

[Camera zooms out to show Sankalp sitting at a desk talking to the camera]

 

We can employ a mathematical technical simulation to build a model of a hospital and use that to try scenarios like, what could happen if I added a bed into the Orthopaedic Ward.

 

[Camera zooms in on Sankalp’s face as he talks and then the image changes to show a bar and also line graphs on a screen]

 

What could happen if I added two staff members for the ICU?

 

[Image changes to show a rear view and then a facial view of Justin working on a computer and text appears: Graphs, Heat Maps, Bubble Plots, Word Clouds]

 

We use a variety of techniques, we use graphs, we use heat maps, we use bubble plots, we use word clouds.

 

[Image changes to show a line graph and then the image changes to show Sankalp talking to the camera and then the image changes to show data on a screen]

 

We use whatever it takes to get the message across to the right audience and we can combine this knowledge with mathematical problem-solving techniques to optimise the entire health system so that we can deliver better patient outcomes, so that the money that we are spending is spent at the right place and delivers the best patient outcomes for all of us.

 

[Image changes to show Justin talking to the camera and then the image changes to show a male wheeling a patient in a bed]

 

Justin Boyle: I really like being able to use mathematics to help people and particularly people that are so sick that they require going to hospital.

 

[Image changes to show a female working on a computer and then the image changes to show Justin talking to the camera]

 

Maths is great because it gives us a way to describe the world and gives us the tools that we need to help understand the world, and also become masters of it.

 

[Image changes to show a male wheeling a patient in a bed and then the image changes to show a female cleaning and remaking a bed and the camera zooms in on the bedclothes]

 

For my work we need to have a good sense of curiosity, critical thinking, reasoning, but then also good communication skills to try to help make the world a better place.

 

[Image changes to show Justin talking to the camera]

 

And really mathematics is integral to all that.

 

[Images move through of people working in a hospital, two females making a bed, Justin talking to the camera and Justin at his desk talking to the camera]

 

In the future we’re going to need a lot of students to be able to analyse all this collected data from the world. The use of predictive analytics and Big Data are going to become more and more prevalent in our communities.

 

[Image changes to show a bandage being put onto a male’s hand and then the image changes to show Justin talking to the camera]

 

Moving in to health care, we’ll be able to work out how to potentially identify genes related to ageing, maybe look at how to fix particular genetic defects by inserting cells back into someone’s genetic code. Now, it doesn’t matter what path you choose after school but it’s likely maths will form part of that in some way.

 

[Music plays and the Maths Inside logo and text appears: Investigating the maths inside, Maths Inside is a project led by University of Technology Sydney, and funded by the Commonwealth Department of Education and Training under the Australian Maths and Sciences Partnership Program, The aim of Maths Inside is to increase engagement of secondary school students in mathematics, by using rich tasks that show the ways it is used in real world applications, To find out more about this project and other AMSPP resources, please go to http://dimensions.aamt.edu.au, Maths Inside 2016 except where otherwise indicated, the Maths Inside materials may be used, reproduced, communicated and adapted free of charge for non-commercial educational purposes provided all acknowledgements associated with the material are retained, Maths Inside is a UTS project in collaboration with CSIRO and AAMT]