Officially the worst flu season since 2009’s pandemic, 2017 recorded more cases, and more deaths, than recent years. A lot of people had to take time off work. More than twice the number of people were admitted to hospital than in typical years.

It has been blamed on an ineffective influenza vaccine and a virus that mutated during the year.

But what if flu outbreaks could be better predicted so that vaccines could be sent to the areas most in need, so that hospitals could be better prepared, so that new or mutated strains could be detected?

This is what Dennis Horton, program lead and data scientist with The Data to Decisions CRC, and in conjunction with the University of Adelaide, is hoping to achieve.

Using disease outbreak modelling based on data from participating GPs, emergency room and laboratory data, and assumptions made about influenza — and potentially incorporating social media if predictive power is found — Horton is hoping ‘big data’, the type used by operations like Netflix and Uber to predict what customers want, could help to prevent the pandemics of tomorrow.

Big data, for those not familiar with the term, is often described in terms of the three Vs — volume, velocity and variety.

Volume refers to the size of the datasets that can now be analysed due to advancements in computing power.

Velocity refers to the speed at which data can be analysed, often in real-time.

Variety refers to the number of different formats that can be analysed, from unstructured text to images to video. Big data analytics is basically a way of extracting value from a variety of data sources and sizes by performing predictive analyses or analysing behaviour.

But the use of this information in healthcare is in its infancy.

When Netflix suggests a movie for you, the algorithms used are much more advanced than big data currently used in healthcare.

“It is starting to be used more in healthcare, but there is still a long way to go,’’ Horton says.

“Health as an industry is lagging behind others in the application of big data. It is an area that is steadily growing and has the potential to make massive improvements in the quality of healthcare.”

Horton hopes to lead that change, and it’s not just the flu he is fighting.

He hopes to use ‘big data’ to reduce cardiac complications, and emergency waiting room times, as well as improving mental health care.

Cardiologists from the Central Adelaide Local Health Network and The Data to Decisions CRC have collaborated to deliver a groundbreaking study using over 100 million healthcare records from more than 1000 hospitals, the first of its kind in Australia.

The project has enabled a nationwide assessment of outcomes of hospital based cardiovascular care for common conditions and procedures using ‘big data’ analysis.

“There has been very little work, if any, performed on the analysis of Australian patient outcomes in the cardiac space,’’ Horton says.

“By monitoring risk-adjusted post-procedure outcomes, such as 30-day mortality or 90-day unplanned readmission, we are able to assess the performance of all hospitals in Australia. “For hospitals performing above the average, we can identify their characteristics and apply them across all hospitals which will improve patient outcomes on the whole.

“The data used is de-identified for patients and hospitals.”

There’s only one thing worse than being sick and it’s sitting in an emergency room on a Saturday night with a lot of other sick people watching infinite numbers of them see a doctor before you.

Data to Decisions has developed an application which gives waiting time information and directions for all South Australian metro hospitals.

The app considers a patient’s current location (or any location specified) and combines this information with emergency waiting times, as well as with Google travel information.

This highlights the most optimal emergency departments according to wait and travel time.

Horton hopes it will help patients get the quickest medical attention and reduce the burden on emergency departments.

“We are looking into building triage estimators into the application that could eventually guide people to other forms of treatment such as after hour clinics,’’ he says.

“This reduces patients going to emergency departments for non-emergency reasons. For example, someone needing stitches could do this at a nearby clinic.”

For those with mental illness, Horton hopes an app, designed in collaboration with Flinders University, could identify possible behaviours — for example, when usual medication or appointment needs are not being met — that may lead to a hospital readmission.

It is estimated that an acute mental health hospitalised patient costs around $1500 per day.

“It is clear that if we can promote out of hospital care for mental health patients then there would be financial benefits to the healthcare system,’’ he says.

“Patients can sign up to the application with their Medicare number and then all their data is automatically generated from Medicare records.

“It analyses all prescription and appointment information to identify any possible behaviours that may lead to a hospital admission. Their clinician can then use this information to treat the patient in the most appropriate way.

“The use of big data analysis in the health space presents exciting opportunities to improve patient care. It is lagging behind many industries but the tide is turning as people realise the value that data analysis can add.

“We have already demonstrated from our completed projects, how much potential exists to improve outcomes for patients, and we hope these projects will continue, along with new ones in the coming years”