Precision Driven Health, New Zealand's largest ICT research project, officially kicked off in Auckland today with the country's largest district health board involved, while two other DHBs are poised to get involved in the public/private partnership.

The $37.8 million programme involves listed healthcare system software developer Orion Health, the University of Auckland, and the Waitemata DHB, with Counties and Canterbury DHBs about to join.

The goal of the seven-year research programme is to position New Zealand at the vanguard of precision health, which enables all information pertinent to a patient's health and well-being - clinical, genetic, environmental and social factors - be to captured, analysed and delivered to healthcare professionals in real-time.

The Ministry of Business, Innovation and Employment made a $14 million research grant with the rest met by industry participants, mainly Orion Health which is putting in more than $23 million while Waitemata DHB is contributing $200,000.


Orion Health chief executive Ian McCrae said the benefit to its shareholders should be world-leading intellectual property the company can incorporate in its software products and sell internationally, while the DHBs will be able to use that IP to deliver more individualised health outcomes to patients in a more cost-effective way, and the university to do further research.

He said while a lot of money was being poured globally into precision medicine research, particularly in the US, New Zealand has a unique advantage because it already leads the world in automation of primary health records.

"New Zealand started automating healthcare 25 years ago while other countries like Singapore are still making notes on paper. New Zealand is a decade or more ahead of some other countries," he said.

"This project is at the vanguard of that potential change."

Science and Innovation Minister Steven Joyce said a lot of people think the cost of superannuation is one of the biggest fiscal risks facing the country due to the ageing population when, in fact, it is the cost of health provision.

That challenge can be turned into an opportunity if research programmes like this one turns up "innovation that squares that circle", he said.

One of the things he's learnt about science and innovation in recent years has been the importance of collaboration and one of New Zealand's real strengths is that it's small enough to allow more cross-discipline collaboration that others struggle to do, Joyce said.

This project is at the vanguard of that potential change.

McCrae said research projects will be chosen based on real life problems the DHB clinicians say need solving, and that he'd be "very disappointed" if the research didn't produce a usable outcome within the next six-to-12 months.

One of the first two projects is being conducted by Peter Sandiford at the Waitemata DHB to develop a prototype health outcome prediction engine that can make precise health outcome predictions tailored to the specific circumstances of individual patients.

The other two-year project by Doug Campbell at the Auckland DHB will look at epidemiology and estimation of long-term surgical mortality, using mathematical models to present a method of describing post-operative risk more accurately for each patient than the current long-term surgical risk calculators can do.

All data used in the research programme will be anonymised or follow the standard research ethical guidelines and required patient approvals.

New Zealand started automating healthcare 25 years ago while other countries like Singapore are still making notes on paper. New Zealand is a decade or more ahead of some other countries.

Waitemata DHB chief executive Dale Bramley said there was more information than ever available that could help patients but clinicians need analysis of that big data so it can be applied to help treat patients individually based on a broad range of circumstances, rather than based on the average outcomes of medical studies.

Those DHBs not involved in the precision health programme will still be able to access the outcomes which will be shared across the public health sector, Bramley said.

Orion Health also released a report today on the application of machine learning in healthcare given estimates that the size of the average cloud-based electronic health record could include as much as six terabytes of data (a quarter of the whole of Wikipedia).

Machine learning is a type of artificial intelligence that enables computers to find hidden insights without being programmed.

Precision Driven Health research director Kevin Ross said in the report that preliminary research on a publicly available dataset of 100,000 anonymised diabetic patient records from 130 US hospitals showed that machine learning processes were 20 per cent better at assessing the readmission risk of patients than the standard risk scoring approach based on length of stay, acute admission, other illnesses, and other emergency visits in recent months.

The machine learning models achieved a greater accuracy because they were able to explore patient and disease specific factors as well, Ross said.