Published in


We’re missing a huge health data opportunity, says Daniel Petre

This article first appeared in the AFR.

Australia is missing a big opportunity by ignoring the potential for Medicare data and machine learning to transform diagnosis and treatment of common diseases, such as heart disease, diabetes, melanoma and oesophageal cancer, says venture capitalist Daniel Petre.

Mr Petre acknowledged valid concerns over health data security, which have flared up because police and tax officials can access health records without a warrant. He said tax officials should never have access to health data and police should need to get a warrant, but the debate about the value of a repository of health data had “lost all perspective”.

“If I fall sick I want the doctor I am seeing to know everything about every treatment I have ever had,” said Mr Petre. “The more she knows the better the chances are that I will get an effective diagnosis and treatment. In terms of data use by third parties there is a big difference between someone getting my health data with my name attached and someone getting health data on a 50-year-old male and using this data to create predictive diagnosis tools that will save lives.”

Using machine learnng and health data would produce a step change in health outcomes, says Daniel Petre. Brook Mitchell

Mr Petre, a director of Innovation Science Australia, the body charged with delivering the Turnbull government’s innovation agenda, said the government’s embrace of the policy did not go far enough and put workers’ futures at risk by failing to prepare them for the upheaval from artificial intelligence, which Bain & Co said could eliminate up to a quarter of US jobs by 2030.

He is one of the speakers at The Australian Financial Review’s Innovation Summit 2018, being held in Sydney on Monday and Tuesday, and a co-founder of venture capital firm AirTree Ventures.

Mr Petre said the government’s decision to pursue a national mission in genomics — recommended in ISA’s Australia 2030: Prosperity through innovation plan — was the future of medicine and a great idea but it would mainly yield solutions to rare diseases in the short term, which wouldn’t massively help population health for some time.

Short term ‘step change’

By contrast, applying artificial intelligence and “machine learning” to the health data of 24 million Australians held by Medicare and the Pharmaceutical Benefits Scheme could deliver a “step change” in population health in the very short term. But the idea didn’t have the popular support or “really cool” image of genomics.

For example, using machine learning on the massive data set of melanoma scans could help GPs and junior dermatologists accurately predict which moles will turn into melanomas and which won’t. Retinography can diagnose diabetes much earlier than traditional tests using a scan of the eye, and oesophageal cancer detection rates can be improved, Mr Petre said. Radiology and pathology could also be made more precise.

China is investing billions in machine learning, Britain has allocated £1.6 billion ($2.8 billion) to artificial intelligence projects using its vast trove of National Health Service records, and the US Food and Drug Administration recently launched a software-based diagnostic tool.

A My Health Record advertisement. Valid concerns over health data security have flared up because police and tax officials can access health records without a warrant. Fairfax Media

“We could be doing some similar things here but we seem to have a problem with the government and the opposition understanding it,” Mr Petre said.

‘Massive health economics issue’

He said Treasurer Scott Morrison should care about it as much as the health minister “because it’s a massive health economics issue”. Reducing false positives and negatives in cancer scans could save hundreds of millions of dollars in unnecessary biopsies and lost wellbeing, and suitable pictures could be taken with a $100 attachment to an iPhone.

“The rest of the world is moving at a rapid rate and if we want to provide global health interventions into those conditions then we are moving too slowly. I think we are being very cautious,” Mr Petre said. Tech-phobic chief executives and older doctors were also resistant to the idea that software could do a better job on medical diagnosis.

On the impact of AI on work, he said boosting the skills of the workforce to cope with upheaval in jobs at all skill levels was more important than giving companies a tax cut. But he said it was impossible to get an informed debate with 30 per cent of the Coalition party room not even accepting the science of climate change.

Examples of melanoma using machine learning on the massive data set of melanoma scans could help GPs and junior dermatologists accurately predict which moles will turn into melanomas and which won’t. Dr Richard Johns FSCCA

On the positive side, the sharp increase in venture capital funds entrusted to AirTree, Blackbird Ventures, Square Peg Capital and other groups should help to create some great new companies. While there was “near zero chance” of Australia developing digital platforms of the size to rival Google and China’s Tencent, “we could have a massive vertical market set of solutions … in the health space and the education space to name just two areas”, Mr Petre said.

Treasurer Scott Morrison should be concerned that we make proper use of health data because “it’s a massive health economics issue”, Daniel Petre says. Erica Canepa



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store


AirTree is a venture capital fund that supports the most ambitious Australian and Kiwi tech founders. First round, multiple rounds.