The Big Cure

 New technology can fix Australia’s ailing health system, but only if    

To fix health care and human services – two government programs as egregiously expensive as they are important – we need to solve a basic property rights problem. It’s not a spending issue. It’s not a management flaw. This is a problem of data ownership and we can solve it with new technology.

Healthcare ($150 billion) and human services ($165 billion) consume about 60 per cent of Australia’s federal and state budgets. Any solutions in these domains that can drive even marginal improvements will make a huge difference to public finances, but most effort in this space is directed at spending constraints or organisational efficiencies.

What is striking about both domains, compared to other sectors of the economy, is how resistant they are to change through productivity growth and innovation. In one sense, this is unsurprising, given they are mostly services, and highly regulated, making change costly. Healthcare also invests heavily in research and development, which drives demand for new products and services, leading to even more spending. But both domains have fundamental architectural design principles that constrain innovation.

A key issue is data, particularly patient or client data. From the beginning, for reasons related both to atavistic technologies and to the prerogatives of statecraft, data records have been almost entirely centralised.

Centralised data—patient records and human services client databases—is widely regarded as a good thing. It promotes legibility, as a single clear record, and facilitates control, as a standardised format that can be centrally managed and monitored. Decentralised records would be potentially chaotic, so instead we have silos of centralized data across each delivery domain.

Yet because many services overlap, there is increasing value in sharing these records. The frontier of public service management is the drive to promote interoperable or joined up data. This is logical, from a managerial efficiency perspective. It is also expensive, requiring extensive re-engineering of legacy database systems.

There is another way to solve this problem. It involves recognising that the solution is not necessarily fully integrated data centralisation, but rather to shift the ownership of the data. At the moment, the government owns the data. What if the patient or client owns their data?

To appreciate why that shift in personal data property rights might be key to revolutionising the innovation potential of these sectors, we need to look to three new technologies: two of which you probably already use many times a day, and one that you may have heard of but not appreciated its significance.

These technologies are the smart phone, artificial intelligence (AI), and the blockchain. My claim is that the combination of these three technologies will revolutionise how healthcare and human services are delivered, and possibly much faster than you think, but only when all three are combined. With just one or two, nothing much changes. For this revolution to happen, we need to change the basic architecture and property rights of data ownership.

The current default gives government ownership of client data, which it then manages. To enable these three technologies to unleash their full potential, the ownership of data needs to shift to the patient or client, who will then give permission to the government and others to access that data. How does this shift engender a revolution?

First, consider what a smartphone is. It is not a phone, or even a portable computer, but a multi-tool that can be used, with targeted apps, to generate vast amounts of real-time biometric and sociometric data. Already, terrabytes of such data are generated, almost none of which enters the public health system.

Second, consider AI, such as IBM’s Watson. It does not replace doctors or nurses, but is a tool that embodies deep knowledge that can input patient generated data to suggest or augment diagnoses. AI can also work to reveal patterns and discover connections in social data that are difficult for humans in the field to see. AI can significantly augment service delivery, when it has access to data.

Third, consider the blockchain, the newest of our three technologies, which is a technology for creating public records.

A blockchain is a cryptographically secure decentralised database. If we move health and client records from centralised government databases onto the blockchain, the benefit is that we incentivise citizens to contribute data to the blockchain on smartphones, in return for those same clients choosing to grant access to that data. They may permit government to have access, but also to others, including AI.

To innovate in healthcare and human services, we need to engage all three technologies. But first we need to change the structure of data ownership from centralised government ownership, the current default, to individual ownership.

With individual ownership of data—giving citizens the power to decide who gains access to their data—each citizen has an incentive to upload as much data as possible to the blockchain in return for the best possible service, some of which may come from AI. Moreover, with blockchain technology, data sharing can be precise, and becomes a bargaining process. This would enable entrepreneurial innovation in this space.

If we want to fix health and social services, data should no longer be stored and owned by government, but should be created, uploaded and managed by citizens. Government would have access to the data because citizens see value in granting access. But first we need a revolution in property rights, shifting data ownership from a default setting of ‘owned by government’ to a new setting in which data is owned by citizens.

The future of healthcare and human services depends on new technology, but this requires a fundamental shift in property rights over personal data. This change, as messy as it is, needs to happen if we are to solve the innovation problem in health and human services.