Microsoft Health Vault
After nearly 12 years in operation, Microsoft has announced it is closing its HealthVault — a service that was originally touted as empowering individuals with their health data.
Starting as it did in the year we (Mydex CIC) was founded, we’ve always taken a close interest in its progress.
Its launch caused quite a stir. That’s because, at that time, the idea that individuals might want to access, use and control their own data seemed almost dangerously revolutionary. Yet here was Microsoft offering a service where “You decide who can see, use, add, and share info, and which health apps have access to it.” The promise that “HealthVault won’t provide your health information to any person, app, or service without your permission,” seemed like a breath of fresh air.
And it also promised to be empowering, saying you could “Share any part of your health record with anyone you choose, whenever you like, to make sure everyone is in the loop.’ The HealthVault would be your account and you would be in control.
Now it’s folding …
But we’re not surprised that Microsoft is pulling the plug on this service. It highlights a key challenge in this space: sometimes jumping half-way across a river is worse than not jumping at all. Microsoft was making important and welcome overtures towards personal data empowerment. But it wasn’t going the whole hog, and the compromises it built into the way HealthVault operated meant it could never really deliver the goods.
Here’s we believe is why
First, even though Microsoft invested heavily in the product, pouring money and sponsorship into the NHS and SMEs to develop on HealthVault and open-sourcing its vast data schema, they always tied it back to proprietary Microsoft protocols and standards. To use HealthVault you had to become a Microsoft partner and adopt their technology. That’s different to creating a safe, secure neutral personal data infrastructure using open standards and protocols, which is what Mydex CIC has and continues to work on.
Second (and along the same lines) HealthVault was not an independent neutral personal health record. It consisted of a single massive database which ran in Microsoft’s cloud servers. Microsoft were in control of the third party access to it, so that anyone wanting to access it and use had to comply with Microsoft’s rules not the individual’s preferences or policies. In contrast, with Mydex CIC each personal data store is a separate, independent repository under the control of the individual. Mydex simply provides a platform for them to engage in new ways with service providers.
Third, it was expensive to use. Many of the services that wanted to use it couldn’t afford it. So it never really reached any scale. (Mydex’s whole approach is ‘cost out’ not ‘cost plus’.)
Fourth, the HealthVault was restricted to health data. That may sound sensible, but in reality people trying to manage a health condition aren’t just managing health data. When they try to get things done, they need to draw data from many aspects of their lives. Mydex CIC personal data stores do not try to set up separate data domains, never mind define in advance and impose where there borders should lie.
It’s this combination of design flaws and commercial priorities that (we believe) meant HealthVault never could achieve its promised potential.
But progress continues
The general idea of letting individuals control and use their own data for their own purposes was great and 12 years later is now making solid progress. Microsoft should be praised for responding to it all those years ago. In its application, however, Microsoft wasn’t prepared to go far enough with HealthVault. As a result, the Vault sank in the middle of the river rather than getting to the other side. Which leaves others free to pick up the baton.
In our experiments with the Digital Health Institute on ‘untethered’ personal health records (that is, health records not tethered to any particular health care provider), we have found significant potential benefits in enabling citizens to be active participants in the collection, storage and sharing of their own health and care data. It enables innovations that deliver improved efficiency and better service outcomes, mostly because personal applications and services can share a common source of accurate and maintained data about the individual (their lives, not just health their records). The DHI’s Demonstration and Simulation Environment is a platform for apps and service providers to test this new approach of self directed and self managed care, enabling citizens to work in partnership with a cluster of service providers.
Health is a hugely complex and highly sensitive aspect of personal data. Because of this, and the scale of already-existing health services, progress towards a new paradigm will be slow. But it is happening. And the potential for improvement, for both individuals and service providers, is vast.
What are the key observations from DHI programme
DHI will be publishing an extensive review of progress shortly what we have observed and been an active participant in is a range of policy review, co-design, technical simulation, live trial and industry engagement activities over the last five years. The policy and co-design drivers for next generation health and care services mandate a person-centred approach that distributes power away from organisations to individuals and their own ‘circles of care’.
New services should be distributed, community based, and must empower the individual to use their own assets to co-manage their own care. These services should be activated on the persons’ terms, and the organisational need to reduce harmful variation must be balanced against the individual’s need for productive variation.
This type of service cannot simply be a repositioning of current organisation-centric services to include more self-management. Technology has, to an extent, democratised many other sectors and services — it should be used as a lever to re-weave the fabric of how health and care services work.
In particular, the data sharing infrastructure crucial to a successful digital health and care service should reflect the chosen policy and user co-designed direction — i.e. distributed, loosely coupled, but tightly integrated — to help build trust between people and enable new types of organisation to emerge to meet diverse needs in balance to the centralised formal health and care service infrastructure.
