Big Health, Small Data

By Doug Given

We all know it: Technology hasn’t made the impact in healthcare that everyone was expecting. The McKinsey Global Institute’s May 2017 update on innovations in digitization, analytics, artificial intelligence, and automation noted that the healthcare sector had not achieved the digitalization benefits outlined in its 2011 big data research. In fact, healthcare “had captured less than 30 percent of the potential value highlighted previously.”

It’s a conundrum: We find incredibly sophisticated technology being used for diagnostics and treatment, but at the same time, less than 20% of payments to healthcare providers are done digitally, and it’s still virtually impossible for people to share their health records across providers. 

Tail Wags Dog

We’ve been approaching digitalization in healthcare from the wrong vantage.

For example, Microsoft has 168,000 enterprise customers in healthcare and a huge sales force, yet scaling its HealthVault personal health record platform floundered—as did Google Health’s personalized health information centralization service. Both companies misjudged HIE business models, underestimated the weight of healthcare technology’s myriad ingrained standards and moved forward before solving for key legacy information stuck in an analog quagmire.

Instead of looking at simplifying healthcare through the lens of “big data”—as so many technology companies do—we need to be looking to achieve “big health” through the lens of small data.

The healthcare sector is extraordinarily fragmented, complex, regulated, regional, and hyper-local. There are an estimated 100+ separate closed healthcare systems in the US (Mayo, Mercy, Kaiser, Carolinas…), and more than 280 health information exchanges. Moreover, upwards of 5,600 hospitals, more than 100 major regional healthcare markets, and about 1M physicians are addressing the healthcare needs of 320M+ Americans, many in underserved and underresourced settings.

All produce data…mountains of it…but in a lava flow of small, disjointed clumps with troves of incompatible semantic and syntactic ontologies. None communicate broadly in a common language or with a similar minimum viable patient data set. Health data is not ambulatory. This is the “small data” problem that must be solved to achieve any impact at scale. But how?

Scaling SWIFTly

The financial services industry presents promising models. A few decades ago, banking needed a way to improve global transaction clearing. So they created SWIFT, which lets financial institutions worldwide send and receive financial messages in a secure, standardized, and reliable environment. The SWIFT co-op gave the financial world a shared data processing system and worldwide communications network. It provided a way to securely exchange small data.

And that network operates well. Different banks, different systems, different languages, different countries, different security protocols, and different currencies. Yet, despite all that complexity and fragmentation, people can easily and securely access funds from their accounts whether at home or traveling internationally. It is quick and painless because SWIFT exists.

SWIFT links 11,000+ financial institutions in more than 200 countries and territories, exchanging 15M+ messages per day on average. It makes people’s financial lives easier, empowering them in real-time, at the point of interaction. Big dollars from small data.

Horizontal Exchange

The SWIFT concept holds obvious potential for the healthcare ecosystem. Creating a horizontal healthcare data exchange operating with shared principles (modeling, say, TCP/IP or HTTP) enables true data liquidity. Obviously, healthcare will require something modified beyond the “trusted handshake” that is SWIFT, but we can leverage advances like blockchain to provide seamless, authenticated, immutable data with the requisite security. Such a solution for the exchange of small data will fuel both innovation and efficiency via ready access to the right health information, in the right place, at the right time — with something akin to the ease with which you can withdraw your California dollars as British pounds from a random ATM on a layover at Heathrow.

With a chronic care burden that gets larger by the day, it would behoove us to create that horizontal exchange sooner rather than later. By empowering our system with technology that affords immediate exchange and secure control of health-related data at any point of interaction, we will capture the full potential of digitalization: enhanced patient/physician relationships, increased quality, and better value from our health system. We can get big health from small data.