Open vs. Closed Data Ecosystems in Healthcare

Earlier this week, Google, Amazon, IBM, Microsoft, Oracle and Salesforce announced their joint commitment to improving healthcare data interoperability. In other words, the biggest players in cloud computing have taken a small — yet critical — step towards an open data ecosystem.

The technology world frequently oscillates between “open” and “closed” systems. Closed systems (like Facebook or iPhones) are designed to work end-to-end, and wall off or minimize collaboration with other parties, typically blocking data portability and interoperability in favor of a smooth end-to-end experience. Open systems (like the Internet and Android) are designed to integrate hundreds of different vendors. From a product perspective, closed systems tend to have smoother end-to-end experiences if a customer is willing to buy in fully to the vendor, while open systems tend to provide more choice & flexibility thanks to access to a best-of-breed set of vendors. From a business perspective, closed systems tend to form powerful monopolies, exerting strong control over all parts of the system — and thus tend to be the preferred business model for most emerging tech platforms.

The technology and data around healthcare has to date been an extremely closed ecosystem. There is very little interoperability between different sources of data; when I show up at a doctor’s office, the odds that they can find my past records from other systems is extremely low. Similarly, the ability for a medical researcher or company to be able to combine data about a patient from several data vendors is very difficult today.

The closed nature of the healthcare data ecosystem is bad for patients, data sources, and consumers of data. While closed ecosystems aren’t inherently bad, the health data ecosystem is too fragmented for patients or companies to be satisfied locking into the walled gardens of a single solution. A consumer could theoretically spend their entire digital lives within the Apple or Facebook walled garden, but that isn’t possible in healthcare: the best hospital systems only have a snapshot of a patient’s life and the largest health data vendors only have a small subset of data about patients. As a result, an open data ecosystem is critical for completeness of information.

In recent years, we have seen both private and public actors begin to advance the cause of an open data ecosystem:

  • The 21st Century Cures Act, which became law in December 2016, encourages interoperability of electronic health record (EHR) systems and discourages information blocking;
  • Health Level Seven (HL7) put forth common data standards, which have been spread in part through a private sector initiative, the Argonaut Project;
  • EHR vendors support clinician-to-clinician exchange of data (e.g., EPIC Care & Share Everywhere), and are aspiring to real-time data flows.

These and other efforts represent significant progress, but there is still a long way to go.

The reasons that the healthcare data ecosystem remain mostly closed are complex, and include legacy systems, on-premise hardware, friction in data contracting, collective action failures, and challenges around regulatory compliance. We also see business models that are aligned with a closed ecosystem:

  • In the past, some EHR vendors were accused of creating barriers to extracting information from their systems in an effort to induce vendor lock;
  • Data aggregators make it difficult to combine their data sets with other data sets for fear of cannibalization

One cost of these closed systems is incompleteness of patient information, but another key distinction is that in a closed system the uses to which data is put are bound by the original data owner’s imagination. In healthcare, that means that the data might never reach innovative people developing potentially transformative healthcare analytics and applications.

Establishing an open data ecosystem is essential if the healthcare system as a whole is going to deliver high quality, cost-effective and seamless care for patients. A few examples might make this clear:

  • Patients suffer when their primary care physicians can’t access data from a specialist’s electronic health record (EHR) system;
  • Innovative treatments are delayed (or, in the worst cases, foregone) when pharmaceutical companies can’t find valuable demographic data that would inform site selection and patient recruitment

To get to a more open system, there will need to be not only strong interoperability standards, but a shift in business models to embracing openness. This week’s announcement only addresses a small part of a very big problem, but (for the moment), the intention is more important than the details. The public commitment of tech leaders to an open data ecosystem is excellent news for everyone who has a stake in the healthcare system, and hopefully builds more momentum towards openness.

Thanks to Bob Borek for his help authoring this article.


Open vs. Closed Data Ecosystems in Healthcare was originally published in Datavant on Medium, where people are continuing the conversation by highlighting and responding to this story.

datavant-logo-stacked-sharp

Copyright 2018 © Datavant