Epic Failure

Note: this post is centered on the healthcare system within the United States.

Ben Thompson’s Daily Update today reminded me of many conversations I’ve had over the years. The article focuses on Epic Systems’ CEO’s attempt to rally support against new government regulation that would require interoperability at a programmatic level between Electronic Medical Records. Epic’s arguments against requiring API access to medical records basically boil down to:

  1. We’ve been giving people access to their data for years so there’s no need to interoperate.
  2. Costs will go up.
  3. Patient privacy will be threatened.

Electronic Medical Records have interested me ever since I was first exposed to them when my wife began medical school. My observation: medical software is a mess. It’s a labyrinthian tangle of rows on rows on rows of tabs, form fields, and constant alerts. It is difficult to learn, and you will have to do it every time you change jobs because every hospital spends millions to have significant customization built for them. It takes up so much care provider time that many practices have begun to hire scribes who follow physicians around and fill out the EHR for them so they can actually look at the patient and speak to them. I doubt Epic or Cerner include the cost of extra staff to fill out forms in their sales pitches to hospitals. The software is such a user experience disaster that people are willing to pay someone else to do it for them.

Before I get into the arguments that the EHR industry is making against interoperability I thought it was important to establish just how bad their product already is. Now that that’s taken care of, let’s dive in:

Data Access

While it is possible to download medical data, it is difficult, not standardized, and not very useful to the lay person.[^1] It is clear that EHR makers are not interested in making the records they provide to patients usable or digestible. If there were true programmatic interoperability then I would anticipate dozens of startups and the big tech companies figuring out ways to provide exactly this kind of data to you. The one thing that they have figured out how to give us easy access to is billing.

It is not typical for patients to know how they can access their medical records and every care provider they visit is responsible for providing the data in the way they see fit. Many patients arrive at a new provider and assume that by supplying their personal information that their medical records from a different provider are easily accessible. This is a reasonable assumption, and if these interoperability regulations had been in place from the start the assumption might be true. However, reality is very different. There is no central repository of medical records and different EHRs do not speak to each other.

For me, argument one fails.

Costs

It’s hard to imagine EHR costs going even higher than Epic and Cerner have pushed them already. Contracts with large care providers run into many millions of dollars. They jealously shut out other vendors and competition and are eager to put their thumbs on the scales when it comes to lobbying and leaning on their healthcare partners to lobby on their behalf. If a new up and comer EHR provider that promised usable software were to bid for a hospital contract at a much lower price than Epic is currently charging them, they would be dead in the water because they would not be able to import the data in a cost-effective way from the existing Epic system. By making the EHRs interoperable this moat could be eliminated. Competition is what drives down prices and currently Epic is pushing for less of it.

For me, argument two fails.

Privacy

There is hardly something more important in healthcare than patient privacy so this is a big one. Ben Thompson argues that this is a real concern, and I agree. However, I think that it is overblown. I would bucket healthcare information with financial information when it comes to how much people care about privacy. I don’t want my banking info spilling out publicly, but the convenience of accessing my account info online is too great to sacrifice for ultimate privacy. I can access my banking information directly, but in addition I make use of 3rd party tools that access my financial info in a variety of ways:[^2]

There’s a reason that similar tools for healthcare aren’t available, because it’s almost impossible to access all the different EHRs that individuals may have built up over the years. Privacy is a real concern, but I think the state of digital financial tools demonstrate that it is not a unique one. There is value, monetary and social, that could be created if a market for 3rd party healthcare data services wasn’t so hobbled.

So for me, argument three fails.

Interoperability FTW

Interoperability would be a huge win for healthcare providers and patients within the US. Despite the monopolies fighting it, the proposed regulation seeks to make real what many patients naively assume already exists, a thriving marketplace of people seeking to build quality software that makes doctors’ lives easier and provides useful insights to patients.

[^1]: I’ve downloaded raw XML and CSVs before.

[^2]: There is no regulation that I am aware of requiring financial systems to be interoperable. However the fact that money moves between institutions, accounts, etc. on a constant basis means that there are industry conventions that are widely understood. Recently VISA acquired Plaid, a service that builds custom integrations for different banks so that a single service can access accounts from different institutions. It is a brute force approach that would be difficult to replicate with EHRs, because the industry is so locked down by one or two players.

Healthcare EHR Software Technology Politics