February 12, 2019

Interoperability is not enough

Watch this article as a video

Summary: interoperability of healthcare data between clinical systems, and even interoperability of healthcare workflows between clinical systems will not break the market forces that prevent innovation in electronic health records.

Interoperability between clinical systems will not bring down the high barriers to entry and the high switching costs of the medical record marketplace.

Introducing big-tech companies into the commercial medical record market will add new vendors, but not fundamentally fix the dynamics that lead to a stagnant marketplace.

Ultimately interoperability not just between clinical systems, but interoperability between the functional components within a clinical system is required.

Interoperability between systems is not enough

Sharing patient data

A lot of the focus today is on interoperability between clinical systems. From one hospital to another. From the clinic to the patient.

The sharing of this information needs to not be by fax, pdf, or mule, but via a format that can be ‘consumed’ and ‘ingested’ into the receiving clinical system.

This means creating, and using, standards to exchange patient data. A process many decades in the making, but progress is finally happening with increased uptake of FHIR, and consumer market forces.

Sharing workflows

There is an effort (though far less adopted) to take interoperability between systems a step further. Instead of only sharing patient information, there is effort to also be able to share FHIR workflows.

For instance, an order can originate on the medical record in the Doctor’s Office, and pass seamlessly into the medical record system used in the nursing home. (See FHIR Workflows for more details).

Both of these steps towards clinical system interoperability are important and long overdue.

However, interoperability between clinical systems is not sufficient to break the EHR marketplace out of stagnation and develop the innovation required to advance healthcare.

Let’s review briefly why progress in electronic health records has been slow.

What's holding EHRs back?

Everyone will give you a different answer to ‘why innovation in electronic health records is slow’. I suspect most of these reasons  are justified.

But for this post, I want to highlight a specific area that has hindered innovation: the size and scope of electronic health record software.

Major electronic health record systems for inpatient and outpatient care perform a number of functions: patient care, medication coordination, billing, scheduling, etc, etc, etc.

The sheer number of things that an EHR does makes it difficult to switch to a new clinical system. Doing so is expensive and time consuming. The system is deeply integrated into the facility’s other systems.

The broad feature set that exists in large EHR vendors is very difficult to build out for a new competitor entering the EHR market. Legacy vendors have been able to build up their feature arsenal for years.  In order for a new company to compete directly they must offer comparable features to the legacy vendors. This is tough to do.

And, even if a new startup competitor could enter the marketplace and compete on features, who would buy from them? Established hospitals and clinics want to purchase software with confidence it will be around and supported in 5 or 10 years time. If there is any chance the company or product may fold, the thought of having to switch to another system early is painful.

Ultimately the electronic health record marketspace is hard to enter, and once your are there, tough to compete in given the high pain required in switching clinical systems.

This is in many ways why EHR companies do not compete in their sales materials directly on quality but instead sell reasons to purchase their product based on ‘market share’. Nobody got fired for buying IBM. (Who cares how good the product is. Everyone else is using it.)

Let’s summarize:  

+ High barriers to market entry
+ High switching costs
= Low innovation
= Low energy market

Won’t Amazon fix everything?

If the problem is that health systems won’t purchase health informatics software from startups, then the Amazon + Berkshire + JP Morgan health venture is extremely exciting.

In theory this partnership provides a sandbox, and the necessary environment to make progress in healthcare information systems and healthcare system design that previously didn’t exist.

With more than 1 million employees between the companies, Amazon has the potential to develop health information systems they can use for care delivery for their own patients. They can act as be both developer and consumer. This gets around the problems listed above. It provides the sandbox needed for innovation, and a willing buyer (themselves).

Given the geographic distribution of their employees, I suspect software targeted at management of outpatient care, likely through distributed systems such as mobile, with occasional in-person visits is where they will start.

Deja Vu?

However, will the entry (actually the re-entry, many have tried and failed before) of large tech companies into the healthcare records market fix the problems of EHR stagnation?

Not entirely. Instead of large traditional EHR vendors competing against each other to sell monolithic clinical systems, we will have large traditional EHR vendors and traditional ‘big-tech’ companies competing to sell large clinical systems.

It is a stepwise, but not a logarithmic improvement.

Competition may have increased, but we have not fixed the EHR market forces. Large clinical systems are still being sold. The equation remains the same:  High barriers to market entry + High switching costs = Low innovation. Low energy market.

Interoperability of data between clinical systems does not fix this equation.

An unfortunate conclusion

Interoperability between clinical systems is long overdue. Patient data needs to be able to flow between hospitals, clinics, and patients.

However, achieving full interoperability of patient data, and even full interoperability of clinical workflows will still not fix the underlying problems with electronic health record systems.

The space is too hard to enter. The systems are too large to switch between.

Even if big-tech gets into EHRs, they are still operating in a broken marketplace.

There are many other problems interoperability between clinical systems does not address.

  • It doesn’t address how the thousands of health startups all doing ‘their own thing’ actually integrate into clinical systems (as opposed to being wiped out by large vendor systems simply replicating their feature sets in their system).
  • It doesn’t propose a clear pathway on how machine learning algorithms are going to be integrated into clinical care for research, and later for purchase.
  • It falls short of being able to develop really innovative healthcare delivery systems that required very tight integration across multiple clinical services & systems. Not just share data and workflows.

How can we fix this problem?

Ultimately interoperability not just between clinical systems, but interoperability between the functional components within a clinical system is required.

That is the topic for a future post. The current solution appears to:

  1. Decompose the monolith that is the EHR into individual functional components.
  2. Create standards between on how these functional components interact.

Note on header image: although Sketch is a great program for mockup design, as the header image shows, it  clearly lacks decent tools for cropping images.

Subscribe for updates

Get an email update from time to time.