This article was originally written in 2015 and links were updated in 2021

---

The headline couldn’t have been more focused and directed against a single EHR software vendor. It screamed in bold, capital letters ‒ EPIC FAIL ‒ but then for some reason, backpedaled with the subheading.

Digitizing America’s medical records was supposed to help patients and save money. Why hasn’t that happened?

Since it’s original publication date (October 2015), the headline has been changed and now reads:

We’ve Spent Billions to Fix Our Medical Records, and They’re Still a Mess. Here’s Why.

It’s a reasonable question for the industry, of course, which is the way the subheading phrases it, but that’s not the responsibility of a single commercial software vendor, regardless of their size or corporate structure.

To date, the MotherJones article by Patrick Caldwell has generated over 260 comments, 1,100 tweets and 3,400 shares on Facebook. Most of the comments are supportive of the Epic Fail premise, but the primary argument ‒ as stated in the 4th paragraph of the 2,300+ word essay ‒ is technically wrong.

But instead of ushering in a new age of secure and easily accessible medical files, Epic has helped create a fragmented system that leaves doctors unable to trade information across practices or hospitals.

Epic has certainly contributed to the challenges of sharing patient data ‒ but to blame them for lack of interop is about the same as blaming Apple for fragmenting the mobile world with iOS. In a nutshell, software is only designed, marketed and successfully sold when it does what customers want. Nowhere is that more true than on the expensive enterprise side of software manufacturing. It only sells to people willing (and able) to buy it ‒ often by the billions of dollars.

To be clear ‒ I’m not an Epic fan anymore than I’m an Apple fan ‒ but there are parallels. The biggest, easiest distinction, of course, is that Apple creates the whole consumers experience (hardware and software), and Epic is exclusive to enterprise software only.

By about paragraph 11 we finally get to the argument that’s at the heart of the whole HITECH Act ‒ which includes Epic and about 400 other software vendors that manufacture and sell electronic health record software.

As it stands, she [Julia Adler-Milstein, a University of Michigan researcher who studies health care IT] says, using Epic is easier than trying to piece together better options from various software vendors. On top of that, Epic will tailor each installation on-site to a customer’s specific needs. What it doesn’t have—and ditto systems created by competitors Cerner and Meditech, the other bigwigs in EHR—is a framework to connect to other facilities using competing EHR systems.

It’s perfect staging for the appearance in the next paragraph of healthcare’s word of the year ‒ interoperability. For such a lengthy article ‒ it’s surprising that the interop word only appears about 5 times, but  that’s really the point of the whole piece ‒ and the biggest charge against Epic.

Without defending Epic, here are some “interop” statistics that Epic President Car Dvorak gave to ONC’s Health IT Policy Committee last year ‒ under oath.  

Epic has strong support for “standard HL7 traditional interfaces,” as well as other transaction types such as NCPDP and ANSI X12, enabling “about 20 billion data transactions per year, over 12,000 different interfaces across our 320 customers, to about 600 other vendor systems ‒ including 88 public health agencies, 18 research societies, 51 immunization registries across 46 states and 17 research registries. As for meaningful use, specifically we support the Direct protocol, which is reasonable for planned transitions of care. That “is clearly better than nothing,” he said, but it’s just as clearly inadequate for unplanned transitions such as ED visits. For those, “we support Healtheway exchange standards ‒ the Connect or NwHIN standards; we think both are vital for solving interoperability in the country,” said Dvorak. More generally, he pointed out that Epic’s Care Everywhere technology enables the exchange of 480,000 consolidated clinical document architecture, or C-CDA, documents each month with products built by other vendors – and that number is on the rise: more than double the amount that changed hands as of January 2014. That exchange ecosystem “includes about 900 hospitals, 20,000 clinics, with another 85 hospitals and 3,000 clinics coming online in the coming year,” said Dvorak.

We connect to 26 other vendors systems, 21 HIEs, 29 HISPs and 28 eHealth Exchange members,” with 20 more coming online soon” he said.

Epic Defends Interoperability Bona Fides by Mike Miliard / Editor, Healthcare IT News

We should demand this type of interop accountability from all the EHR vendors. To date, I’ve only seen it from one ‒ Epic.

Also lost in any discussion is the cost by software manufacturers to actually build the gateways into competing software systems. They can do that and they often do ‒ for a fee ‒ but like every other industry, software engineers expect to be paid for their work. Sorry, but just because it’s healthcare, doesn’t entitle industry observers to “free software.”

“We have what I would call a ‘loosely-defined’ standard that’s subject to interpretation. At the end of the day, for two systems to figure out how to convert and communicate, there’s some level of manpower, hours and efforts that are required. If you’re a for-profit organization, this results in having to charge somebody for their time. There’s a cost for interoperability, and that’s where we get a lot of talk.” Mark Janiszewski ‒ EVP of Product Management for Greenway in The Vendor’s View of Interoperability ‒ Becker’s Health IT and CIO Review

A survey in 2014 suggested that Epic owns about 20% of the inpatient market and about 20% of the ambulatory market. That’s an important distinction because the fragmented market often referenced is really the ambulatory one ‒ where there are newer entrants that tend to have relatively small percentages of the market but are among the most vocal when it comes to screaming about “interoperability.” It’s also easy to make that noise and attract free publicity, but it doesn’t solve any problems.  

Here are two charts highlighting that exact point from a firm called Software Advice. They published their EHR Meaningful Use Market Share for 2014 last year and included two helpful charts. There are other studies, of course, but most of the really current ones are by firms that charge big bucks and then heavily restrict public disclosure.

For our purposes, however, it doesn’t matter that much because the market just isn’t that volatile ‒ especially at the top. Here’s the Software Advice chart for the ambulatory market.

Software Advice was quick to highlight the visibly obvious here.

What’s surprising is how fragmented the market continues to be: five of the market’s top 10 vendors boast a market share of less than 3 percent each, and nearly 40 percent of the market uses an EHR from a vendor that falls outside of the top 10. 

Here’s the Software Advice chart for the inpatient market (Hospitals) that MotherJones could also have included ‒ but elected not to. 

Software Advice was equally quick to highlight the obvious here as well (bold emphasis mine).

When it comes to inpatient hospital attestations, the biggest vendors show greater market dominance: the top 10 account for over 90 percent of the market, while the top three alone account for over half of it.

While it may be popular, the case against any single vendor for the industries lack of interoperability is just technically wrong. The only real mistake here ‒ and it’s a big one ‒ was the original legislation called HITECH ‒ which paid for EHR adoption, but didn’t insist on standards for sharing information. Commercial software companies aren’t responsible for Government failures and by comparison, no one looks at the sizable lead that Android has in the mobile world and blames Google for failing to “interoperate” with Apple.

Commercial software in the healthcare industry isn’t automagically different than commercial software in other industries except that our Government did pay handsomely (to the tune of almost $30 billion) to accelerate the adoption of EHR software. If, as a part of the HITECH Act, the Government had wanted “interoperability,” it could have ‒ and should have mandated it. MotherJones finally acknowledges this deep in the article.

Epic is not the only barrier to a seamless medical records system. Thanks to legislative maneuvering by former Rep. Ron Paul (R-Texas) in 1999, the federal government can’t fund any sort of system with unique health care identification numbers. (Paul saw individual medical IDs as further creep of Big Brother). Social Security numbers aren’t a good fill-in; they’re not on insurance cards, and in April Obama signed a bill that will strike them from Medicare cards in order to reduce identity theft.

Large and influential healthcare IT organizations like CHIME and HIMSS have all recognized the enormous handicap of trying to build interoperability without this critical component. They are also joined by influential CIO’s like Michael Blum at UCSF who called the Congressional ban on establishing a universal patient identifier “the biggest single failure in the history of health IT legislation.” [page 189 of The Digital Doctor by Robert Wachter, MD]

Everyone’s entitled to an opinion of course and it’s certainly popular to blame Epic for all that ails in getting EHR software to share patient information, but almost all the reasons for the blame revolve around safety and quality. The healthcare system we have has been optimized for revenue and profits ‒ not safety and quality ‒ so neither the system or Epic is “broken.” It was actually designed this way and we need to change the design. Hating Epic Systems Corporation may be popular ‒ and it may even feel good ‒ but it’s totally misdirected and won’t change one byte of their software.

The lack of data interoperability in healthcare continues to plague and haunt the entire industry. Much of the challenge falls squarely in the realm of Electronic Health Record (EHR) software, but EHR software is by no means the only category where this challenge is directly affecting patient lives. I ‒ along with countless others ‒ have written extensively about this topic. Last year I did a 5‒part series on just ‘interoperability’ which I published on Forbes. One piece highlighted the importance and direct effect of lack of interoperability from the patient perspective: The View of Digital Health From an ‘Engaged Patient’

I wish I could say things have improved since ‒ they really haven’t ‒ nor are they likely to improve soon. We remain stuck in competing commercial interests, political agendas and very real technical challenges on how to securely share even the most basic health information. While the debate and agendas continues ‒ this year alone has seen the health data breach of 91 million Americans through just 2 high‒profile cases ‒ Anthem and Premera ‒ and that was just in the first quarter. Whatever the privacy argument is against sharing data ‒ it’s all but irrelevant when more than 90 million Americans have their health data breached. That’s effectively 28% of the U.S. population.

The result of all this is sadly theatrical ‒ a kind of Kabuki dance ‒ and reflects the broader inability of many to grasp the true meaning of “patient‒centered” thinking. It’s excruciatingly painful to watch ‒ let alone live through as patients.

The topic of ‘interoperability’ kept percolating throughout last year and finally reached the hallowed halls of Congress. Outraged that the Government had spent almost $30 billion dollars on “infrastructure” software that locked data in proprietary silos, the Government demanded some accountability through the Office of the National Coordinator (ONC). That report was delivered by ONC to Congress on Friday with this odd ‒ uniquely Government title. Report on Health Information Blocking  

Pardon my French, but what the hell is ‘information blocking?’ As a software engineer, I’ve been around software most of my professional career and I had never heard that phrase before ‒ let alone as applied to software interoperability. Sadder still, it also highlights a real naiveté by ONC and Congress on the actual business of software manufacturing. If it’s not naiveté it’s certainly very theatrical.

The report itself is about 39 pages (including 6 pages of appendixes). It’s a painful attempt at laying blame for the lack of interoperability on the big, bad Independent Software Vendors (ISV’s). Even the first hurdle ‒ to define “information blocking” ‒ is an enormous challenge because it attempts to fabricate a definition from thin air. The report does acknowledge the scale and scope of this enormous difficulty outright.

The term ‘information blocking’ presents significant definitional challenges. Many actions that prevent information from being exchanged may be inadvertent, resulting primarily from economic, technological, and practical challenges that have long prevented widespread and effective information sharing. Further, even conscious decisions that prevent information exchange may be motivated by and advance important interests, such as protecting patient safety, that further the potential to improve health and health care. Finally, it is important to acknowledge that certain constraints on the exchange of electronic health information are appropriate and necessary to comply with state and federal privacy laws; this is not considered information blocking.

Undaunted by these challenges, ONC went ahead and fabricated a definition anyway.

Information blocking occurs when persons or entities knowingly and unreasonably interfere with the exchange or use of electronic health information.

With this broadest of definitions, I asked two healthcare legal minds I know for clarification on that gaping hole of a legal phrase called ‘unreasonable.’

Can U define “unreasonable” in @ONC_HealthIT “info blocking” report 2 Congress? http://t.co/ij20SPcjnX cc: @VinceKuraitis @healthblawg — Dan Munro (@danmunro) April 10, 2015

I fully expected the reply.

.@danmunro Agree, ONC language re: “knowingly & unreasonably” is unnecessarily high burden of proof. @ONC_HealthIT @healthblawg — Vince Kuraitis (@VinceKuraitis) April 10, 2015

It’s not just unnecessarily high. Given all the challenges listed outright by ONC ‒ it’s virtually impossible to prove ‘knowingly and unreasonably.’  Assuming you could even level the charge ‒ who would adjudicate the case ‒ and the resulting appeal ‒ at what cost ‒ and to whom? ONC knows this ‒ and acknowledged it indirectly further in their report.

ONC-ACB surveillance activities and other feedback from the field show that although certified health IT is often conformant with the criteria to which it was certified, there is still a substantial amount of permissible variability in the underlying required standards, unique clinical workflow implementations, and numerous types of interfaces to connect multiple systems. This variability has contributed to information sharing challenges and also creates opportunities for developers or health IT implementers to erect unnecessary technical barriers to interoperability and electronic health information exchange.

Commercial software ‒ specific to the healthcare industry ‒ has taken years to mature and there are distinctly separate categories of EHR software. Two of the largest categories are Ambulatory and Inpatient. One could argue that the first and most successful vendor in this unique category of software called EHR is Epic Systems ‒ founded by Judy Faulkner (1979) and still privately held. Here’s a relatively recent breakdown of market share ‒ by vendor ‒ for both categories.

With a firm hold on 20% of both markets, Epic is the Apple (or Microsoft) of EHR software except that it remains privately held. Whether a software manufacturer is privately held ‒ or traded on the public market ‒ makes little difference to the way the business operates. The reality is that most of these EHR systems started (and largely remain) as enterprise billing engines ‒ that now include a fair amount of clinical data ‒ stored in proprietary formats. Anyone ‒ including ONC ‒ who argues that Independent Software Vendors (ISV’s) are obligated to build software connections (or API’s) into competing software solutions ‒ for free ‒ clearly don’t understand the way revenue is captured and reported to the IRS. They also don’t understand that ISV’s are beholden to shareholders as much (if not more) than their customers and their customers are not patients. They’re alignment to shareholders and customers is purely economic. We’ve supported this model to the tune of over 400 ISV’s that now offer “certified” EHR software to some customers ‒ almost entirely in proprietary formats based on each vendor’s commercial interests.

All of which begs the question. Absent any National Standards often used in other industries as a way to effectively neutralize all of these competing interests ‒ who is truly to blame? Congress for actively “blocking” the development of National Standards ‒ or ONC for lacking the technical ability to police an entire industry for “information blocking?”

But that’s also a serious charge. How is Congress actually ‘blocking’ the development of National Standards in healthcare? For insight into that exact dilemma ‒ it’s important to link the release of ONC’s report with another headline from last week.

Congress Continues to Block Nationwide Unique Patient Identifier

There’s that word again ‒ ‘block’ ‒ only this time aimed squarely at Congress ‒ not the EHR ISV community. According to the article, the resulting damage from this one ‘block’ has serious and direct patient consequences.

According to a survey of healthcare CIOs conducted by the College of Healthcare Information Management Executives, error rates due to patient mismatching averaged eight percent and ranged up to 20 percent.  In addition, 19 percent of the respondents indicated that their hospitals had experienced an adverse event during the course of the year due to a patient information mismatch.

The history to this one Congressional ‘block’ goes back to the passage of HIPAA and I wrote about that history as a part of that 5‒part series on interoperability just over one year ago.

Who Stole U.S. Healthcare Interop?

Long forgotten is a similar dilemma in a vastly different industry ‒ motor vehicle manufacturing. For decades, car companies issued their own vehicle identification number. The resulting mess made it virtually impossible for consumers (let alone law enforcement) to tell if a car they were about to buy had been stolen, was a lemon, part of a manufacturer recall or if it had serious damage (manmade or otherwise) somewhere in its history.

A vehicle identification number, commonly abbreviated to VIN, is a unique code including a serial number, used by the automotive industry to identify individual motor vehicles, towed vehicles, motorcycles, scooters and mopeds as defined in ISO 3833. VINs were first used in 1954. From 1954 to 1981, there was no accepted standard for these numbers, so different manufacturers used different formats. In 1981, the National Highway Traffic Safety Administration of the United States standardized the format.It required all over-the-road-vehicles sold to contain a 17-character VIN.  VIN [Wikipedia]

Today, that VIN is stamped on every major part of every vehicle manufactured in a way that makes it easy for consumers and law enforcement to check on the status of any car (or car part) easily and quickly ‒ nationally.

So, there’s really no mystery to this Kabuki dance between ONC, ISV’s and Congress. ONC is firmly stuck between enormous commercial interests and a clearly recalcitrant Congress to simply authorize the development of any National  Standard for healthcare IT (either NPI or more broadly for data interoperability). ISV’s are, of course, competing aggressively for revenue and market share in any way they can. Patients are the collateral damage in these competing interests.

It’s painful to watch and it’s painful to realize the complete and utter disregard by all for basic patient safety ‒ but that is the healthcare system we have built. No, it’s not broken ‒ it’s just been optimized for revenue and profits ‒ not safety and quality. Congress simply refuses to reign in the commercial interests of the ISV’s and ONC is simply an effective buffer to hide their clear ‒ financially motivated ‒ reluctance.

Everyone’s well meaning, of course. Everyone’s doing “the best they can”, but it’s very much like the quote from the recent Ken Burns biography of cancer. With a little substitution we can easily apply the title of the second installment of that epic series directly to the healthcare interoperability dilemma.

The important thing is that the viral National Patient ID theory was not wrong. The environmental Meaningful Use theory was not wrong. The hereditary Vendor API theory was not wrong – they were just insufficient. It was like the blind man and the elephant. They were catching parts of the whole and then all of sudden – if you stepped back – you saw the whole elephant. Siddhartha Mukherjee, MD ‒ Top 10 Quotes From the Ken Burns Documentary: ‘Emperor of All Maladies’

Unlike the war on cancer, we haven’t seen ‒ nor do we appear to be looking for ‒ the whole elephant.

---

This article first appeared on HIT Consultant