athenahealth CEO and co-founder Jonathan Bush doesn’t mince words when he talks about the health IT marketplace. Bush is a mile-a-minute talker with a quick wit and the kind of enthusiasm that can only be described as "unbridled." He's not afraid to tell you that he thinks the major pre-Internet players in the EHR space—players like Cerner and Epic—​are eventually going to "collapse like big black swans."

Bush, and by extension athenahealth, are the industry champions of cloud technology, betting on the fact that the fundamentals that have up until now made interoperability a distant dream will shift just enough to admit the possibility of an "always-on, totally-reliable national healthcare Internet." His strategy is the kind of good old-fashioned market-centric thinking that you might expect from an entrepreneur (and a Bush, although that may be irrelevant).

That strategy is well underway. A pair of notable 2015 acquisitions have springboarded the company into the mammoth acute care market: Early this month, the company announced that it had purchased WebOMR, Beth Israel Deaconess' cloud-based, stage 2-certified EHR. In January, it announced that it had purchased RazorInsights, another cloud-based EHR provider that specializes in solutions for small, rural hospitals. Both of these purchases serve the same master: an eventual network that spans the entire continuum of care.

It's an ambitious plan, but the key to its execution may be nothing more than semantic: an understanding of the difference between the Washington buzzword "interoperability" and actual interoperation.

"Practically nobody does it"

Bush has historically turned an almost jaundiced eye on the business of running a hospital. In the introduction to his 2014 book "Where Does It Hurt? An Entrepreneur's Guide to Fixing Health Care," Bush laments, "I often get very angry about hospitals. They can be horribly inefficient and resistant to change."

Of course, he doesn't lay the blame at the feet of the hospital when it comes to interoperability, but rather on a set of perversely-misaligned incentives. With admissions declining across the country, and value-based reimbursement still finding its feet, many providers have found themselves living with a foot in both worlds, trying to juggle both fee-for-service and quality-based revenue streams. In an uncertain world, their primary concern has become the generation of referrals. This isn't unconscionable—​it's what they have to do in order to survive, Bush says.

But under current data standards there is, if not an incentive to directly hamper data-sharing efforts, an incentive to not work too hard at them. Aside from the mandate of Meaningful Use (and outside of the population-health efforts of ACOs), why open up data when its possession is a competitive advantage?

"The business model is around creating closed systems so that you can charge as much as you want," Bush told Healthcare Dive. "If there's only one system you can see and connect to electronically, you're going to get the business."

There is also the question of whether hospitals even see sufficient demand for data sharing.

"If you are Kaiser, or an academic center, with a predominantly employed medical staff, there's less of a demand for data sharing across different systems," John Kenagy, senior VP and CIO at Legacy Health, told VentureBeat.

Taking advantage of this model is how Epic and Cerner have gained their success, Bush says—"by creating seamless coordination as long as you don't leave the hospital campus." 

This is where that key semantic difference comes in.

"Interoperability," according to Bush, refers to the federally-mandated standards for healthcare IT. (Bush calls the ONC roadmap "a pile of paper" and "bullshit.") Is your EMR capable of talking to another EMR? The answer is probably "sure." Bush uses the example of TV remote controls, which per SCC regulations have to be interoperable with competitors' systems. Yet despite these regulations, most of US homes have four or five getting lost in the couch cushions—one for the sound system, one for the TV, one for the DVD player, one for the Apple TV. There is no incentive for Samsung or Sony to make it easy-to-use with a Toshiba TV, or vice versa.

The same, Bush says, can be said of healthcare. Most vendors are technically interoperable even today—​they have federally-defined CCD messages and sell a product that allows users on other systems to connect. Epic, for example, constantly faces accusations of being a closed system. Those accusations are not strictly true, and Judy Faulkner, Epic's CEO, rejects them outright. (Kenagy notes that "there's a natural inclination to blame Epic because they're just a big target.")

"100% of our customers that are live with our EHR are also live with our Care Everywhere software built in," Faulkner told VentureBeat. (Care Everywhere is Epic's branded data-sharing interface, customizable to each client.) "We have gone back and retrofitted it into old versions so that every one of our customers can send and receive to others, to anyone who uses industry standards, whether they use Epic software or if they use other vendors' software who also follow the standards."

But those kinds of retrofitted applications require the same kind of inconvenient reprogramming as the TV remotes. While the system may be "interoperable," that isn't the same thing as actual interoperation, Bush says.

"Practically nobody does it because it’s so difficult to use and unsexy in its presentation," Bush said. "No one does it because [vendors] really want it to be hard to get access to their data without buying their product."

A three-tiered evolution

Despite the fact that "structurally in the industry, the primary buyer of tech today is the hospital and the hospital has this inbuilt incentive to make it a lot easier for data to flow around its own facilities than out to other ones," Bush says some vendors are moving towards a more open platform. Of the big players, Bush notes that he believes Cerner is moving away from viewing opacity as a competitive advantage "a little more quickly." In the public arena, global consulting firm PricewaterhouseCoopers has pitched a merger of open-source software and commercial applications for the $11-billion Department of Defense Healthcare Management Systems Modernization Electronic Health Record contract. 

At its lowest level, interoperability is the exchange of data between two systems, which frequently breaks down when one party doesn't control both ends of the exchange. Absent a single standard of intercommunication, various EHRs have evolved along different lines, acquiring, storing and transmitting information in different ways.

"There are multiple standards, and we do most...if not all of the ones that are most frequently used," Faulkner told VentureBeat. "You know the old saying… The good thing about standards is you have so many to choose from."

At its highest level, what Bush calls "level three interoperation," two cloud-based companies build one low-cost connection that seamless integrates all of their customers simultaneously—​like Kayak's application programming interface (API) in United's ticketing system. "You didn't have to go customer by customer to connect to United," Bush explains. "Everyone in the country could automatically drive United's ticketing system through Kayak."

But right now, that holistic integration isn’t possible. No single vendor (or group of vendors) has the network to make that kind of seamless flow possible. And moreover, few can compete with the established giants—Epic and Cerner—​when it comes to sophisticated modules, functionalities and use cases (including athenahealth). If you're a large provider that needs to switch to a new EHR, Bush says, there really isn't an available Internet option.

As a result, the "vast majority" of hospitals in the country operate on non-cloud systems, creating an unassailable barrier to single-click interoperability. "You can't write one API into Epic and all the Epic systems are remote controllable by athenaNet," Bush said. "All the copies of Epic are different, sitting in different places on different servers under the governance of different institutions."

"What we need now is a middle level," Bush said. "We have to get the manufacturers to build a single sign-on—​the ability for athenaNet to jump into a given patient record through a link. Then we have to get hospitals to go through the expense of turning it on, one by one."

Bush says athenaNet has 160,000 current connections to different provider systems. "Probably 10,000 of them are to different copies of just 10 different products," Bush said. "It's the only way to get actual interoperation happening amongst these already-interoperable but pre-Internet systems."

Epic told VentureBeat that it is building more APIs, to be released in 2015, so that other systems can access and pull data from different elements of a patient record.

The cloud solution

The healthcare cloud market in the US and Europe is expected to grow between 10% and 30% by 2020, and in the United States reach $3.5 billion by 2020, up from $903 million in 2013. And although security concerns remain, a HIMSS survey from June showed that hospital IT executives are steadily turning to cloud technology in an effort to lower maintenance costs while meeting expanding tech needs.

The current state of affairs won't be sustainable for much longer, Bush believes. As independent physicians continue to shift towards hospital employment and the enterprise segment grows, the need for seamless acute care integration will become unavoidable. Bush describes it as being "hustled" into the acute care space.

56% of athenahealth's revenue already comes from customers that own and run acute care facilities. The company has over 500 hospitals in its client base—​they just don't use the company for their EHR.

If sheer weight is what allows "pre-Internet" companies like Cerner and Epic to maintain their dominance, athenahealth plans to scale up the continuum of care one rung at a time. "If we're going to accomplish the vision we have in mind, every node of healthcare is going to need to be on athenahealth or on something that connects easily to athenaNet and reliably shares information," Bush said. "We need nursing, we need pharmacy, we need every step of care."

Acute care was the next step, and thus the RazorInsights and the Beth Israel OMR purchases—​two platforms with complementary abilities in the acute-care space. RazorInsights is a multitenant platform whose architecture for serving multiple unrelated hospitals Bush calls "very good." Its workflow for handling complex acute care, Bush says, is very limited. Conversely, WebOMR is capable of administering complex care but is not a multitenant platform.

"The strategy isn't to go right to work on the Beth Israel [EHR]," Bush said. "The goal is to take the acute care administrative workflows out of OMR, build them into Razor, and go from serving only 20- to 50-bed hospitals to serving 50- to 100-bed hospitals. Once we're good at that, we’ll go to serving 200-bed hospitals."

It's an almost elegantly simple plan: Find a development partner and a blueprint, then create a scalable model. athenahealth will rewrite the workflows of the Beth Israel product onto RazorInsights' platform, then move the program onto the athenahealth cloud.

"Not much of WebOMR code that you see today will be in existence in a year," Bush said. "Some will, if whole modules work perfectly and are standalone and can be moved over."

"That's the beauty of cloud architecture," he says. "You can keep objects running for years even though on their own they don’t make a great application."