The new Epic

The tech star lost huge federal contracts, but emerged stronger for it

by

In early 2019, Epic Systems, Dane County’s premier tech company, got walloped. The Department of Veterans Affairs backed out of a $624 million contract for scheduling patient appointments at its 100-plus VA hospitals and awarded it to Epic archrival Cerner Corporation.

Chalk it up to a dramatic reversal in fed thinking, prompted by Cerner securing a huge no-bid contract to provide the full panoply of hospital software — clinical, financial and administrative — and not just patient scheduling.

With nary a protest, Epic stepped aside, pulled the plug on what it described as its “wonderfully successful” pilot program at a veterans’ hospital in Columbus, Ohio, and graciously thanked its VA partners for their help.

Three years later, the Cerner installation of patient software is a mess. A steady drip of negative reports culminated in July with a sharply critical inspector general review of Cerner’s performance at its pilot program at the Mann-Grandstaff VA Medical Center in Spokane, Washington.

Cerner’s software “failed to deliver more than 11,000 orders for requested clinical services,” the inspector general found. This created an “extraordinary risk” as a VA doctor put it. The inspector general said patient safety was in question, and there were “multiple events of patient harm.” (The VA’s deputy secretary disputes the accuracy of the 11,000 figure and defends the department’s commitment to patient safety).

What appears to be a disastrous roll-out from Cerner prompts a question: What if Epic, and not Cerner, had landed what is now estimated as a $16 billion project to update the medical records system of not just the VA, but of the entire Department of Defense? How different would life be if Epic, the biggest vendor of electronic health records (EHRs) in the country and the acknowledged industry leader, hadn’t been booted from the project?

For sure, Epic and Dane County would be rolling in new riches. And almost certainly our veterans and military personnel would be better served medically by Epic’s award-winning technology. But observers say a switch to Epic is highly unlikely, though long term can’t be ruled out entirely.

The surprise is that Epic probably is better off for having lost those contracts, because (a) developing software for the government can be very difficult, as Epic already knows from a failed project with the U.S. Coast Guard, and (b) it allowed Epic staff to work on other projects that have diversified Epic’s business strategy in ways that were all but inconceivable 10 years ago.

The company finds itself well positioned competitively, despite the federal setbacks, to fend off its next formidable challenge: a freshly re-booted Cerner marshaling a new attack on Epic’s market dominance.

“Yeah, $624 million is a really big deal [for Epic to lose], and $16 billion would have been a helluva deal [for Epic to win],” says veteran tech investor John Neis of Madison-based Venture Investors. “While Cerner’s stumble out of the gate seems like it could be an opportunity for Epic to step in, my guess is that it isn’t.”

Simply put, Neis says the feds have already poured billions into the troubled Cerner contract, and starting over with Epic would mean writing off a huge expenditure of time and money.

It didn’t hurt that Cerner had friends in high places. As Politico reported, Cerner’s VA project launched in 2017 as an unusual no-bid contract championed by the Trump Administration’s Jared Kushner. The no-bid rationale was this: Putting veterans on the same medical software system as active duty personnel was the smart thing to do for the long-term continuity of their care.

But the project has been mired in serious problems from the get-go. VA physicians told Politico that terrified clinicians were quitting because they feared the malfunctioning Cerner software “would hurt patients, or even kill them” by scrambling their prescription orders and referrals.

Cerner, which the business-software purveyor Oracle bought recently for an eye-popping $28.4 billion, has lost its way and is suffering from “internal decay,” says health-tech commentator Brendan Keeler, whose resume includes six years at Epic.

“Cerner is losing ground to Epic for the premium customers,” he says. And only winning new contracts with “price-sensitive” buyers who flinch at Epic’s premium rates.

This has been going on for years, Keeler adds. And he sees it as an open question if Oracle can deliver on its promise to turn Cerner around. He’s not alone in that assessment. Oracle Cerner did not respond to a request for comment.

Epic and Cerner offer sharply conflicting business models that each tell a fascinating tech story of their own.

Epic Systems was founded in Madison in 1979 by Judy Faulkner, a sui generis computer-programming wizard who at age 79 remains firmly in control of the privately held Epic to this day.

Epic has 12,500 employees worldwide, about 11,000 of them clustered at the Verona headquarters outside of Madison. Revenue reached $3.8 billion in 2021, $2 billion less than #2 EHR vendor Cerner. (By its bigger hospital market share, Epic is the industry leader). Based in Kansas City, Missouri, the publicly traded Cerner has a much bigger international operation and far more employees at 27,000. But in the end Cerner never caught up to Epic, and was laying people off to fatten the bottom line to impress Wall Street when Oracle made its move.

Observers agree that Epic’s great strength is how it built its software modules — everything from outpatient scheduling, to processing lab results, to managing emergency room operations, to handling insurance claims, to scheduling cancer treatments, and more — from the bottom up, so operationally they integrate well.

In contrast, Cerner has bet heavily on acquiring health IT companies and putting them under the Cerner umbrella. Notably this included buying Siemens’ Health Services unit from the German conglomerate in 2015 for $1.3 billion. An early assessment said the takeover got off to “a rough start.” The accounting functions of both the Cerner and Siemens software “rank low” on functionality and development, the newsletter EHR Intelligence advised.

“Cerner was really just a bunch of companies bolted together,” says Mark Bakken, managing partner of Madison-based HealthX Ventures, which invests in digital health startups. “Epic was slowly beating Cerner” with its superior software, he adds.

Keeler concurs: “There’s a huge cost to mergers and acquisitions. People horribly underrate how much effort it takes to fit together different corporate cultures, different data models.”

Faulkner would not be one of these people. Her 10 commandments for guiding Epic’s operation begin with three unequivocal dictates: 1. Do not go public. 2. Do not be acquired. 3. Software must work.

All this is baked into the Epic DNA, including being “fanatically focused on customer success,” as an Epic executive once summed up the company mission.

Today, Epic reports that 253 million Americans have an Epic account. Some 2,100 hospitals use the Epic platform, as do 60,000 medical clinics. Epic says more than 400,000 physicians write their notes in the company’s software.

In the end, Cerner’s record was mixed at best. Once the predator in corporate takeovers, it became the prey when Oracle, the ubiquitous business systems giant, executed a takeover that closed in June. Cerner is now Oracle Cerner, and the man in charge, Silicon Valley legend Larry Ellison, became the latest tech/Fortune 500 giant to promise he will transform American healthcare.

Reputed to be among the 10 richest people on Earth, Ellison was once dissed by a Canadian business writer who said “even in an industry dominated by barbarians, he has earned a reputation as a first-class SOB.”

True or not, Ellison, at age 78, is certainly not lacking ambition or confidence in his plans for the new Oracle Cerner. He is pledging to put together nothing less than “a unified national health record database” that will yield improved medical care for all while lowering medical costs.

That U.S. healthcare needs a good shaking up cannot be denied. We burn through way more money on medical care than other industrialized countries. On an annual per capita basis, U.S. expenditures are more than double the average: $12,318 to $5,829. And while American medicine can be spectacularly good, the overall results are surprisingly meh — the quality of U.S. healthcare ranked last in a study of the 11 richest countries.

Electronic health records were supposed to help change those outcomes.

Replacing the bulging files of paper documents with electronic health records was heralded as a breakthrough for the friction-free sharing of patient information among physicians, hospitals, insurers and patients themselves. Not to mention empowering researchers to use “Big Data” analytical tools to reveal the best treatment outcomes for most any disease.

This was the vision President Barack Obama embraced in 2009 when he rolled out what proved to be an approximate $36 billion federal subsidy for EHR implementation. His promise to the American Medical Association that the electronic records would lower administrative costs, improve care and simplify the work of doctors...well, to put it charitably, it’s still mostly an aspiration. And Ellison knows it.

In June, the mogul laid out his own expansive vision of how Oracle Cerner would usher in a new world of digital health. The key would be solving what he describes as the problem of “healthcare data fragmentation.” Each hospital having its own information system means “there are thousands of them throughout the United States.”

Ellison sees tremendous problems emanating from the fragmentation. As he explained it, most of us have health data scattered over one- to two-dozen separate databases to account for every medical provider we’ve seen over the years. How then does an emergency room doctor quickly get a read on our health history? On a far broader scale, how do public health officials assess the breadth and success of various pandemic treatments?

“We’re going to solve this problem by putting a unified national health record database on top of all of these thousands of separate hospital databases,” Ellison pledged.

The mogul makes it sound easy, but as one health IT correspondent gently put it, Ellison was “light on details” on how other EHR vendors — not just Epic, but smaller ones including Meditech, Allscripts and CPSI — would work with Oracle Cerner.

Perhaps more to the point, why would they want to support what clearly sounds like Larry Ellison’s big plan to put his company in the driver’s seat of a new national health record database?

One presumes this question/threat has registered with Epic, but the company never talks about competitors as competitors. Actions, though, speak loudly. And Epic has been very busy in recent years building the infrastructure that Ellison now dreams about. This speaks volumes about the future of Epic…and the challenge that Oracle Cerner faces.

In January 2012, Steve Dickmann, Epic’s chief administrative officer, provided a rare insider’s look at Epic’s thinking when he addressed the Wisconsin Innovation Network. What was Epic’s market? He explained it was almost exclusively large medical groups, children’s hospitals (“because we like children”) and research institutions.

Put another way, Epic concentrated on selling software to the biggest, most prestigious and affluent organizations: Mayo Clinic, Johns Hopkins Hospital, Massachusetts General, the Kaiser-Permanente network, and Cleveland Clinic among the iconic biggies.

Ten years later, it’s a different story. Now Epic casts a much wider net, selling software to small community hospitals, medium-sized physician practices, and overseas markets in 16 countries.

That’s in addition to partnering with Walmart Health to help provide the retailer’s customers everything from urgent care to dentistry to X-rays to optometry “at transparent prices, regardless of insurance status,” as Walmart put it in a press release.

Epic software also undergirds the nearly 10,000 CVS pharmacies. Closer to home, Epic has paired off with Madison’s Exact Sciences, another bright star in Dane County’s health tech constellation. The cancer-test developer announced in 2018 that Epic would become “a core component” of its information system.

The centerpiece of the new Epic is unequivocally Cosmos, its expansive database of 5.7 billion (and ever climbing) anonymized patient encounters as detailed on Epic software. Not all Epic clients participate. (Some larger health systems are creating their own DBs.) But for the ones who do share, Cosmos becomes a huge resource for physicians to plumb successful treatment protocols tailored to the specific characteristics of their patients.

This represents the extraordinary power of what’s called “the network effect” — that every scrap of new data added to a database can make it stronger, deeper and more valuable.

And, the Cosmos data is proving a proverbial gold mine for research. In August, for instance, University of Maryland researchers analyzing the Cosmos database of 315,000 emergency room cases for drug overdoses discovered how infrequently OD victims are screened for fentanyl: Over five years, only 5 percent were tested for the illicit drug, but of that 5 percent more than 40 percent tested positive.

What’s notable is that for years Epic was regularly accused of impeding data sharing with rival EHR systems. Epic countered it was being vigilant about data security and patient privacy. Now the company champions itself for playing well with others. Of the approximately 11 million medical records it exchanges daily, Epic says almost half originate from rival EHR vendors.

The normally hush-hush Epic offered a look under the Cosmos hood in late September when four enthusiastic members of the Cosmos team talked about their work at a meeting of the Wisconsin Technology Council. Two of them described it as “the coolest” research imaginable. At that point the ever-growing Cosmos database included the health records of 163 million patients from more than 170 Epic-connected health providers.

“To our knowledge, it is the largest medical record database in the U.S. and quite possibly the world,” Epic’s lead data scientist Caleb Cox said.

When asked if Epic would collaborate with Oracle Cerner on a common database, the quartet said it was too early to say. They noted there were just too many unanswered questions. Was the Oracle Cerner data reliable? Had it been properly “de-identified” so that patient privacy was protected? Would it mesh with Epic data? And just as important: Would Oracle Cerner accept Epic’s foundational requirement that while Cosmos data would be shared with researchers, it would not be sold commercially?

They didn’t say as much, but these are huge unanswered questions.

Meanwhile, how Oracle Cerner tackles its software problems while simultaneously challenging Epic’s marketplace dominance is an equally important story in charting the future of electronic health records.

John Neis, the venture investor, says Oracle’s purchase of Cerner is a plausible way for Oracle to enter the fiendishly complicated health-tech field. “It certainly gives Cerner more resources,” he says, adding that Oracle Cerner has to prove “it can create the kind of synergies that work for these kinds of mergers. This is not a business space you waltz into from other sectors.”

As Neis points out, healthcare is tightly regulated, and its economics are tangled in the confusing world of third-party payers — Medicare, Medicaid, health insurers and employers. This means the people receiving the care usually aren’t the people paying for the care (co-pays aside). “It is an extremely complex market,” says Neis.

And there is a strong feeling among healthcare insiders that Epic knows its niche better than anyone else.

Mark Bakken of HealthX credits Faulkner for making the right call to stay private. “Epic plays the long game,” he says. “They make investments that might not pay off for five to 10 years, while Wall Street expects quarterly or annual results.” Fail to meet earnings projections, and you’re kicked to the curb, Bakken adds.

He suggests that Larry Ellison’s takeover of Cerner — and the uncertainty it introduces over what comes next — may prompt some Cerner customers to consider jumping to Epic when system upgrade time comes.

Tech commentator Brendan Keeler makes the same point. “When you buy million-dollar enterprise software you seek stability,” he says of the costly EHR systems. “When you look at Cerner the last five years, you don’t see it. Instead you see Oracle purchasing Cerner without a clear plan to improve the core product.”

That’s par for the Ellison course, says Tom Erickson, founding director of UW-Madison’s School of Computer, Data & Information Sciences. A former Boston tech executive, Erickson says he regularly saw Oracle buy a company, send out its sales force to sell Oracle business products to a new circle of customers…and that was pretty much it.

“It was almost never a situation of Oracle saying ‘We’re going to do something innovative and sustain this company as a leader,’” Erickson recalls. Instead, “they kind of retreated into what I call ‘the Oracle abyss.’”

As for Oracle Cerner’s ambitious plans to create a national database, he sees it simply as an Ellison ploy to undercut Epic’s dominance. “Cosmos is more comprehensive than anything he can put together,” Erickson says. “Quite frankly, if I were Epic I would just ignore it. You’ve got the market leadership. People will come to you for the data.”

Let’s end with the starting point of this story: How Epic losing a $16 billion Department of Defense contract and the smaller VA contract became the best thing that could have happened to it.

“There was some reflection and realization at Epic that maybe they were lucky to not be the dog that caught the car,” says a cautious tech-industry executive who asked to be quoted anonymously. The scale of the defense department project and its specialized requirements were just too great for even Epic to undertake. It would have drained focus from everything else Epic was successfully doing to diversify and expand its reach, the source says.

Erickson agrees. He tells the story of a company he ran winning a huge contract that wound up “oversizing” his operation. “We spent so much time and effort trying to make it work,” he recalls. “But it compromised our other customers and our workforce.”

That’s the problem with big federal contracts, he says. They are so large and the government so insistent on enforcing its many arcane protocols that the contract can turn into a deadly quagmire for an unprepared company.

How odd it is to conclude by saying that Epic won big by losing big. But that’s the situation for now until we see what comes next from Larry Ellison and Oracle Cerner.