EHR Spending Continues, But Jury Still Out on ROI

As healthcare leaders continue to pour money into electronic health record systems, many (but not all) find that their return on investment remains elusive.

This article appears in the January/February 2014 issue of HealthLeaders magazine.

Of all the capital investments made by healthcare leaders, none appears more unproved than the move from paper to electronic health records. In the past decade, billions in public and private money, boosted by nearly $20 billion in federal incentives over the past three years, have been poured into EHRs. What has that money bought?

On Capitol Hill last summer, with healthcare's cost curve just barely starting to bend, patience was in short supply. Concerns about lack of interoperability among purchased EHRs mounted. Lawmakers grilled providers and vendors, seeking hard numbers on EHR return on investment—numbers that remain in short supply.

For their part, most healthcare leaders maintain the investment is worth it, pointing to a vast array of anecdotes and scattered studies showing the value of the EHR. In a few relatively isolated cases, they say, tangible ROI is achievable within five years. And yet others caution that true ROI requires even more patience and say the bulk of the rewards can take up to 10 years to achieve.

Still, disagreement persists on how the healthcare industry is going about the move to EHRs. Critics say the existing generation of EHRs, built for billing in a fee-for-service world, is ill-suited to delivering value-based coordination across the continuum of care. Patients, newly empowered by a plethora of apps that help them manage their own care, still struggle to access their EHR-based records, stymied by technological limitations and the same lack of interoperability that frustrates clinicians. And the federal government, just now deploying stage 2 of meaningful use, must still craft guidelines that move from basic interoperability to measures that truly impact patient outcomes.

Thus, instead of the nation reaching a goal next year—first set out by President George W. Bush in 2004—of every American having an interoperable electronic health record, it could take as long as another decade before healthcare can hang its mission accomplished banner on the EHR.

"If you asked me the question in 2007, as we were installing an electronic medical record and starting a health information exchange, 'How long do you think before you will have a functional, smooth health information system?' I might have been bold enough to say, 'Well, 2012,' " says William Park, MD, senior general surgeon and former chief medical officer at North Hawaii Community Hospital, a 20-staffed-bed rural hospital in Waimea. The facility, which reported 2012 net patient service revenue of $47 million, is a participant in the Hawai'i Island Beacon Community, one of 17 such nationwide programs developing advanced health information technology through federal funding.

"But at this point I would say I'm skeptical about 2014, and certainly not courageous enough to say 2016," Park says. "I think it's going to take a decade before we have a truly intelligent, connected system that will benefit patient healthcare. We have bits and pieces of success. I think we have little silos of effort. There seems to be some general understanding about where we need to get to, but it varies depending on who you speak to."

A big obstacle for Park and others is the lack of a national health information exchange infrastructure. This project is being worked on actively through a variety of public and private efforts. But for now, the biggest EHR success stories usually involve the efforts of individual providers, and of those, only a few have already crossed the chasm between older fee-for-service payment models and newer value-based-care payment models.

Hunting for real ROI

One such provider is Sentara Healthcare, the Norfolk, Va.–based system of 11 acute care hospitals that reported 2012 net patient service revenue of $2.8 billion. From the outset of its efforts to implement its eCare EHR at seven hospitals in the Hampton Roads area of Virginia from 2008 to 2010, Sentara leaders have had return on investment in mind.

The key to ROI is to start with a baseline and "redesign your thought system and processes to leverage the value of electronic records, or any IT solution," says Howard Kern, president and chief operating officer of Sentara.

Kern draws inspiration from information systems first used more than a decade ago to improve manufacturing and business processes in the auto industry. An oft-cited 1990 Harvard Business Review article, "Don't Automate, Obliterate," described how Ford first implemented information systems and ended up with an accounts payable department head count reduction of 20%, but when it redesigned the department's processes around IT it was able to reduce head count by 75%.

"Ford simply overlaid the IT system on top of their existing process and automated an inefficient process," Kern says.

Thus, when Sentara began its electronic record work with Epic Systems in 2007, it started 18 process redesign teams, focusing on everything from clinical delivery to medical records.

"We had clinical leaders and management leaders working side by side with Epic technical staff before we implemented anything," he says. "We had done complete redesigns around each of these major areas, so when we implemented the system, it was with a design that was built to drive value in the healthcare environment, both in inpatient and outpatient environments."

Although Sentara organized separate teams for the medical office and its hospitals, leaders also designed the system so that processes would look and feel the same in the inpatient and the outpatient environments, Kern says.

"The value there is that we got much more efficiency and much more value from the way physicians and nurses were able to interact with the system, and from continuity of care with respect to how the patient experience would be with the system."

To measure ROI, the most important metric watched was inpatient length of stay. Because Sentara has largely left fee-for-service payment behind, the shorter the length of stay, the more money the system saves. Kern says Sentara has a combination of DRG and per case reimbursement. "Very little is fee-for-service in our world anymore, and we have a fair amount capitated in the sense that we have our own health plan, and that's probably about 20% of our business," he says.

"We were able to document in a way directly attributable to the electronic medical record how length of stay was reduced in a range of areas," Kern says. In the first year, savings were difficult to measure. "We had to sort things out and make sure we had all our systems stabilized," he says.

But after the second year operating the redesigned system, Sentara has seen a $54 million annual cost savings over operations across the seven hospitals, compared to operations before implementing Epic. Inpatient length of stay only accounted for $15.5 million of that savings. According to Sentara, the rest was achieved through increased outpatient procedures, increased unit efficiency and retention of RNs, and reduced costs associated with transcription services, medical record supplies, medical records personnel, health plans, and IT maintenance.

On the investment side, Sentara's total equipment and systems cost was around $180 million, Kern says. "Our total going-in cost in terms of implementation, training, and support for the back office function from the old system totaled up to about $250 million overall," he says.

Thus, at the current rate of operating cost reductions, Epic will have paid for itself completely at Sentara before the company reaches the end of the fifth year of the $54 million annual savings—2014 at the latest.

"If we didn't do the baseline process redesign and really made sure we were harvesting the value, two things happen," Kern says. "One is you don't really redesign, and there's no pressure then, or burning platform, to redesign the processes to get to that value. Two, you're layering on a lot of cost, from an expensive electronic record, and you don't have anything to show for it, other than the depreciation, maintenance, and licensing fees. Whether you're cost-reimbursed or reimbursed per case or capitation, that's just not good to do from a good business and clinical operation perspective."

The July 2013 HealthLeaders Media Intelligence Report found that 16% of healthcare leaders saw no clinical quality improvement from EHRs, and the leaders interviewed for this story believe that those systems must not have implemented the kind of process redesign that Kern and others speak of.

"Just to give you an example, we avoided 117,400 potential medical medication errors due to the medication bar coding element of the electronic medical record," Kern says. New processes and the EHR reduced medication verification from a 59-minute process to a 5-minute process, and overall medication administration time dropped from 132 minutes to 38 minutes.

Under the pre-Epic system, "we had to actually photocopy the physician's written order, even after the nurse entered it into the nurse order entry system with the old system that we had," Kern says. "The nurse had to then take a photocopy of that physician's written order, send it to the pharmacy, and the pharmacists had to enter the order after reviewing the physician's written order.

"When we redesigned the system around the physician order entry, all of a sudden all that went away, because the physician's order is already in the system and the pharmacist doesn't even have to verify it. It streamlined so much, and from a quality point of view it also took out a huge number of errors and potential errors."

Redesigning EHRs

Skeptics continue to point fingers at inadequate EHR software. One notable physician practice startup recently decided to write its own EHR instead.

"With our system, I can drill down and see if one doc is doing better than the other," says Rushika Fernandopulle, MD, CEO of Iora Health, a Cambridge, Mass.–based primary care practice that manages 7,000 patients in six locations around the United States. The same is true for patients, he says. For instance, Iora's EHR software can also quickly identify patients whose hypertension is under control. Clinical averages are around the 50% mark in the United States, while good practices achieve 60% or 70%, he says. At Iora Health, 90% control is achievable.

The net result at Iora Health's clinics, within a one-year time frame: "We've seen [a] 50% drop in emergency room visits, 40% drop in hospitalization, 25% drop in specialty visits," and a 10% total drop in healthcare spending, Fernandopulle says.

"The big caveat is very few if any of the EHRs out there in the market do any of what I just talked about because they're so focused on this coding and billing stuff," Fernandopulle says. "We've had to build our own, which does things from a completely different point of view."

Iora Health's experience highlights the fact that the definition of ROI is different for fee-for-service systems than for value-based healthcare. In fee-for-service, an EHR that bills most effectively is the one with the greatest ROI, while just the opposite can be true in accountable care.

In Iora Health's business model, insurers such as the Culinary Health Fund of Las Vegas pay a flat per-member, per-month fee to Iora Health. "We don't have to ever submit a claim or a bill," Fernandopulle says. The payer power of the Health Fund is such that Las Vegas hospitals are compelled to report hospitalizations of Iora Health patients to the clinic, thus helping engage primary care physicians in controlling hospitalization costs and reducing readmissions.

Fernandopulle's belief that the better ROI lies in discarding existing EHR software entirely is definitely a minority opinion. The vast majority of healthcare leaders continue to emphasize the greater importance of process redesign.

"We often talk about how important it is to redesign workflows and improve your processes before you automate them," says Gary S. Kaplan, MD, chairman and CEO of Virginia Mason Medical Center, a Seattle-based integrated health system with 2012 net patient service revenue of $875 million, a 336-bed acute care hospital, a multispecialty group practice of more than 450 employed physicians, and a network of regional clinics. "In fact, we jokingly say that if you don't do that, you're really just moving garbage at the speed of light, and you're magnifying inefficiency and accelerating it in some respects."

At Virginia Mason, though, a thorough evaluation of the ROI of EHR technology is simply not a priority, Kaplan says. "We view our investment in the technology as really a cost of doing business," he says.

"In general, we don't use ROI for a lot of our capital expenditures," adds Suzanne Anderson, executive vice president, chief information officer, and chief financial officer at Virginia Mason.

"We have a much larger capital appetite than we can afford, as most healthcare providers do, and so being good stewards of our resources, we need to really evaluate what it's going to do for our business in general, and not just what the economic return is," Anderson says.

And yet Kaplan regards the Virginia Mason Production System, in conjunction with Cerner EHR software, as contributing to higher-quality care that is both safer and more efficient, such as providing timely diagnosis of a curable colon cancer. A key enabler is Cerner's health maintenance module, the EHR screen that greets primary care physicians and call center operators.

"They can actually use that technology to help ensure that our patients have appropriate previsit testing, are up to date, and compliant with all of the recommended screening guidelines," Kaplan says. "We think of this entire wired system as a way of facilitating integrative, coordinated, seamless, appropriate care."

"One of the things that we say here is that our IT projects are really only 20% technical, and the other 80% is adaptive change and integrating the operational systems with the technology," Anderson says. "We spend much more time on that operational technology integration, so that when we go live, it's just a seamless part of our system."

An example occurred when Virginia Mason implemented bar code medication administration. Up to 18 months of preparation ensured that nearly all medications could be scanned by bar code readers that compared the codes to previously entered physician orders stored in the EHR, she says.

Because the bar code process worked the first time, "the nurses were not going to do some of the common workarounds that might otherwise lead to unsafe care," she adds.

Although Virginia Mason chooses not to focus on ROI, Kaplan does share that because of the combination of Virginia Mason Production System's redesigned Toyota-inspired workflows, underpinned by EHR technology, "at least double-digits [of] millions of dollars have been saved."

"We've been on this EHR journey for more than a decade," Anderson says. "EHRs are not in their infancy but are probably in their teen years, and are not fully developed the way that they will need to be in the future to manage new ways of looking at value and population health. We've made a conscious decision not to be a developer, but to work with vendors, so that we don't get out of sequence with the rest of the country. We continue to work closely with Cerner to move to more mobile applications with them, and we will continue to push them in that direction."

Enabling improvement and value

Take the EHR's ability to manage a population and add its capacity to stop duplication of radiology, lab, and genetic tests; to suggest less expensive medications to a prescribing physician; and to track readmissions at scale, and the savings add up, says Marlon Priest, MD, executive vice president and chief medical officer at Bon Secours Health System, a Marriottsville, Md.–based not-for-profit Catholic health system that reported 2012 total net revenue of $3.4 billion and owns, manages, or has joint-venture oversight of 19 acute care hospitals across six states, along with long-term care, assisted living, and retirement communities.

"If you use the EHR right, the record will remind you that Mrs. Jones has a Foley catheter on day one, she had a procedure yesterday, and by protocol it should come out on day two," Priest says, "but it often gets left in because nobody looks under the sheets."

Reduce the amount of time those catheters are left in, and it is possible to realize cost savings from exchanging a hip replacement with no infection for the same procedure with an infection, Priest says.

"Did the electronic health record do it all? It didn't do all of it, but we never argued that," he says. "We argued that the record makes your ability to recognize and transform care better. But we started our transformation journey just before we started our electronic record journey."

In Bon Secours' case, that journey began in 2006 with the assembly of order sets and care plans, "things we know we need to do, that we've never really been able to do in paper well," Priest says. "Let the EHR be an enabler."

As for hard numbers, Priest says, "We've spent to date probably $350 million total, and if I do the math, we've got our money out of it."

In 2012 alone, Priest's team was responsible for generating $50 million in savings through ConnectCare, Bon Secours' EHR. He estimates the savings in 2013 could be $60 million.

A small example illustrates why. Nurses starting blood transfusions, IV fluids, or chemotherapy are supposed to record both the start time and stop time of treatment. A DRG fee gets paid regardless of the stop time, but that only covers the drug itself, not its administration.

Now suppose that the procedure takes place during a shift change. In a paper-based record system, the new nurse is reluctant to record a new stop time for fear that the first nurse already recorded a stop time and that the entry of a second stop time could have the appearance of double-billing for the administration. In that paper system, if no stop time is ever recorded, the administration billing cannot occur, and the hospital leaves money on the table. If, instead, all the recordkeeping is captured electronically, the transfusion/administration stop time always gets captured, which allows for an accurate billing of the administration, Priest says. "We built a system so no matter how many times people record the stop time, it took the first one," he says.

Net revenue gained from this change in just one of Bon Secours' markets was $1 million per year, Priest says. Overall revenue increase in 2013 due to better coding and documentation compliance will top $25 million, he adds.

Coordination around the C-suite has been essential to transforming Bon Secours through technology, Priest says.

At the start, Bon Secours President and CEO Richard Statuto and Priest realized the system had the cash to go any number of ways. They realized they wouldn't be gaining much market share just by adding hospital capacity.

Instead, Bon Secours drove clinical and financial innovation by assembling multigenerational teams and tackled critical issues. In the clinical setting, sepsis was the first target and the EHR was key to defining, measuring, and communicating the problem throughout the care team, Priest says.

The first goal, in 2007, was modest—to save $4 million annually by reducing the costs caused by avoidable charges due to sepsis. "As a $3 billion company, it's pocket change," Priest says, but nonetheless, "everybody thought it was a great idea." And the efforts continue.

"Six months into the year, nobody is on target. Finally the CFO and I wrote a communication to the unit CEOs: 'Guys, this is 10% of our bonus this year. This is an expectation the board has, and you know they're going to ask us to talk about people who died, so why don't we get on with it.' They hit the number."

Still, at many providers, the balancing act is to provide these kinds of EHR-driven savings without cannibalizing systems that still rely on fee-for-service for the bulk of their revenues, until they reach the tipping point and switch entirely to value-based care.

"You've got to have a way to give docs a chance to make the same amount of money that they were making last year, but doing something better for their patients," Priest says.

Part of the answer is surrounding those doctors with a knowledgeable healthcare informatics team. "We just hired a young man who was a surgeon and gave up his career entirely to get a master's in health information, because he was convinced this was the next wave," Priest says. "He's doing a lot of our decision-support work now. He's beginning to put together technical people, particularly pharmacists and nurses, to think about care."

When EHRs collide

Mergers and acquisitions of hospitals and healthcare systems—a trend that shows no signs of slowing down—can confound the gains an EHR can achieve. Incompatible EHRs in a merged organization must either be ripped out and replaced by the acquiring system's EHR, or integrated through health information exchange technology or enterprise data warehouse technology. Neither course is easy, as many technology leaders can attest.

Virginia Mason's stand-alone growth enables it to have a clearer picture of the value of its IT enhancements. "Most of our growth has been organic growth," Kaplan says. "Where people may not have had the same sort of systems, we just integrated them into our Cerner platform."

One place where multiple EHRs exist today, and will continue to do so for some time, is North Shore-LIJ Health System, which reported 2012 total revenue of $6.6 billion and includes 16 hospitals and nearly 400 ambulatory physician practices.

"Most of our hospitals are either deploying or have deployed Allscripts' Sunrise suite," says Michael Oppenheim, MD, vice president and chief medical information officer of North Shore-LIJ. "Most of our medical group is deploying Allscripts Enterprise." Two hospitals that joined the system were already using the McKesson platform when acquired by North Shore-LIJ.

"Because of the scramble to get all of our hospitals that are still on paper onto the core platform, at this point those two hospitals are continuing to move forward with their McKesson deployments," Oppenheim says.

To leverage interoperability between the different vendors' EHRs, North Shore-LIJ is also deploying InterSystems health information exchange software.

This HIE software's notification capabilities can alert primary care physicians when one of their patients has arrived at an emergency department or a specialist somewhere in the system, Oppenheim says.

"It's been very effective at getting the practice to communicate with the hospital," he says. "It's also raised a lot of interesting questions that we're struggling with around the volume of patients who are using the emergency departments during business hours and could potentially be handled in an office."

In the final analysis, whether ROI is measurable or not, whether it is sufficiently fast or not, many healthcare leaders agree that the EHR remains the essential platform that will drive the Institute for Healthcare Improvement's Triple Aim: improving patient experience of care, improving population health, and reducing the per capita cost of healthcare.

"We know the consequences if we don't make the investment," says Lloyd Dean, president and CEO of Dignity Health, a 21-state network with FY 2013 net patient revenue of $10.5 billion, 39 acute care hospitals with 8,400 beds, 9,000 active physicians, and 55,000 employees.

"We know that in the current systems, the fact is that system A doesn't talk to B, patients have to input data multiple times, physicians have to wait to get critical data to help make decisions in terms of the patient care; we know also that we get into rebilling and we get inefficient information if we don't change the status quo," Dean says.

"If you look at moving data to a place where either the patient, the physician, the clinical team, [or] the provider teams can make decisions, then it's a good investment. I think we absolutely have to continue."

Like any system implementation, some providers have had immediate successes while others have not, Dean says.

"Some people have stumbled because the planning, the system, or the culture environment wasn't right, or the software just did not do what it had hoped to do," Dean says. "But is that sufficient enough to throw the whole system out, or to condemn the investments that we're making in EHRs? I think not. There are enough pilot indications that say it is scalable."

Reprint HLR0214-2

This article appears in the January/February 2014 issue of HealthLeaders magazine.