David Whiles, CIO of 320-bed Midland Memorial Hospital in Texas, is not easily deterred. That's why, when faced with the sunsetting of his hospital's information systems (none of them clinical), and desiring to move towards an EMR, he was not defeated by learning that a standard commercial product would cost his county hospital $18 to $20 million.

“At first,” he says, “we were hit with the sticker shock. That was simply impossible; we had no finances to pay for that kind of project at that time,” he recalls. In other words, money - or the lack of it - was a fundamental barrier to initiating an EMR implementation at Midland Memorial.

Fortunately, Whiles was able to learn about the open-source VistA EMR through Internet searches. The solution, which emerged out of a complex series of developments, is a commercialized form of the EMR operating in all Veterans Administration (VA) hospitals across the country. (The VistA acronym stands for Veterans Health Information System and Technology Architecture. The final “A” is capitalized to distinguish it from Microsoft Corporation's Vista platform.)

Several years ago, the source code for the VA's system was made publicly available online. Since then, a handful of companies have commercialized it in some form, though only two have stuck with the concept -Carlsbad, Calif.-based Medsphere Systems Corporation, and Juno Beach, Fla.-based DSS Inc.; and only Medsphere has installed its version in full-service allopathic hospitals (DSS has been focusing primarily on behavioral and other types of non-med/surg hospitals, and on clinics).

For some CIOs, though, customer base is not a preeminent consideration. Whiles and his colleagues forged ahead even though no hospital had yet gone live with open-source VistA, as the commercialized version of the VA's VistA is known. So Whiles hired a consulting firm, which validated the feasibility of self-development with VistA, and signed a contract with Medsphere in December 2004, beginning the process of implementation in January 2005. The organization gradually rolled out the system across various departments and clinical areas until a full go-live in June 2006, eliminating all paper charts by February 2007. The cost? About $6.4 million overall; encompassing price-guaranteed consulting from Medsphere, including ongoing consulting help with upgrades and improvements in the system.

In particular, Whiles says, “We don't pay for add-ons. With any other vendor - when asking for a little report to be run or a small change to be made - you'd have to pay additional fees; whereas everything they've done for us has been taken care of within the context of prearranged contract terms.”

Of course, there is a downside - implementing an open-source EMR can be more like self-development than implementing a commercial EMR product. Whiles and his colleagues knew that going in, and took on the challenge of being the first inpatient hospital to fully implement open-source, working more or less as a beta site with Medsphere.

“We've certainly contributed to Medsphere's success in partnering with them,” Whiles reflects. “But we recognized going in that we were going to be doing that (development) work, and that was just part of the deal.” Still, he says, “EMR implementation is never easy, though there are probably some products that are more difficult to implement than others.”

Meanwhile, Whiles is highly satisfied with the Medsphere open-source VistA system's functionality. “I would put open-source VistA up against Epic or any product any day,” he opines. “No system is a panacea, and neither is open VistA. There are certain functional areas that are not well-developed, simply because of the background, such as women's and children's services, OB/GYN, and pediatrics, labor and delivery; but that's simply because the VA doesn't provide those services.” But as other hospitals begin to implement Medsphere open-source VistA, he says, those clinical areas will add functionality over time.

Who will pick up the thread?

Nationwide, only a small number of hospitals, including several inpatient hospitals (the rest being rural critical-access hospitals, Indian Health Service hospitals, and behavioral hospitals, as well as a small number of nursing homes) have adopted some version of open-source VistA, either from Medsphere or from DSS Inc., despite a strong dose of attention that the open-source EMR concept received several years ago when the VA publicly released the source code. But if the open-source concept received an initial burst of attention, why has it not seen wider adoption?

On the one hand, IT leaders at the handful of inpatient hospitals that have implemented open-source say they are very happy with their choice. Among these are the 99-bed Memorial Hospital of Sweetwater County in Rock Springs, Wyo.; the 49-bed Welch Community Hospital in Welch, W.Va., part of a seven-hospital public system that has created a unified EMR; and the 479-bed Lutheran Medical Center in Brooklyn, N.Y., which was making final preparations for an organization-wide go-live at press time (the full, final go-live was scheduled for the end of this year).

In addition, several hospitals have signed contracts with Medsphere in the past few months, confirms Mike Doyle, the company's president and CEO. And DSS Inc., Medsphere's lone active competitor in the open-source VistA space, has five behavioral hospitals running that company's version of open-source VistA, confirms Mark Byers, president of the firm.

Implementers appear quite satisfied with the process, and the result. “The implementation was definitely not problematic,” says Hartsel Bryant, Welch's clinical applications specialist and the day-to-day manager for the system. “This is a highly customizable piece of software,” Bryant says of the Medsphere open-source product. “It's very user-friendly, and we're definitely getting our money's worth.” Also, open-sourcers point out, they have access to some of the knowledge that the VA's health system acquired in developing its original VistA solution, simply by virtue of the functionalities already contained within the VA's software itself.

What's more, Bryant's hospital is part of a seven-hospital system run by West Virginia's Department of Health and Human Resources; and, with Medsphere's help, that system has built a unified EMR across the seven facilities, of which two are behavioral health facilities, four are long-term care facilities, and the seventh of which, Welch, encompasses both 59 acute-care beds and 59 long-term care beds.

But if open-source VistA is a strong choice for some hospitals, why haven't more signed up? The reasons, industry observers say, are complex.

“I think there are three issues,” says Erica Drazen, Sc.D., managing partner in the Waltham, Mass.-based Emerging Practices Healthcare Group of the Falls Church, Va.-based CSC. “One is a perception issue; most people don't think their hospital is anything like a VA hospital. And the second issue is the reality that VA hospitals don't have the same reimbursement issues as other hospitals. And as we move further into pay-for-performance schemas, you have to integrate billing and clinical functions more and more closely, or you get a lot of waste. The third issue is that, with open-source, you are kind of on your own. You get the software, and you don't have a very big community of users. So it takes a special kind of organization to implement open-source.”

In addition, says John Halamka, M.D., CIO of Beth Israel Deaconess Medical Center in Boston, “Software licensing fees are a small part of the entire cost of software implementation.” And because so much work falls onto the purchaser of any open-source system, Halamka urges CIOs considering the open-source path to, “Make sure you have a sound implementation strategy, with project management, workflow redesign, process improvement, business analysis, and training resources well-specified before any purchase.”

Certainly, those CIOs who have implemented open-source VistA, or are in the process of doing so, seem to be a hardy and intrepid bunch. Take for example Steve Art, the senior vice president and CIO at Lutheran Medical Center in Brooklyn, N.Y. Art's whole attitude is one of can-do and self-reliance when it comes to software.

“One of the great things about open-source,” he says, “is that I no longer have to rely on my vendor to make changes to my software. And I am indeed making changes to it right now,” in the final ramp-up to go-live. “Also, there's a community around open-source. We have an agreement that we will make changes as we'd like to the product, and submit those to Medsphere, and if others like it, Medsphere will package those changes and make them available to the rest of the community. In the traditional vendor environment, if I make changes, the vendor charges me for those changes, and then charges its other customers for it. In this case, the better off I am, the better off the community is.”

Indeed, Art says, he relishes the open-source environment, with its do-as-you-like atmosphere. “Here, the people from the vendor company know they don't have control. It's up to me and other clients to make the product better, and they're relying on all of us to do that. In a traditional environment, I don't know the other customers of a particular vendor, nor do they know me; nor does the vendor encourage such communications or familiarity.”

Of course, the open-source environment requires the availability of individuals who have strong technical ability, as well as pluck. That's how Memorial Hospital of Sweetwater County got its open-source VistA implementation done, confirms Linda Simmons, R.N., M.S.N., the hospital's vice president of operations and its de facto CIO. In the case of Sweetwater County, a former programmer from the VA who had helped design aspects of Vista joined the Sweetwater staff to help implement open-source VistA, before going back into programming. Without that programmer's involvement, Simmons says, her small hospital, with only seven IT professionals, could not have been successful.

Nonetheless, Simmons, Bryant, and Whiles all say they're very pleased with VistA's overall functionality, while they work to build out the solution in areas such as obstetrics and pediatrics.

Tipping point?

While Medsphere and DSS Inc. executives say they are optimistic that open-source adoption will accelerate, and while the IT leaders of those hospitals that have implemented open-source praise its capabilities, some say a cataclysmic market change in not in the offing.

“The point is, innovation in the marketplace is hard,” says Joe Bormel, M.D., M.P.H., vice president and CMO at the Reston, Va.-based QuadraMed Corporation and an HCI Blogger. “The availability of an open-source platform, and the competitive pressure from that, is unlikely to make either the incumbent vendors or the innovative ones perform better in the marketplace. I think competition and having a variety of offerings out there is fundamentally good. But in terms of having a bright, exciting future with substantially more valuable, meaning cheaper, products; open-source doesn't seem to be the panacea it might be.”

Nonetheless, for organizations that face financial barriers to commercial EMR adoption, open-source remains a viable alternative. In fact, Midland Memorial's Whiles says, “Any hospital should consider this possibility. This really is a system that could work for any size hospital. And when you look at the open-source model, in terms of the contributions from a wide variety of sources that are involved, I think you get a better and ultimately more diverse product in the end.”