Sharing Electronic Medical Records Still Too Hard

From article posted on InformationWeek June 17, 2013

The average patient can’t fathom why the sharing of electronic medical records is so hard. But those inside healthcare aren’t thrilled either with the state of electronic record interoperability, as several smart discussions at this week’s Digital Healthcare Conference in Madison, Wis., showed.

“If we’re this far into this implementation across the country, and we still have this level of discordance, shame on us,” said Dr. Frank Byrne, president of Wisconsin’s St. Mary’s Hospital. “How did we get here and how do we get out? Because we’ve created barriers.

Epic CEO and founder Judy Faulkner highlighted some of the obstacles to data sharing, from patients wanting to control such sharing, to difficulty training clinicians, to the many technical challenges. While data-definition standards in theory should make sharing easier, “the standards are only describing a very, very small subset of the data that’s really there,” Faulkner said.

Here are a few of the broad questions about interoperability discussed at the DHC 2013 event, with input drawn from several speakers and sessions. What’s missing? Share your thoughts in the comments below.

Q: Shouldn’t data standards allow easier sharing across vendor systems?

Faulkner sounded a skeptical chord about today’s standards solving interoperability. There are about 110,00 data elements in the electronic health record, she said, and existing data-definition standards don’t come close to capturing the full scope of what’s in electronic records. What’s more, Faulkner said she is a bit worried about standards as a double-edged sword — standards might improve communications, but limit innovation and new ideas. “That has to be always balanced as well,” she said.

A more hopeful view of standards came from Jamie Ferguson, VP of health IT strategy and policy for Kaiser Permanente, one of the nation’s largest managed healthcare groups. Ferguson said standards are “perfectly good” for standardizing close to two thirds of the needed records. But he said EHRs generally aren’t implemented well based on the standards.

Byrne said that as a practical matter, interoperability is working locally for St. Mary’s and neighboring health system around Madison because many use Epic EHRs, making data compatibility easier. Direct Epic-to-Epic data exchanges will be most common, he said, but it’s also exploring and supporting other exchange-based options. Byrne, Faulkner and Ferguson shared a panel discussing interoperability.

Q: Why isn’t there an API culture in healthcare?

Big software platforms in other industries use application programming interfaces (APIs) to allow integration and development of add-on applications by third-party software makers. APIs have fueled the boom in mobile app development.

Although mobile app development has picked up in healthcare, it would benefit from more extensive APIs for medical record data, said Judy Murphy, deputy coordinator of programs and policy for the Office of the National Coordinator (ONC) for Health IT, in a separate presentation at DHC 2013. “Many, many of the electronic health records are still proprietary and closed, they don’t publish APIs, they don’t allow app developers to access their information,” Murphy said. “And that’s part of what we’re trying to change.” She cited open government data efforts such as Medicare’s and Medicaid’s that have led to new mobile apps, such as iBlueButton.

Faulkner emphasized the openness Epic does allow: Epic releases its source code to customers, and will train providers’ developers on the system and all the ways they can pull out data for their use. “What we don’t do is release that to other vendors,” she said.

Q: Why aren’t there better ties to specialist systems, from optometry to cardiology, to easily update the EHR?

The question came from Scott Jens, founder of RevolutionEHR for optometrists. Faulkner asked how many optometry records vendors there are, and Jens said about 30. That kind of software sprawl exists in every specialty practice area. If Epic worked on interfaces for all the vendors that wanted to integrate, “we would do no more development on our software,” Faulkner said. “All we would do would be interfacing to the other vendors. … We would need thousands of programmers just to be on top of that.”

Q: Is some kind of central repository for some core health data the answer?

Such a repository would face the same challenge Epic does in contemplating integration, said Faulkner: Does that repository have an army of developers to write and maintain all the necessary interfaces to the systems that will contribute data?

Ferguson at Kaiser Permanente was unequivocal: “Bad idea.” One, he said, a central repository is a huge breach target. Two, the opportunity for conflicts of interest are insurmountable. Three, the expense is unsustainable to maintain a big central repository and normalize all that data. Instead, the better option is standards-based exchange efforts, such as the national e-health exchange started by ONC.

Q: Is the answer a simpler download to a personal record, so people do their own aggregating of health data sources?

Kaiser Permanente has let people download their own record for years, but there isn’t an easy way to transfer that mass of data into a third-party, consumer record. “The technical barriers are way too high today,” Ferguson said.

The list of independent, consumer-focused personal health records that have flopped is long and distinguished, including both Microsoft and Google. Byrne is betting on EHR vendors providing access via PCs or smartphones, such as Epic’s MyChart service. Given the parade of failed personal health record startups, “I like my chances better picking up my iPhone and pulling up MyChart, and I can do that today,” he said.

Q: What other barriers are there to interoperability?

Financial incentives are one barrier, Ferguson said, because fee-for-service medicine doesn’t provide the incentive for sharing information. Accountable care and integrated care models have “native incentives for having complete information and sharing,” he said.

Faulkner listed several barriers. One is patient control, which Epic hit with its earliest efforts to allow data transfers. “What we found right away is people wanted to share with people they felt comfortable sharing with and not with others,” she said.

Another is lack of training, she said. Emergency rooms generally are steeped in how to exchange data and gather what they need, but there are many other areas that could use Epic’s interoperability platform where they aren’t trained in data exchange. Ferguson seconded the training obstacle, saying it’s particularly tough in settings where clinicians only rarely exchange data and thus struggle to remember how to do it. “That’s one of the reasons we have such low exchange rates even where the technical capability exists,” he said.

This isn’t a comprehensive list of barriers to interoperability, only some of the highlights from a good discussion. Are there others you would add? Please share them in the comments below.