As an InformationWeek staffer recently assigned to this beat after only occasionally covering health IT in the past, I was surprised how unanimously and passionately dissatisfied most doctors are with the usability of this software, which they see as draining rather than enhancing their productivity. I'm sure there are exceptions where doctors are more enthusiastic about technology, the software they are using is higher quality, or a little of both. But if you open the door to a conversation about how horrible medical records software is, you'll get an earful.
Here's what I'm basing this on. Having spent the past few months writing about massive open online courses (MOOCs) for the education beat, I was happy to discover a Coursera course on Health Informatics in the Cloud starting at just about the time that I needed to come up to speed on my new beat. I'm happy to say the instructor, Georgia Tech's Mark L. Braunstein, MD, will be contributing to InformationWeek as a columnist, so watch for that. Braunstein has spent most of his career in healthcare IT, so I think it's fair to say he's a true believer in the potential and the necessity of digitizing medical information. However, when I turned to the course discussion forums I found a message thread titled "Health IT Doesn't Fix Problems -- Good Health IT Does."
[ Case in point: In One ER, EHR Takes 44% Of Doctors' Time. ]
One of my classmates was a pediatrician named Dave Denton, and the point of his discussion was that he hadn't seen a heck of a lot of good health IT.
"I use several EHRs in my clinic and hospital," wrote Denton, who practices in Portneuf, Idaho. "None of them allow transmission of data between systems. They all are encumbered by poor graphic user interfaces that make it hard to see patient data in a way that makes sense and helps patient care. It is actually much harder to take care of sick patients in the ICU with our new hospital system. They also tend to hide the pertinent by scattering it through the program and displaying all types of ancillary data, time stamps, and formatted notes that are inserted to ensure appropriate coding, but get in the way of seeing what is important. I have made several errors by failing to find the correct information in the chart because it was buried in the note. Finding what is important has become a treasure hunt."
In a follow-up post on dysfunctional workflow imposed by the software, he added, "I am not one of the physicians striving to resist change and implementation of EHR. There are so many valuable things this technology can bring to medicine. I am the chairman of our information systems committee in our hospital striving to make things work, but frustrated by the poor quality of products that have been forced into the marked by unrealistic timelines."
In his posts and in an interview where he elaborated on the same themes, he made it clear that he sees the potential for health IT even though he has so far been disappointed by the reality. At about the same time that I was tuning into Denton's rants in the classroom forum, a similar very active, overwhelmingly negative conversation popped up on the LinkedIn HIMSS group, under the title "Can we turn EHR dissatisfaction around?" (Hat tip to Jennifer Bresnick, an editor at Xtelligent Media, for getting the conversation started.)
The bottom-line answer from most (but not all) of the doctors and healthcare workers chiming in on that discussion was, essentially, no, not going to happen as long as there is such a mismatch between how the software in healthcare works and how the people in healthcare work. Clearly, the government agencies promoting the technology are in thrall to the software vendors, or why would they be pushing this so hard?
Typical lament: "How did IT get more powerful than the people who actually care for patients? The answer is that IT charges by the hour, while Docs can be made to do more work for less money, that is, add 2 hours work to every day without additional compensation. If we asked IT to do the hard work, we'd have to pay them." I don't want to quote by name without permission, but another commenter identified as a medical director for a healthcare group pointed out that any drug or medical device would have to be proven in FDA testing before being adopted into a hospital, while EHR software "impacts quality of care, and expecting to improve it 'on the job' causes delays in care, complications and death. Piece of advice, if anyone of you or your family has to go to a hospital make sure someone stays with the patient because the nurses will be on the computer."
Offline, I'd been hearing something similar from a friend in my Toastmasters group who is an emergency room physician. Practicing for a speech to a professional organization he is active in, he put almost as much emphasis on the productivity drain from electronic recordkeeping (and the coming horrors of ICD-10 expanding the number of codes to remember) as he did on the reimbursement squeeze aspects of Obamacare.
In both of the online conversations I referenced, there were technologists defending the good intentions of IT. They proposed solutions like better object-oriented software to separate generalized features from the requirements of a specific institution, or better use of either voice recognition or human-powered transcription to get the doctors away from doing so much data entry. But even those arguing that good EHR software is possible tended to acknowledge that a lot of EHR software is crummy. I hear the same thing from EHR vendors when they're talking about other people's products.
So far, the only doctors I've heard say nice things about their EHRs have been hand-picked references of the software vendors. (Again, I'm relatively new on the beat, so there's time for that to change.)
It would be easy to shrug off a lot of the complaints from doctors and nurses as the same old change management story, where people complain about whatever they are unfamiliar with. Healthcare providers aren't the first workforce to complain about the shift from paper to computers and being asked to change their work habits to match workflow dictated by software. The logistics workers who complained about the advent of supply-chain automation have either adapted or found another line of work by now, and, once the bugs were worked out, the efficiencies delivered by the automation became impossible to argue with. Eventually, EHR may deliver on all its promises of efficiency and patient safety. But has it been a mistake to rush it?
Obama partisans might also be tempted to see an anti-Obama agenda in this criticism, since this is another element of his healthcare reforms, along with the changes in health insurance regulation. But the sentiment is too unanimous to be that alone.
On the phone, Denton reiterated that he sees "a huge potential for benefit" from health IT. The problem: "Some of it was pushed before it was ready for primetime because of meaningful use."
That is, before the federal government introduced its system of immediate incentives for EHR implementation, combined with eventual penalties for lack of use, his hospital was beginning to introduce the technology, but slowly. "It was happening slowly because the products weren't very usable," he added. The HITECH Act establishing the meaningful use goals was signed into law in 2009 as part of economic stimulus measures, although it can be seen as a companion to the Affordable Care Act healthcare reforms that followed in 2010. Suddenly, EHR implementation became something that had to move forward at a steady clip, whether the software was usable or not, he said.
If not for government intervention, the technology would have been phased in more slowly and carefully, with time for course corrections along the way, he believes.
Meanwhile, meaningful use "hijacked some of the development" going into making EHR products better, Denton said. Or that's part of his theory of how things have gone so wrong. Instead of working on user interface improvements, vendors diverted effort into satisfying the checklist of government requirements to get their products certified as supporting meaningful use, he suspects.
One of the justifications for implementing health IT is that it should reduce medical errors caused by sloppy paperwork and unreadable prescription slips. Again, Denton sees "the potential for that to exist in a well-designed system," but generic software that tries to address every population and medical specialty can actually have the opposite effect, he said. User interface tricks meant to improve productivity can introduce errors. Auto-complete search, where a blank on a form will be filled in with a suggested match when a physician types in the first few letters of the name of a drug, can make it easy to pick the next drug down on the list, after the one intended. As a result, he has seen cardiac drugs prescribed where a painkiller was intended, for example. "So it introduces new potential errors." Even where the software is trying to head off potential errors, for example by detecting potential drug interactions, it is "really easy to get fatigued by all the popups" and stop paying attention.
His hospital uses a McKesson product, Paradigm, and has sunk enough money into it that the choice is unlikely to change. So as head of an IT oversight committee, Denton's role is to seek ways of improving that product's implementation. Yet even though the EHR and related systems at the hospital are all supposed to be HL-7 compliant, they don't share data, and the hospital is still struggling to be able to produce the "continuity of care" documents that healthcare providers are supposed to be able to exchange online in the new world of digital IT.
"Even at our own clinic, we've felt stuck," Denton added. His pediatric practice had to junk the first cloud-based EHR it tried as simply unusable, he said. Adding insult to injury, he wound up paying $10,000 to break the contract and get an export of his own data. Part of the deal was a "gag order," so he can't name the cloud software firm, he said, although "it's not one of the bigger ones." Since then, he has moved to OfficePracticum, but doesn't sound too enthused about that, either.
One reason for his disappointment is that as a medical student interning at Intermountain Healthcare in the 1990s, he worked with an early electronic medical records system custom built for its hospitals and was favorably impressed. While it may not have had every feature of today's EHRs, "it worked in a way that allowed you to really do your job better," he said. "But it's unique to them, and they spent years and years developing it."
In contrast, commercial EHR software has to be generic enough to work in many hospitals and all specialties. The result is a compromised design that doesn't serve anyone's needs really well, Denton said.
Thus, Denton came into the Coursera informatics course wanting to study up on how all this technology is supposed to work. I had to ask if he thought our professor, Dr. Braunstein, was describing a different world of health IT than the one he was living in.
"I think his world exists," Denton said. There seem to be select places in the U.S., probably academic medical hospitals with the resources to invest in perfecting their IT, where positive results from EHR and health information exchange between providers is a reality. But the effect of meaningful use has been to spread the technology across the country into communities that lack those resources. "It's probably a matter of trying to do too much at once," he said.