Electronic health records haven't exactly been touted as health care's Holy Grail. But there are high expectations for the technology's ability to fix some of health care's major comorbidities -- information unavailability, poor interprovider communication and inconsistent (or no) documentation -- and thereby improve patient safety across the board. Too high, some experts claim.
"If EHR systems didn't have value, no one would use them," Edward Fotsch -- CEO of PDR Network, a distributor of drug-labeling and safety systems -- said, adding, "There's no doubt that HIT -- and EHRs -- are keeping patients from being injured.
"But the expectations are way too high. We presume that because it's a computer, it's perfect, and because it's digital, it's safe. Of course it isn't," Fotsch said.
Those unreasonable expectations are one factor in some safety problems emerging in the EHR realm. Drug-dosing errors resulting from weight-conversion or -limit features, medication list discrepancies among systems, treatment delays resulting from faulty provider-computer interaction and important free-text comments that don't cross interfaces are among the hazards increasingly being reported.
Missed-diagnosis and wrong-patient errors still occur in the EHR era, research shows, and some systems make it very difficult for users to amend entries or simply delete incorrect information. To do so, users often have to resort to the equivalent of footnotes, which may be located far below the entry they seek to modify. And increasingly, the one-note-fits-most functionality many EHRs employ -- "click-tation," it's been called -- can produce notes so boring and boilerplate-sounding that recipients miss the important finding.
PDR Network's recently launched EHREvent reporting system is becoming more robust by the month, but its reports to date likely represent only the tip of the iceberg, Fotsch said.
Then there's the elephant in the room: EHR alert overrides. Research has shown that 60% to 95% of drug-drug interaction, drug-allergy and other alerts generated by clinical decision support systems are bypassed or ignored by annoyed users.
"Computers will do a better job at most things [than humans], but not at knowing which alerts to ignore," Fotsch said. "And all [drug-drug interactions] are not equal -- some are serious, and some are not."
Sociotechnical Issues Undermine Gains
Overzealous clinical decision support systems and alert-fatigued providers have become the poster children for EHR technology's safety downsides. They're also fertile fodder for opponents of the rapid-adoption push that's ensued since the government allocated $20 billion for health IT implementation. But EHR safety problems are far more complex and multifactorial than many reports indicate, according to Shobha Phansalkar, a senior medical informatician with Partners Healthcare in Boston who evaluates and develops decision-support systems.
"Many of the errors we see are not because of the EHR system malfunctioning but because people haven't spent the time to really think about how a provider interacts with that system," Phansalkar said, adding that a recent Institute of Medicine report "shows the unintended consequences of not taking workflow into account."
That November 2011 IOM report, titled "Health IT and Patient Safety -- Building Safer Systems for Better Care," effectively concluded that the jury is out and the literature is inconclusive on whether EHRs have actually conferred patient safety gains. But it clearly pointed to the sociotechnical factors -- including poor user-interface design, user-proficiency deficits, and roadblocks to interoperability and data exchange across EHR systems -- that may be compromising the safety gains the technology promises.
Those inherent problems will have to be addressed with the next iteration of products, Phansalkar maintained, but in the net, patients are safer with EHRs than without them.
Don't Dismiss the Progress
"Safety has improved. The primary gains are in having a medium of communication when many teams are involved in a patient's care," Phansalkar said, "and in having one place where all of the patient information resides. But sociotechnical factors are really the issue, and those are at the helm of the next [safety] gain with EHR adoption."
If those gains are to be realized, users must do more than just grouse, Mary Griskewicz -- senior director of health information systems for the Healthcare Information and Management Systems Society -- maintains. "It's up to the physicians and their organizations to work with the vendors and tell them which alerts are appropriate for their system, and what the workflow issues are," she said. "Vendors have user committees, and large organizations have CMIOs. People need to use them."
Dean Sittig -- a researcher at the University of Texas School of Biomedical Informatics in Houston -- concurs with Phansalkar that EHRs make the health care system safer. But he cautions that EHRs' safety improvement potential, for now, is overrated; systems are too nascent in development and too variable in quality to close the myriad safety gaps that persist.
"When I hear President Obama or government officials talk like the EHR is going to remove all errors, it's frustrating. The point is, we're not going to zero," said Sittig, who testified before the IOM panel that produced the recent report. "EHRs are like cars. They're inherently a good thing, but they need some seatbelts, wipers and headlights to make them safer. And we're not there yet," he said.
In addition, the "safety" of the users themselves is presenting new safety challenges. Training may be insufficient, or users may think that because they've worked on one system, they'll be able to transition to another one easily. "Previous EHR users have an advantage, but that doesn't mean they'll know how to enter complex orders on the next system," Sittig explained, "and we're seeing new kinds of errors because of this."
Increasingly, two big-picture issues -- the lack of EHR standards and fact that no EHR safety oversight body exists -- are conspiring to make the entire EHR arena less safe, some health IT experts contend.
In its report, IOM calls for establishing a federal entity to evaluate EHR product safety; track and report safety issues; and investigate IT-associated injuries, deaths or unsafe conditions.
Unfortunately, the train has already left the station, according to Ross Koppel, a health IT researcher and sociologist at the University of Pennsylvania who reviewed the IOM report. "The IOM report says that vendors' approach to making usable software has been way behind, and everybody agrees that usability remains horrific," Koppel said. "It should not take 17 clicks to find two pieces of information that should be contiguous."
Koppel noted that in some ways patients are safer because of EHRs -- they remove handwriting problems, oblige doctors to indicate the route of the drug and speed the order to the pharmacy. "But it's not a slam dunk, and we're just now starting to find out about and deal with the new safety problems," he said.
Vendors' hold-harmless and non-disclosure clauses -- which have deterred open reporting of EHR safety problems or sharing of information among users and vendors -- aren't helping, Koppel added.
While health IT leaders say the next step toward improved EHR safety is quantifying and qualifying errors and system problems, the current state is that we don't know what we don't know. But we're on the road to finding out, according to Hardeep Singh, chief of the Houston VA's health policy and quality program.
"We've had a big measurement problem in health care, and unfortunately or fortunately, the EHR makes it easier to measure. You start finding quality and safety problems you never found before because you didn't have the technological infrastructure to find them," Singh said. "If we want to get to the next level of quality and safety, we have to extract that EHR data and use it."