CDC on EHR errors: Enough's enough

Problems go way beyond outdated software
By Evan Schuman
06:30 AM
Share
The CDC has found that EHR hiccups have their roots in poor design.

The Centers for Disease Control and Prevention does not routinely get involved in telling hospitals how to run operations, but with increasing reports of EHR deployment problems, the Atlanta-based operation now sees the need to act.

"Some hospital laboratories have legacy information systems that do not even have the ability to use current coding," said Megan Sawchuk, the lead health scientist for CDC's office of public health scientific services, which is in the division of laboratory programs, standards and services.

The problems, though, go way beyond outdated software. There are also issues involving staff time and expertise. One key example is the huge number of codes and the maddening fact that different medical facilities use different codes for the same tests. This is particularly problematic for physicians who work in multiple hospitals and practices. "We have to develop a simpler coding system that balances the clinician’s need to consistently order the right test with the laboratory’s need to show unique aspects of testing when necessary," Sawchuk said. "There are pros and cons on each side, but ultimately we want to make it easy for clinicians to order and interpret the right test for the patient."

[See also: Object of beauty, or ungainly nuisance?]

Fixing this problem is critical, but the answer isn't obvious. Creating consistent codes will require staff to make a lot of changes. "Staff need to have the time to learn and do the coding. In small settings -- especially rural hospitals—they are very unlikely to have that extra personnel," she said.

The CDC, which documented some of its concerns in a report issued in May, is finding that many of the EHR hiccups are not only impacting patients, but have their roots in system design. Specifically, the lack of IT participation in design details prior to rollout.

"End users are not yet sufficiently involved in the early stages of EHR design, including innovation in display design and workflow analysis, and that creates problems with their ability to effectively understand and use the information," Sawchuk said. She cited an example that tied a display preference with a patient not being treated for a life-threatening disease.

Laboratory data is typically displayed in reverse chronological order, with the newest results on top. One EHR had the default setting as chronological.  This caused a patient to have a critical delay in her care when her most recent and abnormal pap smear was shown at the bottom of the list. The physician reviewing the historical results only looked at the top of the list and assumed those were the most recent details. "Had a clinician provided input during development of the system, or even during implementation, this patient safety event could have been avoided," Sawchuk said.

Today there is no single place for heath IT issues such as this to be reported, including “near misses,” Sawchuk argues, pointing to the best option being the Health IT Safety Center proposed through FDASIA.

Sawchuk argues that such data centralization is needed because it's the only way to get the data that is needed to fix the system. "At this point, many people have shared anecdotes about safety concerns with their EHRs, but it is difficult to decipher the real from perceived concerns, or even large-scale, systematic concerns from random, singular concerns," she said. "Until aggregated, credible data is available, it will be a challenge for all stakeholders to prioritize their resources to address the most important concerns."

[See also: 5 health IT insiders offer their takes on EHR usability.]

The CDC's position makes sense. The problem looks deceptively easy to resolve given widespread agreement that naming and preference standardization is needed. The answer lies is getting hospitals to agree on the consistent format and that funding everyone using anything else to make a huge number of changes. The fact that this should have to be done only once is of little consolation to short-staffed hospital administrators who have no resources to make such changes.

In short, agreeing that a change is needed is not the same as having the wherewithal—meaning money—to do the work to change it. Unfortunately, making this optional—or on a "when you have the time" schedule—is no longer viable.