Improved EHR usability requires a new look at how these systems work


The usability of electronic health records has become a major concern, partly because of safety problems that may, in some cases, be attributed to poor EHR design. In addition, observers have pointed out that physicians are more likely to use EHRs that are well designed and easy to use.

These are among the reasons why the National Institute of Standards and Technology (NIST) devised a technical guidance document to help vendors and other parties evaluate the usability of EHR systems. NIST offers some valuable ideas on a technical level, but it does not address the main reasons why many physicians still find EHRs unusable or less usable than they should be.

One major issue is documentation--in other words, data entry by clinicians. Many physicians, for example, find it difficult to input data using point-and-click templates; most users find it challenging to get data into the system when an EHR is new and few patient visits have been documented on it; and some EHRs require physicians to switch to different portions of the record when they have to document their treatment of multiple problems.

This list only hints at the deeper issues: First, most general-purpose EHRs have good primary care templates, but they're not really strong across all of the major specialties. For example, KLAS recently surveyed physicians about their experiences with 18 ambulatory-care products. On a scale of one to 10, EHRs for internists and family physicians scored 7.6 and 7.5, respectively. By comparison, EHRs for urologists scored 6.4, and those for nephrologists, 6.2. Oncology and ophthalmology EHRs were both rated at 5.8. Of the leading EHRs examined, only Epic--designed for large multispecialty groups--scored well across a wide range of specialties.

Then there's the problem of templates and data views that don't reflect the way doctors think or the typical routine they normally follow during patient visits. Some EHRs are much better than others in this respect, according to physicians I've interviewed. But a recent KevinMD discussion shows that inept documentation methods still irritate many physicians. One doctor commented, "One often has to enter the same info 2-3 times. It doesn't cross-populate. And if you don't enter the info the way the computer wants you to enter it then it just isn't happening. No subtleties of diagnosis or physical exam are allowed."

Some consultants from Booz Allen Hamilton recently pointed out that EHR design has not changed much since 1982. In an analysis they performed for the California Healthcare Foundation, they found that

  • EHRs were originally designed to maximize billing, so their developers paid little attention to maximizing quality and efficiency
  • It's still difficult to design programs that capture the complexity of healthcare while being user friendly
  • EHRs were not designed for information exchange, because most of them were--and still are--used to capture patient data within individual offices and enterprises, not across business boundaries.

The EHR approach to data entry has also not changed fundamentally in decades. But there are signs that change is coming, partly because of rapid improvements in voice recognition software and natural language processing. In addition, EHR vendors are developing mobile-native applications that will make it easier for physicians and nurses to enter and view data on their iPads.

All of this should be combined with the sort of usability evaluations that NIST is proposing. But at the same time, clinicians and vendors ought to join forces to rethink how EHRs work. There may be no single solution that will meet everyone's needs. Instead, the solution may be to provide physicians with greater flexibility in how they can use EHRs. - Ken