In 2014 Medscape conducted a survey of more than 18,000 physicians and posed a deceptively simple question – “Are you using an EHR?” Eighty three percent of respondents answered with an outright “yes,” four percent were in the process of implementing an EHR, six percent planned to buy or start using one within the next 1-2 years, and seven percent had no plans to make use of any form of EHR.
On the surface, an aggregate of over ninety percent of physicians either using or working towards using an EMR would seem to be a major milestone in modern medicine. However, when we start to drill deeper into the data some disquieting themes emerge. A staggering 70% of respondents felt that EMR’s decreased their face-to-face time with patients, 57% felt their ability to see more patients was effected, 27% felt their ability to respond to patient issues was compromised, and 26% felt their ability to effectively manage treatment plans was effected.
So why are physicians unhappy with their EHRs?
One often posited answer is that we are at an early stage of technological development for electronic medical records. Although the first documented medical record, an Egyptian surgery record on papyrus, can be traced to 1600 BC, the concept of electronic medical records is barely fifty years old. In modern terms the concept of the Problem Oriented Medical Record, as introduced by Larry Weed, is barely forty-five years old.
An often used comparison point in terms of technological age is the cell phone. Invented in 1973 by Martin Cooper of Motorola, the DynaTAC 8000x weighed 2.5 pounds and sported a single-line, text only LED screen and a price tag of $3,995. Forty years later we have the iPhone 5s, a cell phone with a finger print sensor, a 64-bit chip, a video capable camera that produces extraordinary shots, a beautiful high resolution screen, ultra-fast LTE wireless, every app imaginable, and a price tag of $199 under plan or $649 at cost. A cell phone so simple that my two-year-old daughter was found playing games on it one sunny afternoon.
At some level the cell phone analogy is a good one, but many would argue that the comparison is invalid because cell phones are not safety critical systems and are thus inherently less complex. However, the apple iPhone has a large code base that would rival any Epic or Cerner implementation and is most likely tested to about the same level. It is when we consider the hospital system as a whole, including connections from diagnostic devices and other information systems such that the complexity grows out of control.
From an end-user perspective, the complexity of the surrounding system is largely irrelevant. Their world is based on screens inside modules of the EHR that are far removed from the HL7 feeds and medical device integrations. A physician’s dissatisfaction with EMR systems is concerned with the fact that the system disrupts flow with the patient and “gets in the way.”
How then does the EMR disrupt work? Most often the disruption comes in the form of workflow or data entry issues. Here’s a common example: There is literally no information that cannot be stored in a traditional paper based patient folder. In fact, the author has seen folders containing photographs, micro-cassettes, carefully folded radiology films, and many other strange contrivances. The effort to move this data into the folder is minimal and the administrative controls are simple. Retrieval is even easier; the folder is placed in a cheap plastic carrier outside the exam room for each patient. Try doing that in an EMR!
Of course, this system of the patient folder evolved at a time when a patient was likely to see few specialists and was paired with a family Doctor who treated across generations. Fast-forward to today’s medical system where specialist treatment accounts for upwards of 50% of patient care. A chronically ill patient with diabetes mellitus can reasonably expect to see more than twenty providers in their lifetime, many of whom will want to see and edit the medical record. Clearly the chart of old is no longer a viable option.
For some percentage of physicians the chart of old remains the gold standard and EMRs are just harder to use. Certainly, the 7% of physicians who simply refuse to install EMR’s regardless of the financial benefits/risks of Meaningful Use are likely to correlate highly with this category. Even gold standard systems such as those at the VA have been reported as having “poor usability, time-consuming data entry, inability to exchange health information, and degradation of clinical documentation (Friedberg et al, RAND_RR439).”
In our view, usability is perhaps the most important barrier to acceptance of the EHR. Returning to our comparison with the cell phone, nobody would ever have called the DynaTAC 8000x or the next twenty years of its evolution “usable.” It is only in recent years that cell phones would be considered to have reached a usable form factor, and only then after many years of flops and miss-starts. The work of Ives and Jobs set a standard for usability based on years of product testing and an unrelenting quest for perfection at a point when the market had largely stabilized on the flip phone.
In 2013, a KLAS report on Acute Care usability found that none of the six EHRs the report examined scored above a 4. Epic took that high score, followed by Cerner (3.7), Siemens (3.7), AllScripts (3.5), McKesson Paragon (3.4) and Meditech v.6 (3.0). The report stated that Epic: “Wins over physicians during demos. A prescriptive approach to implementation ensures go-live success. Overall adoption of, and highest usability ratings for meaningful use functionality.”
Usability arises from careful work with the users of the system to understand, focus, and arrive at their true needs. This has not historically occurred within EMR design. In fact, a 2012 blog post on KevinMd described the approach to design as follows: “Rather than working with physicians to design the technologies and drive adoption, the experience (and almost universally the perception) is that the technology has been thrust upon physicians by administrators.”
So, where to from here?
It is clear that the current messy user experience in most EHR products is not a short-term problem. Vendors are so consumed with execution and delivery that they have no time, much less the inclination, to change their products to improve user experiences. Practices that use web-based solutions (not imposed on them by a health network) can switch providers and as such are best positioned for adopting new technology as it arises. Enterprises and health networks face large conversion costs and should expect the current problems with physician experience to continue for at least five years.
In summary, why is EHR so disliked by physicians? In part, because of poorly planned user interfaces not designed for physicians by physicians. Only when the focus returns to medicine and clinical workflow becomes a carefully designed, thoroughly “user experience tested” part of the commercial EHR products will we see this change for the better.