Techtalk: Why Today's EHRs Are Destined to Fail

I should start by admitting my bias; while I am a tree-hugging, Birkenstock-wearing practicing family physician, I am also the developer and owner of an EHR and am actually one of those trying...

I should start by admitting my bias; while I am a tree-hugging, Birkenstock-wearing practicing family physician, I am also the developer and owner of an EHR and am actually one of those trying to sell you on the idea of adopting an EHR. And while I truly believe that computerized records will be in each of our offices by the end of the decade, the vast majority of products available today are the Betamax of medical records: cool concept, neat to look at, yet destined to fail because their underlying design is flawed.

To understand the fundamental flaws in EHR software, we need to know the disparate groups who have come together to rescue us from our scribbled notes: the business executives (a.k.a. the suits) and the developers (a.k.a. the coders). Both groups have their strengths, but both are woefully unaware of what the doctor—patient relationship is all about. So, add to these a few doctor consultants, and you end up with today’s EHR vendor: a company and product that has all sorts of neat bells and whistles but has been built on a flawed assumption that suits and coders know best how we should be documenting our visits. There are three serious problems with today’s EHRs: their system architecture, their usability (or rather their inability to be easily used), and their prohibitive cost.

Software Architecture

The coders have designed the program to collect and store each grain of data as a separate item. This granularity of data is collected as a so-called controlled vocabulary, and each data point can be codified and identified. This concept is now being touted by many experts as the best thing since sliced bread. This controlled vocabulary—the technology of turning the doctor—patient encounter into an inventory of individual data points—is the best way to record a visit, believe these experts (experts, I should note, who don’t have to see and document 25 patients/day to pay their malpractice premiums).

The concept of a controlled vocabulary is an example of a good concept pushed well beyond its potential usefulness. Ideally, it makes sense to have certain data easily extractable and quantifiable. It is certainly reasonable, for example, that the data regarding a patient’s diagnosis and medications be distilled down to fundamental terms that can be easily extracted, shared, and researched. Likewise, many procedures can be easily coded in a manner that provides for analysis, which free text—the opposite of a controlled vocabulary—just doesn’t allow.

That said, have we not learned from our experiences with the controlled vocabulary of the ICD-9-CM that trying to describe even one critical piece of data in a codified way—the patient’s diagnosis—is problematic and mostly results in the physician just picking the closest code that will get the visit paid? I find it hard enough to find the correct four- or five-digit code for run-of-the-mill hypertension. Now, the suits and coders want me to distill every sentence of my patient’s history and physical into individual data points too?

Academicians argue that this granular data is necessary to allow research on all sorts of sign/symptom constellations and to protect the public health so we can avert communicable disease epidemics. Ironically, the EHR may well limit the coming plague, since most of the afflicted patients will still be in the exam room while the doctor is trying to figure out how to get their EHR to document the pustules’ duration, location, and modifying features.

To make matters worse, many of the controlled vocabulary dictionaries are proprietary and require a royalty to use (eg, the AMA charges a royalty for use of its controlled vocabulary, the CPT codes; hence one of the reasons you have to pay to buy coding books every year). It is truly amazing that not only are we being told we have to use a new vocabulary to document an encounter, but we’re being charged a fee to use the dictionary.


In stark contrast to technologies that have been shown to improve our efficiency, most EHRs today will dramatically slow down our ability to document a note. Don’t believe me? Just attend a vendor documentation challenge, and experience a level of frustration you likely haven’t felt since you last filled out a patient’s workman’s comp forms. At the recent TEPR documentation challenge, for example, 90% of the products demonstrated couldn’t even finish a progress note in seven minutes (the challenge was to document a simple follow-up visit on a patient with diabetes and hypertension). But thankfully, most of those that failed let you pull all your patients who have blue eyes.

In addition to the dramatic slow-down when using controlled vocabularies to document (eg, see how many clicks are needed just to document “dry cough x 3 days”), most of today’s high-tech solutions also require navigating window after window and screen after screen just to document a brief encounter. Why the suits and coders feel the chief complaint, history of present illness, and every other section of the note should be accessible only from separate windows is beyond me. Many of these vendors will shower you with white papers and pseudo-scientific research showing an increase in efficiency and miraculous returns on investments. Yet, if one listens closely to the statistics, it becomes clear that something isn’t quite right. It seems that the research shows that 33% of practices had an increase in efficiency, 33% of practices were slowed down, and the other practices reverted back to paper charts.


The most serious flaw of today’s EHR, and in my mind the most offensive, is the unconscionable price we are being asked to pay for these mediocre products. We literally have been forced to request bids and hire consultants to negotiate a contract with EHR vendors to buy something that, in most instances, costs more than a new car and will actually end up upsetting us more than if we had paid $100,000 for a Pinto.

There is no lemon law to protect the buyer of today’s EHRs. And as our medical establishment encourages, cajoles, manipulates, and in some cases, downright forces us to adopt these overpriced technologies, it is our practice and our patients who will suffer. We are told to spend time and money being trained by EHR vendors on the “right way” to document an encounter, which boils down to nothing more than a glorified, yet overly complicated, inventory system to catalogue patient signs and symptoms. Sadly, this unintuitive software itself becomes the main focus of the encounter, and our patients watch quietly as we interact with our computers.

Dr. Bertman is a Clinical Assistant Professor of Family Medicine at Brown University, and the founder and president of both Amazing Charts (Electronic Health Records) and (a medical website for patients with embarrassing health concerns). He is in private practice in Hope Valley, RI.