As I read Dr. Kenneth G. Adler's article “Successful EHR Implementations: Attitude Is Everything” [November/December 2010], I considered my own experience with electronic health record systems (EHRs). I've used three different systems and have always been committed to successful implementation, just as Dr. Adler recommends. After using one particular EHR for 18 months, I even spent over four hours with the vendor's representatives discussing useful and producible improvements that I believed were needed. The representatives listened intently, took notes and were very polite. In the following year, they implemented exactly zero of these improvements.

Vendors' most common excuse for not changing something simple is, “What if a different provider wants to arrange the problem list (or other feature) differently?” If they want users' suggestions, then they need to act on them or at least have a good reason why they won't. I feel somewhat displeased about paying $60,000 for a poor program and spending time trying to help the vendor improve it only to be charged for system updates that fix problems like bugs and logistical issues. As expensive as the updates are, you would think vendors could afford to pay us to help develop their products, not continually charge us for them. Adobe Photoshop is a much more elegant and sophisticated program, but it only costs $700. Why are EHRs so expensive?

These are the problems that make EHR software especially difficult to use:

First, many physicians recall producing – and many physicians still produce – paper-based documentation that includes, simply, diagnoses, treatment (frequently in shorthand) and any patient instructions, labs or other follow-up. This information can fit on a small sticky note. This approach has served physicians well for many years, so why don't EHR vendors take a similar approach? If I diagnose only “bronchitis,” that means my eye exam is normal. If the eye exam isn't normal, I make the appropriate additional diagnosis. The EHR I use now requires that I document all normal findings because the software does not respond properly to the diagnosis or problems handled in the office visit. There is currently no EHR that performs in the manner that most physicians are used to. I heard somewhere that the software is supposed to be tailored to the user, not the other way around.

Another problem is that although my computer's word-processing software can correct as I type, mark misspelled words and expand my abbreviations on the fly (even my cell phone can pick “probable” words as I text), my EHR offers few, if any, of these standard word-processing features – and it costs 600 times as much.

And finally, wouldn't it be nice if physicians did not have to invest as much time up front preparing their EHRs to do things the way that most of us do them all the time? Instead, we receive these systems as blank slates – to allow us to specify one of the supposed infinite number of ways to describe each medication, each procedure, each follow up, etc. How many different ways can you write a sig for amlodipine? As far as I am aware, 99.5 percent or more of the time, it is written as 1 tablet, by mouth, once a day. Why not build in defaults with the option to make changes later so that we can get to work?