As a physician who trained in the days of paper, started programming on mainframes, has practiced in myriad institutions including a vertically integrated multispecialty system with and EMR, I agree with everything the doctor wrote and strongly recommend reading it.

Oh, and I now work for a health IT company.

The fact is that electronic records have numerous liabilities and the article mentions only a few. Computers are really just huge adding machines, and deal best with massive computing and sorting of numerical data. A medical record is not mathematical. It contains a lot of quantitative data but it is first and foremost a narrative, a story, with most of the embedded data being analog and "fuzzy" rather than discretely digital.

Forms unfortunately warp substance and digitizing the record shifts the practice away from clinical medicine.

Like fascism and communism in the 20th century, secession in 1861, and prohibition, the EMR is a bad idea whose idea has come--that is to say foolish inevitability rules here.

Not the comment about typewriters in the Smithsonian. Many offices keep a typewriter around. Why? Because the typewriter is less susceptible to failure, and is much more flexible and faster for filling out short forms.

Does the pizza parlor guy need to record not just a few sales numbers, but a description of each customer, a summary of their conversations, details of their purchase and how they ate it, and weave that together in a narrative that they can recall in couple seconds of review on that person entering the store years later? When he has to do that, and finds a computer system that works for him in doing so, then we will have a valid comparison. Oh, and he also has to be able to be sure of presenting an unaltered and original legal record upon being sued decades in the future when the hardware with which it was originally created has been changed several times.