The High Cost of Specificity

When entering a drug allergy in an electronic medical record, you have to click on the name of the drug, you can’t just type it in. You usually have to start typing, but then when the name is suggested, you must click it, because today’s EMRs can’t read. That name you click on, which looks just like the name you intended to finish typing, is linked invisibly to an identifier you can’t see.

This is of everyday clinical importance, because this is how the computer knows and tries to stop you if you try to prescribe a drug your patient is allergic to. If the computer was smarter, it would be able to read your typed PENICILLIN instead of the hidden code behind it that is how the computer really knows what you are trying to tell it. And not even a computer that knows English would get it if you misspelled the drug name as PENICILLIAN or something else just slightly different.

But there is another kind of specificity that really doesn’t help us primary care doctors much at all. It may be of help for some specialists, but the real purpose of it is data mining for actuarial or research purposes. For example, if people who have had a certain operation done a certain way end up with a very specific problem years later, researchers can figure that out by querying EMRs for such connections. And insurance companies can adjust their premiums for health, life or disability policies based on such data.

I only need to know that my patient doesn’t have his gallbladder or appendix anymore in case he develops belly pain or that he had some sort of revascularization of his left leg in case he gets pain in either leg down the road. I personally have very little use for knowing how big their abdominal scar is; laparoscopic or open -ectomy. If I need to know, I can always look.

So when I try to enter a past medical history, I want to be able to freetext it because it is so much quicker than starting to type something about revascularization of left leg and be given a list of choices that are usually abbreviated to such a degree that as a generalist physician, I may not even know what they’re talking about, and what the distinction is between the many choices.

This specificity obviously makes life easier for researchers and people planning or budgeting, but it makes my life harder. Nobody is building time into my day to do what works for them. It takes me longer to do it their way, so I have to choose between spending less time talking with my patients about what they came in for, or taking work home with me at the end of the day.

Either way, I bear the cost of building a database with little utility for me and my colleagues on the frontlines of primary care.