Under my skin

There's a satisfying post on WhiteCoat where he rants against patient-reported allergies. A sample:

When I ask patients about their medical allergies, more often than not patients suffer from at least one. During a recent shift, I had 17 people who told me that they had medication allergies. When someone has an allergy, I always ask what the allergic reaction is. The responses I received included the following:

* Seven people had allergies to various medications (most often penicillin) because their parents told them they had a reaction as a child. They didn’t know what the reaction was, but they have never taken the medication since.

* Four people had nausea and vomiting with medications that typically cause nausea and vomiting as one of their side effects...

He's right -- a lot of people have unwarranted concerns about mild or entirely predictable reactions, and sometimes this can be frustrating on a busy shift. But I also like the commenter who wrote:

I’m not sure how you think the patient is supposed to know which things actually require medical attention, especially when doctors and nurses refuse to give any guidelines over the phone. 'Come on in, and if you aren’t seriously ill, then we can make fun of you on the blog tomorrow.'

Patient perceptions of allergies is a subset of a larger issue facing all of emergency medicine -- patient perception of disease. We don't expect patients to triage themselves, or figure out which symptoms are worrisome and which are benign. That's our job. I try to look at proper allergy reporting as another opportunity for patient education (my favorite is explaining why someone can't be allergic to the iodine atom).

More importantly, from the informatics perspective, allergy reporting is a big frustration as well (and one we can actually do something about, ourselves). Patient-reported allergies find their way into every EMR, and trigger the most inane alerts and stops, forever. If a patient reported vomiting once after codeine, every subsequent doctor who sees this patient will have to jump through electronic alert hoops just to order IV morphine. It doesn't matter if the patient is taking oxycontin and wears three fentanyl patches. The same goes for antibiotics -- I think most lay folks would be surprised that we have to wrestle, years later, with the inherited family warning of about penicillin reactions, even when ordering a 4th-generation cephalosporin with essentially no cross-reactivity...

There's no intelligence built into the system, yet, I think because everyone's afraid that if a patient has a bad outcome because that 14th medication alert was eliminated, they'd be liable. This line of thinking ignores the notion that bad outcomes are probably happening because there are so many useless alerts, they all tend to be ignored.

Someone told me recently (perhaps it was Dr. Reider?) that non-clinical folks involved in setting up electronic health information exchanges thought that communicated allergies to new providers would be the top priority, and were surprised when physicians considered allergies to be less important than, say, recent EKG's, imaging, current med lists, and the like.

I wonder if this attitude toward allergy records is because we don't think most allergies are that serious, because we can most often treat whatever arises... or because we're overcome with alert fatigue.

Whatever the reason, there's no doubt in my mind that if we had an intelligent, efficient system to process patient-generated allergy reports, we'd be less frustrated with this information, and more sympathetic to the patient's concerns.