Cognitive Bias and Antibiotic Overuse

Piotr Marcinski/123RF.com

Attempts to correct the problem of antibiotic overuse have typically revolved around education, with very little attention paid to psychology. A Google search for ‘antibiotic overuse’ brings up articles from the CDC and Mayo Clinic with titles like “Antibiotics: Misuse Puts You and Others at Risk” and “Antibiotics Aren't Always the Answer.” To put it bluntly, this education and awareness tactic isn’t working. I’ll go out on a limb and say that doctors know that prescribing antibiotics for viral illnesses don’t cure them, and I find that patients know this as well. Yet the problem continues without any signs of ending. Awareness and education haven’t solved the problem. I think a new way of thinking is in order.

While I’m no psychology expert, I think that approaching the issue from a psychology and cognitive bias perspective would be more fruitful. Wikipedia defines cognitive bias as “a systematic pattern of deviation from norm or rationality in judgment, whereby inferences about other people and situations may be drawn in an illogical fashion. Individuals create their own ‘subjective social reality’ from their perception of the input.” Clearly, prescribing an antibiotic for a viral infection would be a “deviation from rationality in judgment.” But let’s think about how the “inference about [the situation]” was “drawn in an illogical fashion.”

As with most sub-optimal decisions, the problem starts with uncertainty. Obviously, a doctor wouldn’t prescribe an antibiotic for a common cold, and most patients wouldn’t accept an antibiotic for a common cold-- if in fact they knew it was a cold. Unfortunately, while patients aren’t concerned about colds, they are concerned that they have (or will have) a ‘sinus infection,’ and this is the root of the problem.

The website of the American Academy of Allergy, Asthma & Immunology states, “A bad cold is often mistaken for sinusitis (sinus disease). Many symptoms are the same, including headache or facial pain, runny nose and nasal congestion. Unlike a cold, sinus disease symptoms may be caused by bacterial infections. It often requires treatment with antibiotics (drugs that kill the germs causing the infection).” [http://acaai.org/allergies/types/sinus-infection]

What the ACAAI is trying to say in the above statement is that sinus infections are much less common than colds, even though they have similar symptoms. That fact, unfortunately, is buried by the rest of the statement, which essentially says, “If you have a sinus infection, you will need an antibiotic.” While this may technically be true, the likelihood that a runny/stuffy nose represents a ‘sinus infection’ (which often requires an antibiotic’) instead of a viral cold (which can’t be resolved by any intervention and will likely take another 7-10 days to resolve) is far under-emphasized. I do my best to explain this to patients, but the lack of any pathognomonic signs/symptoms or diagnostic tests to rule in a cold and/or rule out a ‘sinus infection’ makes it a difficult concept for them to understand. I try to explain that common colds are ‘common’—much, MUCH more so than a bacterial infection spontaneously developing in the sinus cavities, but I am not always believed. Can I definitively say that their symptoms aren’t because of a sinus infection? No, I can only say a sinus infection is very unlikely. Unfortunately, that is not enough for many patients, and they will still insist on an antibiotic. This falls under a type of cognitive bias: “neglect of probability.”

Couple this with another cognitive bias: illusory correlation. As the name implies, illusory correlation is the phenomenon of perceiving a relationship between variables, even when no such relationship exists. This is the basis for the placebo effect. If someone has a cold, sees a medical professional who prescribes an antibiotic, and then over the next couple of days notices that the symptoms improve-- well, invariably they are going to attribute the improvement of their symptoms to taking the antibiotic (instead of other factors, i.e. the natural improvement over time or the NSAIDs/decongestants/other medicines intended for relief of symptoms that were suggested by the doctor at the same time). Many times I’ve been told, “The last time I had these symptoms, my doctor told me I had a sinus infection. I felt a lot better a day after starting the antibiotic.” I am skeptical that symptoms of headache, runny/stuffy nose, sinus pressure due to bacterial infection would improve so dramatically and so quickly once the antibiotic was started. But getting a patient to understand that they don’t necessarily need an antibiotic the next time (and probably didn’t need the first time) they have similar symptoms is a near impossible task.

So, the problem of an antibiotic being prescribed in the first place deserves a closer look. More on this another time.