Isn’t it strange how the incidence of certain pathologies always seems to increase right after you’ve read up on them? Think about it for a moment. At teaching, someone gives a presentation on aortic dissection, then suddenly you’ve diagnosed three in a week. You leaf through the guidelines on chicken pox in pregnancy and, even though you’ve never seen a case before, over the next few days you start to get worried that some spotty kid has really got it in for mothers-to-be. The sound of the latest St.Emlyn’s podcast on sepsis has only just left your ears and, what a coincidence, your patients just can’t get enough of being super-infected bug-breeding machines.
Sepsis is one of the things we see quite often, so that last one was a bit of a stretch (though a shameless plug for the podcast – have a listen later!). But, if you’ve been in this situation, and I’m sure at least three of you have, it could feel like you’re some kind of mystic. You might think one of two things. Either you can predict the future, knowing exactly which patients will walk into your emergency department and therefore what to learn in advance, or you can make people selectively ill based on your knowledge at the time. Hopefully it’s not the latter, and I’m sure you’ll be sad to hear that it’s probably not the former.
The Baader-Meinhof Phenomenon.
In 2006, a linguistics professor called Arnold M. Zwicky coined the term frequency illusion to describe the concept that for a short period of time after a person notices or reads something, they start to see it everywhere or believe it happens more often than it should. For example, seeing posters for a certain movie everywhere after someone tells you about it, or after driving your new car off the forecourt, noticing that everyone else on the road has also bought the same one.
About 12 years before it was given a proper sciencey name, a commenter on the St. Paul Pioneer Press’ online discussion group came up with “the Baader-Meinhof phenomenon” after hearing about the 1970’s left wing German terrorist group of the same name (which eventually became the Red Army Faction) twice in 24 hours. Frequency illusion is still more commonly known by this name.
The illusion draws together two processes, selective attention and confirmation bias. The first is something familiar to anyone who has teenage children, well known for only paying attention to things that are important to them, ignoring absolutely everything else. Confirmation bias is a tendency to look for things or interpret data to support our hypotheses, and disregard anything that goes against them.
Pull it together and you get a scenario where we ignore anything we don’t want to see, using what we do see to support our own conclusions. These influence what we do next. Whilst this may not be such a big deal in the world outside the emergency department, within it could be a different matter.
How is frequency illusion relevant?
As Emergency Physicians, we rely on our skills in history-taking and examination. We listen to our patients, ask questions, get answers, and sift through the important positives and negatives to rule in or rule out diagnoses on our differentials list. Most of our work comes from talking to our patients, and based on this we decide what investigations we order and what management to initiate.
We’re also constantly learning. Through examinations, on-the-job learning, teaching and education events, and even the odd thing we read online, there’s always something new to learn and take into our clinical practice. More often than not (or maybe it’s just frequency illusion within frequency illusion), you’ll feel like you’ve seen more cases of something after you’ve learnt about it than before.
Frequency illusion could account for a perceived increase in patients presenting with certain problems right after we’ve done some reading about them. However, there is another possible explanation: we just didn’t consider the diagnosis as a potential before. When we diagnose or learn about an interesting or rare condition, such as an aortic dissection, inevitably it will come higher up on our possible list of differentials for future patients presenting with the main symptom – in this case, chest pain. We may not see another for months or years, but certainly for the next few patients we will be on high alert. It may also influence our investigation and management of our patients. I suspect this would be doubly so if the diagnosis had been initially missed or not considered.
The incidence of a disease doesn’t suddenly multiply just because we’ve read about it though, and it’s also important to remember to still think horses when we hear hoofbeats, rather than leaping straight to a conclusion about zebras.
(The Baader-Meinhof Cavalry)
The dangers of illusion and bias.
When you pick up the next card in the box, it’s hard not to look at the triage notes before you see the patient, which can sometimes lead to you having an idea in your head what the diagnosis will be from this outset. Knowing what you want the diagnosis to be, whether it’s a simple viral infection in the patient presenting with sore throat, or a subarachnoid haemorrhage in the patient presenting with headache, could change how you approach the patient. Breaking frequency illusion down into its component parts, if a clinician looks to make a particular diagnosis, they may just hear what they want to. They may selectively listen to the parts of the history that fit, ignoring or steering away from those that don’t. This then seemingly supports the diagnosis of a disease, confirming to the clinician that they have got it right, leading to further investigations and a decision on disposition.
This could have knock-on effects when referring patients to other specialties, or discussing a case with a senior, who may be biased by the referring clinician towards a particular diagnosis. It can be hard to find a way around this without starting again from scratch. If there is any doubt, it’s always best to go back to the beginning and retrace the steps yourself.
Expectation can play a part in investigation as well. The Baader-Meinhof phenomenon could lead a clinician to increase their perceived pre-test probability of a particular diagnosis. This could then influence the interpretation of diagnostic tests. For example, the PERC rule takes into account clinician gestalt, with an assumed pre-test probability of less than 15%. A physician who has recently read a paper on pulmonary embolism and seen a case soon after may perceive a higher probability of embolism in their next patient. This could lead them to dismiss using PERC, and move straight to diagnostics and imaging, with a possibility of finding benign, incidental pathology.
How can we get around this?
First of all, knowing about these issues can help. Recognising the potential to seek validation of a pre-formed diagnosis can aid the clinician to actively challenge this.
We all take focused histories, but in the pressures of a busy department, sometimes they can become too focused. It’s important to recognise that when this happens, it’s easy to miss parts of the history that could steer you away from a particular diagnosis. Challenge yourself. Are you fixed on diagnosing a patient, or patienting a diagnosis? Ask questions that would not only confirm your hypothesis, but also contradict it.
If you do this, and something doesn’t fit, you still might be attached to your diagnosis, unable to think of anything else as a differential. If you can recognise this, then it’s important to ask a colleague for another opinion on the case. Of course, they may be biased by your thoughts on a diagnosis so let them know that you’re concerned you’ve missed something.
Next time you read up on something, just remember that you’ll probably see it more often for a short period of time afterwards. You might even make a few diagnoses. You might also want to diagnose it, and that could influence the way you interact with, and take a history from, the next patient presenting with a key symptom.
But, also remember that you can’t influence the incidence of disease (legally). You can get too attached to a diagnosis, and be inattentive to parts of the history that would otherwise make you reconsider. Subconsciously, you’ll always want to make an effort to confirm your own hypotheses, but knowing that your subconscious is doing this can help you to overcome it. A second opinion may confirm your diagnosis, but it could also help you see things in a different way.
Oh, and don’t worry if you’d never heard of the Baader-Meinhof phenomenon before, I’m sure it won’t be too long until you see it again.
You can read more about Prof Zwicky’s musings on illusions here.
Title picture from cal3star at deviant art.
Before you go please don’t forget to…