Saturday, April 13, 2013

How much data should you give in a medical education case presentation?

Many people like to use cases to teach principles in medical education.  This makes a lot of sense as it highlights how whatever information being presented will be relevant even to a busy practitioner.  Clinical cases are also commonly used in assessments to assure that students and other learners understand a concept as it applies in a clinical setting.  Most of these cases whether taken from a real case the presenter has seen or whether it is made up by the presenter has a lot of information in it.  This is to help lead the learner usually towards a specific diagnosis.  I think there is value to this approach, and I think it can be used appropriately.  However, I have noticed a paradox being set up with the amount of information presented in a case in a presentation or assessment.  The paradox is the more information presented, the less broad clinical thinking is, and the more the case focuses on a single disease state, and the less it focuses on clinical reasoning.

To begin my explanation of this, let me explain what I have commonly seen, and thus what people expect from a case presentation.  Usually a fairly complete history and physical examination is presented with the level of detail normally encountered in a chart note.  Sometimes the information is irrelevant, but typically there are key pieces hidden in the data to allow the person to find the correct answer.  The goal of the learner is to find the key piece of information which will move them to the correct disease.  One example would be a kid with developmental delay who is presented with a really long history.  Somewhere hidden in the history is the physical finding of a 'cherry red spot'.  For those who are not neurologists, even years after taking boards, there is likely a monosynaptic arc in your temporal lobe which automatically just screamed, "Tay Sachs disease."  If it didn't scream this before, it has now been re-cemented in your memory banks to reside there from many more years.  The trouble with this is that it can lead to premature closure - most neurologists, especially pediatric neurologists, would argue that there are actually a number of other conditions with a 'cherry red spot,' and if all you think it Tay Sachs, you may miss the actual diagnosis.  In this case, the overy complex medical case presentations (or books like 'First Aid') may be doing a bit of a disservice as they are designed to help you pass MCQ tests, and not to help necessarily in the clinic.  I'm not arguing the all medical knowledge is bad to know, nor am I arguing that a well-developed schema of important differentiating features of diseases is irrelevant.  In some cases, learning about a single disease entity is the goal of the learning session, and this may be entirely appropriate.

What I am arguing is that we should consider using cases which don't have the hidden key word which points to a specific diagnosis if you are trying to teach about a broad diagnostic topic like developmental delay in a kid.  I would argue that a very skeleton case with even just a chief complaint, duration of illness, and cursory demographics can be effectively used as a case presentation.  This works well for either introducing a topic or as an assessment to be sure your instructional experience has been successful.  The first thing I hear from audience members when I put up a slide with not very much information is, 'I need more history, and a physical.'  Why?  I think part of it is that it is true that you do need more information, but part of it is they are not used to a format of learning where the answer is not supplied on a silver platter.  Think back to how many case presentations you have been given that didn't have that hidden factoid in it somewhere.  But, I'm not trying to teach about the factoids with these cases, I'm trying to help people learn about the depth of their illness scripts.

In other words, everyone knows that an expert clinician in a particular area has an ordered differential diagnosis in their head way before the hidden factoid is presented.  However, we don't spend much time in education with ambiguity.  We like to spend a little time on the differential, and a lot of time talking about the really rare diagnosis or rare variant of a disease.  I'd like people to spend more time on thinking about how they formulate a differential.  I'm more interested in what my medical students (or other learners) can give me as a differential diagnosis for a person with a chief complaint of parkinsonism, than whether they can ferret out the early apraxia and language dysfunction and correctly diagnosis CBD (corticbasal degeneration).  What I'm more interested in is can they come up with an ordered reasonable differential with a little bit of information, and decide which questions and physical exam findings will be needed to work through that differential.  In doing so, we build a much more robust illness script for those diseases.  Now I have a script for developmental delay for Tay Sachs which is deeper than just 'cherry red spot.'  I think we just need to use these techniques more as medical educators so our audience comes to expect working with ambiguity and not certainty.  The main reason for this, is that real life is ambiguous.  Real patients are ambiguous.  And in real life, I'm honestly not sure my fundoscopic exam skills would be good enough in an toddler to actually see the 'cherry red spot.'  So, if all I have is that in my illness script when presented with a kid with developmental delay, I'm hosed.


  1. Hello,

    I came across your blog post through @nlafferty. I've been reading about the language we use in case presentations recently for some research I'm writing up. I've been thinking about how medical students see what patients write online in forums and blogs so it is interesting to contrast that with the kind of medical talk that happens in presentations.

    Research such as this by Apker and Eggly takes a sociological view of the discourse. So it is really interesting to see you approach this from a pedagogical standpoint. Do you know any research in this area?

    Many thanks,
    Anne Marie

    1. Thanks for the comment. I'll have to look at @nlafferty's blog post that you sent me on Twitter. I'd start by reading this by Judy Bowen - It's on clinical reasoning and diagnostic reasoning. I probably should have cited her work in retrospect as I've talked with her about these ideas before, and she's done nice work on clinical reasoning. The internal medicine morning report is set up to try to highlight these principles.