Getting comfortable with Uncertainty

Medical students are usually taught a rather comprehensive approach to history and physical exam.  They appear to go through a mental check list of all the things that need to be asked and examined.  They start off by following this check list regardless of presentation.  They may also think that a test is needed to rule in or rule out each condition on the differential.  It takes them a while to realize that a good history and physical can diagnose almost 70-80% of unknown cases presenting in the out-patient clinic and that even when one is not sure of the diagnosis, one does not always have to order tests.

When they first show up in clinics students tend to be quite confused that the clinical world is very different from the one they were training for.  The experienced doctor seems to ask questions in a very different order than what they expected and seems to be quite comfortable with uncertainty, tending to order far fewer tests than they would have thought.

What can we do to ease the transition to the clinical years?

Some techniques I have tried are:

1. I teach a mantra that "Time is a diagnostic tool".  This is particularly true in outpatient medicine.  Some conditions can best diagnosed by waiting and watching.  Either they will go away or they will show new features which will help diagnose them.  
2. Empiric treatment is also a diagnostic tool.  Sometimes a short course of medication can help - if the condition gets better (or it does not) it can point to the diagnosis.
3. I give them a framework and ask them to try and place the patients possible diagnoses (from the differential) into one of 3 buckets (categories):
1. Potentially serious (e.g. life threatening) and urgent (needs to be diagnosed or ruled out quickly)
2. Potentially serious but not (very) urgent  - this is on a scale 
3. Likely not serious and not urgent

I then ask them to see if they can eliminate any condition on their differential that falls into category 1.  If that cannot be done, they need to absolutely do some thing right away including ordering a test.  If they can, then they have time and then they can try still order a test or try empiric treatment or wait and watch.

Forcing a student to try and place the likely possible conditions in these categories helps them ask the questions that matter rather than go down a checklist of HPI and ROS and PMH that are often irrelevant.

They students are often confused that they spend a long time with the patient asking questions and doing a head to toe exam and then when the preceptors talk to the patient, they ask one or two pointed questions or check a couple of physical exam findings that change the management completely (see table).  Discussing this with the student early in their clinical rotation can help smooth their transition and reduce their frustration.

Syndrome	Category 1	Category 2	Category 3
Low back pain	Epidural metastases (history of cancer), abscess (Systemic features of infection), (Spine tenderness), (Nocturnal pain)	Sciatica (Straight leg raise test)	Muscle strain (paraspinal tenderness, history of unusual exertion)
Chest pain	Acute coronary syndrome (previous stress test/cath,), aortic dissection  features of affected arteries or nerves e.g horner/recc laryngeal etc)	GERD, viral pericarditis (systemic features, rub, EKG findings)	Rib/muscle strain (reproducible tenderness)
Headache	Aneurysm (neuro findings check the pupils disc), Temporal arteritis (age, jaw claudication, eye symptoms, scalp tenderness)	Migraine (photophonophobia, triggers, caffeine)	Scalp hematoma from minor injury

The items in parentheses are just examples of types of questions or exam findings that may not be part of the standard check list that a student my use. These not meant to be inclusive of all such questions. Also there are several guidelines that students can be directed to e.g. the guidelines regarding which head injuries should get a CT scan.

[This is not to say that there is no role for the comprehensive H and P.  Often the students will discover something important about a patient that their primary care provider was unaware of.  This is especially important in someone who has an unresolved symptom/s even after being seen by multiple providers/consultants.  A fresh look at the case with a systematic approach can reveal clues to the answer.

Clinicians will sometimes miss a diagnosis but if the patient is appropriately instructed regarding any red flags and followed closely, the prognosis will usually not be any different.]

Making a case for importance of clinical skills - PollEverywhere + Human Histogram

Setting:A course for Educators in the Health Professions.
Some of the audience was international and some did not have a clinical background.
There were about 75 people in the audience.

I was trying to make a case for the importance of clinical skills in helping make the correct diagnosis - this was to be done in about 10 min before moving on to some tools to help teach clinical skills. What would be the best way to make a convincing case while engaging the audience?

This is what I came up with:
Step 1. 
Do a poll asking the audience
"What percentage of cases in an ambulatory practice can be correctly diagnosed with a good medical interview and physical examination?"
Here are the results of the poll (done on http://www.polleverywhere.com).

Create your own sms poll at Poll Everywhere

Step 2. 
Provide some references that show that the correct answer is 70-80%.

Step 3. 
Do a Human Histogram - I learned this simple but very effective tool from Ken Locke a clinican educator from Toronto - he used this at a recent Society of General Internal Medicine workshop.  The following slides show how I used it.  FIRST ASK EVERYONE TO STAND UP.  (make a wisecrack about this being "active" learning")

Everyone stayed standing

2-3 people sat down

Almost everyone sat down

I asked the audience to interpret what they had just observed.  There was no doubt in anyone's mind how one or two critical bits of information can help confirm a diagnosis and rule out another.  They also saw how medical interviewing is a very dynamic process of deductive reasoning and how the physician is constantly matching the story with patterns of disease and looking for pieces to complete the jigsaw puzzle.  This is very different from the checklist approach of history of present illness, review of symptoms etc. that we teach our medical students. 
This helped set the stage for discussing various options for making sure our medical trainees get to learn this deductive approach.

Looking back I was struck by how using a web-based polling system where the audience submitted their responses using laptops, smartphones and SMS and combining it with a simple classroom technique of the Human Histogram worked out so well. 

In case you are wondering, Poll Everywhere is a free tool for limited audience sizes and has some great features for submission of responses via web, sms, twitter, and for starting and stopping the poll, clearing results and sharing the data.  You can also embed the poll on a blog and share results with the readers.