The Haversian Canal

« CBL Powercleaning | Main | Covenant House in New Orleans »

June 5, 2006

What to expect on an NBME shelf exam

The NBME advises their question writers to write mainly what they call hinge questions, that is, the you are given some information, and you have to know the diagnosis in order to then answer the question. The anatomy of questions is generally something like this:

Misdirection, misdirection, misdirection, misdirection, misdirection, misdirection, misdirection, misdirection, misdirection, misdirection, misdirection, misdirection, misdirection, misdirection. Misdirection, misdirection, misdirection, misdirection, misdirection, misdirection, misdirection, misdirection, misdirection, misdirection, misdirection, misdirection, misdirection, misdirection. _____question, _______, question, ________ which _______________ is most likely ___________ question, question ____________?

A. Distractor

B. Distractor

C. Answer

D. Distractor

E. Distractor

F. Distractor

G. Distractor

H. Distractor

I finished the Biochem shelf with about 10 minutes to go but they'd already frozen us in our seats (before the 10 minute warning anyone who finished could leave), so I thumbed back through looking for trends. That's the trend. I'd say 85 to 90% of questions is used the phrase which is most likely in a compound, complex, or compound-complex question.

Unlike the MCAT, which is easily parsed using Bloom's taxonomy, I'm not sure Bloom will be of much use on the USMLE. Perhaps, but I need to review it some more. Virtually all questions require the test-taker to walk one step beyond the information in the question, the answer hinges on you knowing something that's not in the question. For example:

A thirty-year old man presents with dyspnea on exertion. Chest radiogram reveals an enlarged cardiac silouhette. If a viral cardiomyopathy is diagnosed, which hormone is most likely increased the least in the circulation of peripheral muscles on ELISA?

A. Renin

B. Angiotensinogen

C. Angiotensin II

D. Angiotensin I

E. Aldosterone

See, you really don't need to know, for the purpose of this question, what the symptoms of a viral cardiomyopathy are. That's misdirection. It might be helpful to appreciate that the heart is weaker, but, what you really need to understand is how the above hormones interact. Renin is likely to go up. Angiotensinogen production may go up, but there may not be an observable increase in concentration because the extra renin is converting it to Angiotensin II. I don't even know what Angiotensin I is off the top of my head. Aldosterone will go up if renin and ang II go up. Peripheral muscle, eh, not really useful. ELISA, eh, not really useful. But, also, do you see how this might really be the question you'd face in the clinic (okay, maybe not in the clinic, but maybe in a clinical trial)? Which enzymes might you test for to diagnose this patient? What results do you predict?

Also notice how a rather awkward reversal was introduced, which hormone is most likely increased the least, to preserve the is most likely to structure. It could have easily been rewritten as is least likely to, but that would mess with the QA people reviewing the questions.

Finally, don't bother copying somone else's answers. During my ten minutes of intel collection, I couldn't match any patterns from one answersheet to another. I mean, I was in the back of the room (Kaplan tactic) and stared quite conspicuously. I wasn't trying to change my answers, I was doing intel.

Also, why do you suppose the proctors give the tests to the students at their seats, rather than having the students come get the tests? Wouldn't that be more efficient? Well, think of all the times the proctors walked those aisles and rows, and looked at things. Casually. While all the students stayed still. Gives a good opportunity to catch conspicuous cheating, doesn't it?

Posted by Niels Olson at June 5, 2006 10:46 PM

Comments

Posted by: GruntDoc at June 11, 2006 1:15 AM

Post a comment