Tag Archives: rants

Anatomy of a Board Question

Posted on by 1 comment

In an attempt to displace some of the frustration of studying for boards, I decided to provide an insight into why USMLE STEP 1 is such a difficult exam for medical students – such that, on average, 5-6 weeks need to be dedicated solely to the task of preparation, in addition to the time spent throughout the course of second year.  The best way to illustrate the misery of this test is simply to describe the average question that may appear on it.  Multiply this by 322, and you have the 7-hour exam that is STEP 1.

To ease the pain ever so slightly, I’ll cast it tn the context of a history question:

In the 20th century, a prominent political figure became heavily involved in the production of a new automobile known as the “People’s Car.”  A few years later, this man’s favorite pet, “Blondi”, delivered several offspring.  He named one of them “Wolf.”  Of what species or breed was “Wolf”?
A)  Canis lupus
B)  Oryctolagus cuniculus
C)  Doberman Pinscher
D)  Felis catus
E)  Golden Retriever
F)  German Shepherd
G)  Thylacinus Cynocephalus
H)  Bradypus variegatus
I)  Bloodhound
J)  Chihuahua

Now, the first question you may have after attempting to answer this is “What does this have to do with history?”  Understandably, trivia is not a favorite pastime for most people.  If this is the case, the logical approach might be to try answering one or more of the separate components of this question, hoping that this would lead to the answer.  An automobile fan may recognize that “People’s Car” is actually an English translation for Volkswagen, and thus come at the question from this angle – who was behind the success of this company?  Sadly, there are no people in the answer choices.  “Well,” you say, “I can still get this.  After all, I know the genus and species of the Wolf – and Canis Lupus IS an answer choice.”  You would be right — and yet ultimately still wrong.  Who names their wolf “Wolf”?  At this point, you will probably throw up your hands in fury and, rather than continue attempting to answer the question, complain about why it was even asked.  Who cares about the pet of a car company owner?

Now, is the purpose of this question really to test your knowledge of cars or famous pets? Of course not.  Obviously, the question was testing your understanding of historical leaders, with the simple task of identifying Germany’s leader in the mid 1900s.  Because everyone knows that this was Adolf Hitler, it would be silly to have this be an answer choice – no one would get the question wrong, and the bell curve would be all nasty and skewed to the right.  Examiners worship the uniformity of the curve like a cult, and do everything in their power to maintain it – even if it means asking the name of Hitler’s pet dog, who happened to be a German Shepard.  What, you never learned that?  Clearly, you should have studied a couple more weeks.

With the example out of the way, one can begin to identify a structure; a method to the madness:

Diagram of a board question.

Anatomy of a board question.

Knowing the structure gives you some idea of what it takes to successfully work through a question, but it is even more useful if we distill this structure into a couple of High-Yield rules:

  1. The question and answer will rarely be what is explicitly asked.
  2. The obvious answer is almost always wrong.
  3. The seemingly logical answer is also generally wrong.
  4. On a rare occasion, the obvious or logical answer will be right.

Now, the first 3 rules can arguably be explained away as methods to test your critical thinking.  Of course, the 4th is there just to play with your mind.  Similar to the way your high school biology teacher would have 10 answers in a row be “C” on your multiple choice test, just to troll you.  Combine these rules with the multi-step nature of the question, in which you must jump through a series of mental hoops just to figure out what is being asked.  Now, add 6-10 answer choices and a time limit.  The result: a multiple-choice question in which a failure to successfully complete any one of the steps necessary to answer it results in failure.  Guessing doesn’t work out too well, either.

To prove that this isn’t all hyperbole, I actually have a real question to submit as evidence:

A sample USMLE STEP 1 question

A sample USMLE STEP 1 question

I hope you can see the resemblance to my example.  Perhaps the detailed analysis and rules actually allowed you to come to the correct answer.  Of course, knowing the rules only gets you part of the way there.  You still need to know that the Erythropoietin Receptor is a Type I cytokine receptor, meaning that it is associated with a non-receptor tyrosine kinase, and that a JAK2 mutation of this non-receptor tyrosine kinase will lead to a clonal proliferation of hematopoietic stem cells, the disease for which is called Polycythemia Vera.  No problem, right?  Or at least… nothing a few more weeks of studying First Aid can’t fix.

The problem I have with these questions, is that they’re essentially training us naive young physicians to miss the forest for the trees.  You spend so much time memorizing that the creator of the Volkswagen had a German Shepard named Blondi that you forget the fact that this individual was named Adolf Hitler and was responsible for the Second World War.  The goal of these so-called “Licensing Exams” has become badly warped.  Rather than asking questions that test knowledge a competent physician should have, similar to the way an FAA exam would ask student pilots about proper navigation techniques, the USMLE examiners have somehow decided that the purpose of their test is to stratify medical students into the most symmetric bell curve possible, regardless of the cost.  The intent behind many questions is “Let’s try to ask them something they might not know” instead of “Let’s ask something that they should know.”  I’m sure this approach makes residency directors happy, as they have a convenient 3-digit number available to sort applicants with, rather than relying on more subjective determinants of ability, such as clinical performance, letters of recommendation, and impressions made during interviews.  I suppose I don’t blame them – who in their right mind wants to read 1200 CVs, anyways?

USMLE Arms Race

Posted on by 0 comment

And so it begins.

Today marks the start of the dreaded ISP, or Intensive Study Period – that time when second year medical students attempt to refresh or re-learn everything from the past two years of their schooling.  We have just one month spend diligently cramming for the most important test of our lives.  The USMLE Step 1 is the first of three national licensing examinations for U.S. MD students, but its real importance lies in determining our career options.  Do well and the world is your oyster.  End up at the low end of the curve…well let’s just say your hopes of matching into a Dermatology or Plastic Surgery residency are rather stricken.  If you were planning to practice Family Medicine somewhere in BFE – well then it may just be your lucky day.


The real travesty of the high selection pressure imposed by STEP 1 is that it has come to resemble an information and economic arms-race not unlike the duel between the U.S. and U.S.S.R. during the Cold War – metaphorically, at least.  Medical students are some of the most competitive and hard-working people on the planet, but here this works against us.  A positive-feedback loop was created, in which students willing to pursue ever-more diminishing returns to memorize factoids that might appear on the boards slowly started to have an impact on average scores.  To compensate for this apparent grade inflation, the designers of the test had to find even more obscure, clinically irrelevant details to maintain their precious bell curve, and so forth.

Then the internet hit.  Now everyone has access to vast expanses of all these formerly secret and exclusive nuggets of trivia, carefully compiled by a host of new companies who have found themselves in the business equivalent of an all-you-can-eat buffet.  Desperate, anal-retentive med students represent the ultimate seller’s market.  $50 soft-cover test prep book?  Shut up and take my money.  $100 online pathology review lecture series?  Sign me up.  $300 question bank?  How could I hope to do well without it?  The list goes on.  I didn’t even mention the $1500-$3000 live prep courses available…ouch.


So now we are in a situation where an exam costing $560 requires an additional $400 at a minimum to prepare for.  I am unashamedly following the herd like a sheep in front of a cane, sporting my trusty 2013 edition First Aid ($40), Pathoma ($100) and USMLE WORLD question bank ($300).  Baaa.

I wish I could say it gets better, but sometime next year we’ll be taking STEP 2, an exam with both a written and practical portion (in Chicago, no less).  Cost?  Well, lets see – $1750.  Looks like the government will be making record returns on its 6.8% investment next year.  Pity it’s at my expense.

Category: Medicine | Tags: , , ,

My perspective on Automation

Have you ever worked or studied in a building with automated motion-sensing lights?  The kind that turn off after an absurdly short amount of time, forcing you to get up and make wild gestures, like some Native American performing a rain-dance; begging for electricity to make its return to the cold, fluorescent lights above?


How about using the “thermostats” in similar industrial settings – the kind that are conveniently placed in every room, yet seem to do nothing when you actually try to use them.  The ones at my school go as far as to paradoxically disobey commands, blowing a cold, air conditioned squall as soon as one dares to increase the setpoint to something above frigid.

Although I have since thwarted both of these sources of daily ire, finding the manufacturer’s  product manuals and methodically crippling the settings of the occupancy sensors that were at the root of their willful disobedience, the behavior of these “smart” appliances epitomize all that can go wrong when implementing Automation of any sort.

My Take

Automation is meant to save time and money, but should NEVER do so at the cost of convenience.  If you find yourself having to perform new or additional tasks after implementing a “smart” system, YOU’RE DOING IT WRONG.  As with most problems involving three different-but-related factors, you can pick any two to improve at the cost of the third. Want something that is both convenient and saves you time?  Expect to pay for it.

Using the automated lighting example, this solution would involve using a motion sensor to turn the lights ON only.  Thus, you would never have to use the light-switch – saving you time. The lights would never turn off when you didn’t want them to, maintaining convenience   Your electricity bill, however, would skyrocket.  It’s just not a perfect world.

Practically, the solution involves balancing the three factors, rather than pursuing just two of them and completely sacrificing the third.  Set a longer delay on the occupancy sensor before it turns off, and crank up the sensitivity to the point at which it can practically see you exhale a deep breath, or blink an eye.  It may still screw up, but not nearly as often.  The electric bill will still be higher, but only slightly.

Category: Technology | Tags: , ,