Answering USMLE Type Questions – Part II

Share if you like this post
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  

We will now continue with Part II of our series on “Answering USMLE Type Questions”. In Part I we discussed the 3 most common problems people have with USMLE type Questions. In part II we will discuss various strategies to correct these weak points. There are strategies you implement during the preparation phase and other strategies you do during the exam itself.

However,  the strategies during preparation is more effective than those you employ during the examination and therefore it makes sense to fix these problems before the actual examination day.

The main challenge in clinical vignettes is to be able to diagnose the case as fast as possible. As a medical student, the way we learned clinical cases starts with a diagnosis, say Myocardial Infarction. Then we study the signs and symptoms that accompany myocardial infarction followed by laboratory tests that suggest or confirm the diagnosis. Lastly we study therapeutic interventions. That is how clinical vignettes are presented in Underground Clinical Vignettes, which is fine for medical students having their first taste of medical cases. Unfortunately for most people who are ready to take the USMLE, they’ve already gone through that process and will probably need only a short review and it will still not help them with diagnosing clinical vignettes. The reason is that clinical vignettes are presented the other way around with signs and symptoms as clues while you come up with the diagnosis.

Which is why when you start your clinical rounds, you are taught a good method for diagnosing cases which is through the use of differential diagnosis. Doing differential diagnosis is an effective way of tackling any clinical case and coming up with a diagnosis. But for the purpose of answering the USMLE clinical vignettes, it may not be a good strategy. The reason mainly being the extraordinary amount of time it takes to build and run through a series of differentials and come out with the diagnosis. The USMLE gives you around a minute to do that and still come up with an answer to the question itself. Don’t get me wrong, you need to learn how to do differentials to be effective in diagnosing clinical cases, but for the purpose of the USMLE, it is not enough.

You must be able to do what experienced clinicians usually do in clinical practice. It is called pattern recognition. Contrary to expectations, most experienced clinicians do not do differential diagnosis when they are faced with a clinical case, they recognize a pattern of signs and symptoms and jump to the diagnosis in most cases. However, it does not mean they do not know or never do differentials. For difficult and unfamiliar cases, they do revert to doing differentials, but since about 90 to 95% of clinical practice deals with common cases, they rarely do differentials in actual clinical practice.

So how do you get to the point of doing differentials to being able to do pattern recognition, reserving differentials for more difficult and rarer cases? For most clinicians, it is usually secondary to experience. After year of seeing thousands of cases, they can easily see the patterns of signs and symptoms that indicate a particular diagnosis. However, relying on pattern recognition alone in clinical practice can result in malpractice and even fatalities. It is one of the most common reasons why extremely competent and experienced doctors can fail miserably and misdiagnose cases. This is especially true in critical cases. For example, suspected cases of Acute MI must always be differentiated from dissecting aortic aneurysm. Failing to do so can prove fatal as in the case of actor John Ritter. I have a family friend who is a pediatrician who died of undiagnosed acute MI despite the fact that she was in the hospital at that time being treated for acute cholecystitis under the care of a team of physicians one of whom is the top cardiologists in my country.

Being able to use pattern recognition is also dependent on the prevalence of a disease in your practice. In the case of the USMLE, it also means how common a particular case will appear in the examination, i.e. how high yield that particular topic is, which is different for Step 1, Step 2 CK and Step 3. As I have stated in other posts, high yield topics in Step 1 includes cases which may not be common but illustrates important basic science principles. For example, oat cell carcinoma of the lungs is more important in step 1 than squamous cell carcinoma because oat cell carcinoma demonstrates the important principle of paraneoplastic syndrome. In Step 2 CK and Step 3, the reverse is true with squamous cell  carcinoma more important because it is more common.

Which is why, some of those medical questions asked during residency interview can be so ridiculous. In one interview, I was told to diagnose a theoretical case they will give me by asking relevant questions on history and physical examination findings. I was specifically told that this case is fairly common in the hospital and there is no reason that I would not have encountered this in practice. The case was a female, who had a prolonged partial thromboplastin time. Of course the first thing that comes to mind are intrinsic coagulation factor deficiency of which the most common is factor VIII. However, being female eliminates Factor VIII and IX. So I thought of anti-phospholipid syndrome. However, all other findings were normal, no history of thrombosis or even failed pregnancy. If I were probably a student without clinical experience and the interviewer had not emphasized it was a common medical condition, I would have run through all coagulation factors in the intrinsic pathway, all of whom are quite rare. So when he finally announced it was a case of factor XII deficiency, I can’t help but feel either he is very, very stupid, or very, very sadistic. Factor XII deficiency has an incidence of 1 in 1,000,000. Therefore, there will be at most 300 cases in the whole of US. Now even if all the cases is concentrated in Brooklyn, New York. 1 in 10,000 does not make it fairly common. However, the incident is not reflective of my competence as a physician, but on him and the institution he represents.

So, if you have extensive clinical experience and have been in practice for quite some time, you should have superb pattern recognition skills and minimal difficulty with clinical vignettes. If you are a medical student with very limited clinical experience, you may still be struggling with doing differentials and therefore, clinical vignettes can be a significant source of difficulty for you. Most people are probably somewhere in between.

The basis for both differential diagnosis and pattern recognition are what I call key words, or words that will evoke a diagnosis or a group of diagnosis for consideration. For example, given microcytic anemia, you will think about all disease process that presents with microcytic anemia. Then if you encounter a second key word for example blood loss or equivalents of blood loss (young women without iron supplementation or old guy with signs and symptoms of colonic CA, which can be +FOBT, alternating constipation and diarrhea or even a family history of polyps) this should evoke iron deficiency anemia. The presence of increased Hgb A2 should bring to mind thalassemia trait. In pattern recognition, you take in the different signs and symptoms and together you are able to pinpoint a specific diagnosis without going through the step by step process in differential diagnosis which is time consuming.

In the USMLE where you have limited time to make a diagnosis and then answer the question which is usually about pathophysiologic mechanisms or other basic science concept related to the disease concerned, doing differentials is too time consuming. In my own experience, I was doing pattern recognition in about 80% of clinical vignette cases, resorting only to differentials in less than 20% of clinical vignette cases. That can save an inordinate amount of time.

So what are the steps you can take to prepare yourself to better able to handle clinical vignettes? Actually we will follow the natural steps we take in becoming competent in handling clinical cases as we outlined above. However, we have to consciously drill ourselves using study tools instead of depending on time and experience to teach us the skill.

The first step is to know the signs, symptoms and relevant laboratory results that define a disease entity. In medical school, we usually encounter these as we study individual diseases in pathology and microbiology and immunology in basic science as well as in the clinical sciences. For the purpose of the USMLE, you need to accelerate the process. However, your only choice currently is to either make your own study tools, or use Underground Clinical Vignettes which at between US$ 16.95 to US$ 19.95 per book with a total of 9 books for Step 1 is quite steep and outrageously expensive. Which is why I am building a High Yield Clinical Vignette course in my prep site at http://prep.askdoc-usmle.com. As of this writing, this portion of the course is still being constructed. Look to my blog for future announcement. It will eventually cover over 900 disease process with their clinical description.

Share
Like This Page!
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  

2 thoughts on “Answering USMLE Type Questions – Part II

  • July 6, 2009 at 6:25 pm
    Permalink

    once again great analysis of the enemy
    and as u said in your e-book that know your enemy
    it’s only u on the whole net (as far as my links) who descibes this enemy so greatly that has no match and also how to deal with it..
    keep up the good work ,it really helps..

  • January 21, 2010 at 11:21 am
    Permalink

    Hi,

    I really like your work, I am wondering do you think it is true to assume that pattern recognition is a quick lite differential diagnosis?
    I was also wondering do you think that pattern recognition is sufficient for step 2 and 3?

Leave a Reply

Your email address will not be published. Required fields are marked *