Patients seem to love tests. I guess people love the
technology. Maybe there’s the allure of something shiny, mysterious, cold and
alien. Maybe it’s the trust in something shiny, mysterious, cold, and alien
over the all-too-familiar, warm flesh and blood. Clinicians are just humans
too, like patients… so how could they possibly know things the patient hasn’t
deduced themself?
Even a junior clinician who has reached a simple
diagnosis through their history and physical has heard the patients say
something like, “But how do you know? You didn’t do any tests?” What the
patient does not realize is that the
history and physical exam are tests as well. The palpation of the abdomen,
the inspection of a person’s gait, the travel history, the color of the stool,
the frequency of symptoms, etc. are all tests in themselves. However, they are
something that seemingly any other person could do... and you the clinician are just flesh and blood doing these relatively simple things.
Getting diagnosis can be like detective work. We think
highly of fictional detectives who solve the crime by paying attention to the
subtle, sometimes-under-your-nose details that others miss. It’s amazing reading about detectives who are able to
solve a crime by weaving a cohesive picture by seeing deeper into the mundane
details that at first seem unimportant… that is until the detective puts the
pieces together. So why is it the opposite of amazing when a clinician does
something similar and finds the culprit through collecting simple clues and
deductive reasoning?
The history, physical, and tests are just data gathering.
These are data gathering tools to bring the
clues into a cohesive picture together, like a detective trying to solve a
crime. All of these have statistical aspects: sensitivity, specificity, false
positive rate, and false negative rate. All
are pretty good in the right circumstances and all are flawed.
Often we consider the sensitivity, specificity, etc. exclusively
when we are talking about data we obtain through technology. As a clinician
gains experience, they realize that their history and physical exam have the
same characteristics as well. Many times this realization is intuitive, not
explicit.
For example in the case of acute appendicitis, a CT scan
certainly has a measurable sensitivity, and specificity. A small, older study I
found says both the sensitivity and specificity are around 90%. Another
small, older study found that the sensitivity of rebound tenderness was
94%.
For a more recent example, this study showed
evidence that for vertebra-basilar strokes the HINTS exam was more sensitive
than MRI. The study showed HINTS was 100% sensitive.
Putting the potential flaws of these studies aside, the
point is that we should treat the data
we obtain via history and physical examination just as we would treat data we
get from a machine. Whether from a history and physical or measurements
from a machine, we can assess the statistical aspects of the data we collect.
Data is still just data regardless of its source and it will have
its appropriate weight in the patient’s overall picture.
So whenever you question whether your history and
physical are enough, and question if you need to do tests, remember that your
history and physical exam are tests too.
I discuss this topic and more in my upcoming book on
clinical decision making. I expect it to be coming out later this year.
