Understanding 2x2 epidemiology tables for
calculating effectiveness of tests
Both clinical practice and clinical or health studies are tied closely with “tests”. Clinicians
frequently order lab or radiologic tests, for example, to help them make a diagnosis.
Researchers also must perform tests for their studies. For example, to find study subjects with
certain diseases, researchers first need to perform tests to find who has the disease and who
does not. Or, to measure the effectiveness of a certain drug against a disease, they need to
perform tests to determine who has been cured after using the drug. The list goes on, but the
point is, tests play a very important role in clinical practice as well as research.
Now with such an important role that tests play, one might ask, how effective or accurate are
those tests? For example, let’s say you have a sore throat and go to the doctor. He does not see
any white or yellow spots on your throat suggestive of a strep throat, but other signs and
symptoms make him suspicious of strep infection. He suggests using a quick strep diagnostic kit,
which would give you the results right away. You agree to the test, and the result is “positive”.
Does it mean that you DEFINITELY have strep infection? The answer is no; you LIKELY have strep
infection, but that is good enough for the doctor to put you on antibiotics. Of course, a
“culture” would have been more accurate, but that would take 3 days, and in the trade‐off
between accuracy and speed here, the doctor has chosen speed.
In the example above, the fact that you LIKELY had strep throat, was enough for the doctor to
start therapy. But this is not always the case. Imagine a blood donor that gives their blood, the
blood undergoes routine tests, and the test shows HIV positive. Similar to the example above,
the screening test that was used here is not the most accurate test. The question is, is it enough
to consider that individual HIV positive? No! This is a much more serious situation than strep
throat, and you don’t want to label or treat someone based on “likelihood” in this case. You
want to be as confident as you possibly can. Therefore, you send the sample for a second, more
expensive, but more accurate test. Now let’s assume the result of the second test comes back
positive. Can you be ABSOLUTELY SURE that the person has HIV? Well, the answer is still no!
but you are now closer to 100% confidence. You can virtually never be 100% sure, but you can
get close to it. For this reason, the doctors will consider other signs, symptoms, tests, patient’s
lifestyle, etc. to arrive at a conclusion. Sometimes, they may decide to repeat the test a second
or even third time, to be sure.