How can doctors minimize unnecessary testing?

A recent comment raised a minor controversy about the strategy of minimizing tests.  I actually do not think that the disagreement is that great, but I feel like exploring the issue.

This is the sentence that triggered the comment, courtesy of primary care physician Rob Lamberts:

Order as few tests as possible.  No test should be ordered for informational purposes only; the question, “What will I do with these results?” should always be answerable.  If it is not, the test should not be done.


Like the Supreme Court we have two options.  We can try to ascertain original intent.  I guess I could write Dr. Rob and ask him, but that would not be as fun.  Or we could develop our own interpretation of this paragraph.  I favor B.

The key here is “What will I do with these results?”

What are the possibilities from tests?

  1. We use tests to make diagnoses
  2. We use tests to exclude diagnoses
  3. We use tests to follow and adjust treatment
  4. We use tests to estimate prognosis
  5. Sometimes we order tests so that in the future we will have a baseline

The real question here is the phrase “informational purposes only.” Information is often valuable, it depends on the clinical situation and the patient.

I would amend this paragraph to include an avoidance of unnecessary tests.  What is an unnecessary test?  I would suggest that we should consider a test unnecessary when we already have the information, the same test recently, or the probability that the test would provide useful information is very low.

Perhaps one or two examples will help.  A patient sprains his ankle playing basketball.  You apply the Ottawa Ankle Rules and determine that the patient does not need an X-ray.   But you order an X-ray anyway.  That is an unnecessary test.

An 18-year-old woman comes to student health complaining of a sore throat.  She is coughing, has had no fever, has no exudates and no anterior nodes.  Doing a rapid strep test is unnecessary, the history and physical have excluded the diagnosis.

A patient comes repeatedly to the ER with chest pain.  The patient gives the same history each time.  We have 3 previous chest CTs that exclude pulmonary embolism, yet another one is done.  Not only is the test unnecessary, but it could actually harm the patient in the future.

One cannot easily develop rules for testing because the presentations that we see vary so greatly.  Test ordering becomes an art and the application of evidence.  I sometimes order tests for information, if that is the only way to gather the information.

We should clearly minimize testing, especially expensive testing and thus part of the art of good medicine is the appropriate use of testing.  So I will declare that Dr. Rob’s original intent fits that concept.

Robert Centor is an internal medicine physician who blogs at DB’s Medical Rants.

Submit a guest post and be heard.

View 14 Comments >