|Seven Sloppy Uses of Medical Tests
Gary E Cordingley, MD, PhD
Medical tests are valuable in diagnosing illnesses and monitoring treatments.
However, used unwisely, they can cause more harm than good.
Even excellent tools can be mis-used. Here are seven "sins" of medical testing:
1. Ordering the wrong test for the right condition.
If I had a nickel for every time a doctor ordered a carotid artery test in a patient with a fainting spell, I could fund
my retirement several times over. And this is despite the fact that problems with the carotid arteries (the pulsating
blood-vessels in the front of the neck) are incapable of generating fainting spells! Narrowed or blocked carotid
arteries are capable of generating many other symptoms -- including paralysis on one side of the body or loss of
speech -- but not unconsciousness. Yet this test is frequently ordered in a knee-jerk fashion for people with
fainting spells. Moreover, when the artery is found to be narrowed, it sometimes triggers a needless and risky
operation on the affected artery. All because of a test that shouldn't have been ordered in the first place!
2. Treating the test instead of the patient.
There are situations in which a tool gets confused with a goal. One example of this is in the treatment of people
with epileptic seizures. Most people with seizures do well with the help of seizure-suppressing medications. The
amount, or level, of some of these medications can be measured in the bloodstream and there are circumstances
in which it is useful to do so. A drug level can be a useful tool. But it's only a tool, and nothing more.
The goals of seizure treatment are simple -- no seizures and no side-effects. What could be more
straightforward? However, some physicians appear to believe that the goal of treatment is to produce a certain
drug level on a lab report. When this occurs, trouble can ensue. For example, a patient might be doing great on a
certain dose of a medication that stops his or her seizures without causing side effects. (How can one improve on
that?) But then a doctor, ordering a drug level because it seems like the right thing to do, feels compelled by the
number appearing on the lab slip to lower the dose of medication. When this occurs, a seizure sometimes results.
This is a seizure that didn't need to happen.
3. Using a test as a substitute for interacting with the patient.
I have great respect for emergency physicians. Having done emergency work myself, I know it's not an easy job.
Emergency physicians work in a fish bowl, subject to criticism and second-guessing for decisions made in crisis
situations and under pressure of time. That said, one gains the impression that sometimes they order thousands
of dollars worth of tests based on a 30-second interview and a cursory exam. Yet there are cases in which, if a
few more questions had been asked of the patient or family, the diagnostic possibilities and choice of tests would
4. Ordering irrelevant tests.
There are certain tests -- like a chloride level in spinal fluid or blood-levels of some of the newer
seizure-preventing drugs -- that are not known to be useful for anything. But they get ordered anyway.
5. Forgetting that tests are imperfect.
All tests -- from high-tech scans to lowly blood measurements -- have false-positives (overcalls) and
false-negatives (undercalls). But sometimes test-results are handled as if they're perfect and never wrong. As an
example, sometimes patients have attacks for which the descriptions are compelling for a diagnosis of seizures,
but then have normal electroencephalograms (brain-wave tests). Electoencephalograms can be very helpful, but
it's possible for a patient who really does have seizures to have a normal tracing. Yet it's not unusual to encounter
cases where patients' normal brain-wave tests kept them from receiving the treatments they needed.
6. Forgetting that there aren't tests for every medical condition.
When patients report hard-to-diagnose symptoms to their doctors, medical tests are often ordered. Sometimes all
the test-results are normal. Does this mean there is nothing wrong with the patient? Not necessarily. There are
many conditions -- such as migraine, Parkinson's disease, fibromyalgia and restless legs syndrome -- for which
conventional tests show no abnormality. We just don't have tests for everything. So it can happen that the tests
are normal, but the patient isn't.
7. Failing to order tests that could affect treatment.
One axiom of medical management is that a test should only be done if its different outcomes would lead to
different plans of action. If the plan of action is the same no matter how the test turns out, then why do the test?
There's a flip side to this axiom. If a test's different outcomes would indeed lead to different plans of action, then
the test really should be done, or at least be strongly considered. So, when it comes to ordering a test, there can
be sins of omission as well as sins of commission.
It is tragic when a patient develops progressive memory loss and confusion. But it's even more tragic when it is
assumed that the cause is Alzheimer's disease (for which there is no good treatment) when it's really due to
something else for which good treatment is available. A risk-free head scan and a small assortment of blood tests
can check for a number of curable conditions, but sometimes these tests are omitted.
(C) 2006 by Gary Cordingley