Medical tests are performed for either of two possible reasons-diagnosis or screening. Diagnostic tests are designed to investigate symptoms or monitor illness. Screening tests are performed to detect disorders when no symptoms or known illnesses are present. They are often done as part of routine preventive care. In this article, we'll explore some pros and cons of screening tests. Let's start with a hypothetical scenario:

Imagine you are planning to undergo a relatively minor surgical procedure. Your physician orders a routine preoperative chest x-ray, even though you have no symptoms of a chest condition. In other words, the chest x-ray is being performed for screening, not diagnosis.

To your surprise, the x-ray shows an abnormality. Your physician postpones surgery while additional testing is performed-perhaps blood tests, a CAT scan and maybe even a biopsy. Finally you are told that the original apparent abnormality was a false alarm. How could this have happened? Shouldn't you expect medical tests to be more reliable?

Well, not necessarily. All tests have weaknesses as well as strengths. When deciding to perform a test for screening, whether it is a chest x-ray, a mammogram, a bone density test, a PSA test, an EKG, blood chemistries or anything else, several factors must be considered. Let's look at a few:

Accuracy
The accuracy of most tests can be expressed in complicated statistical terms, but the basic concepts are simple. First, an accurate test should be able to correctly identify an abnormality when present, by showing a positive result. This is called its sensitivity. Second, the test should show a normal or negative result in individuals who don't have any abnormality. This property of a test is called its specificity.

In our hypothetical example, it turns out that chest x-rays lack sensitivity, specificity or both for many of the lung or heart conditions they are designed to identify. In other words, as tests go, chest x-rays are not highly accurate.

The probability of an abnormality
Suppose a healthy high school athlete were to undergo a cardiac test that showed a previous heart attack. Even if the test was highly accurate, the odds are that the result was wrong anyway because a heart attack in a healthy teen is very rare. A mistaken result of this kind is called a false positive.

Similarly, a screening chest x-ray done on someone with a low chance of a lung or heart disorder will often yield a misleading result. That's why experts generally agree that chest-rays should be performed selectively, not routinely.

The seriousness and treatability of the condition(s) that might be detected
Cancer is the best example of a serious condition for which screening tests are often performed. If there is a chance of finding an unrecognized early cancer, even an imperfect screening test may be worthwhile.

The Pap smear for cervical cancer is an excellent example. A single Pap smear is not very sensitive-i.e., it often fails to detect precancerous or cancerous cells. Yet regular Pap screening has reduced deaths from cervical cancer by 70% over the past 50 years.

Screening chest x-rays, by contrast have not been successful at reducing deaths from lung cancer, even if performed on smokers every year.

The risks to patients
Most screening tests themselves pose little harm. However, the steps taken in response to abnormal results may well cause harm. First, a false positive result can lead to unnecessary but hazardous invasive testing. Second, harm can result from overdiagnosis of a condition that would not cause problems.

This latter concern fuels the continuing debate over prostate cancer screening. Autopsies reveal prostate cancer in 30 to 70% of elderly men who died of something else. Detecting these cancers during life would have often have led to risky surgeries or radiation treatment for no benefit.

The cost of screening
Even highly accurate and safe tests may still be too expensive to perform widely. For example, cost is a big issue in colorectal cancer screening. The stool test for occult blood is cheap but not very accurate. Supplementing the stool test with flexible sigmoidoscopy improves accuracy but adds cost. More expensive still is full colonoscopy, even though it is clearly superior to the other two methods.

The readiness to act on an abnormality
A screening test is senseless if there is no intention of treating an abnormal finding. For example, even though annual mammograms in women over 50 are important for early detection of breast cancer, it would be inappropriate to perform mammography on a woman with a terminal illness.

With all these factors to consider, how can we know which screening tests are appropriate for prevention and health maintenance, and which are not? Part of the answer is to consult one's physician, because each of us is different and no single rule applies to all. In addition, there are many resources for educating oneself about this subject; an excellent one is listed below. In any case, it's wise to remember that just because a screening test can be done doesn't mean that it should.

Where to get more information:
The United States Preventive Services Task Force, part of the Agency on Healthcare Quality and Research (AHRQ), is a panel of experts who review the evidence for screening and preventive measures. They can be reached at (301) 427-1364 or online at http://www.ahrq.gov/clinic/uspstfix.htm.

May 2004