Introduction

Reviews

Think Twice About Surprising Results

Timothy B. McCall, M.D.

When evaluating surprising test results, consider that an abnormal test result is sometimes a “false positive—” that is, abnormal even though you don’t have the condition in question. False positive results might more appropriately be called false alarms, because unexpectedly abnormal results can alarm both doctors and patients. False positives can also lead to unnecessary tests or treatment. Let me illustrate with a case from my own family.

My sister Debbie called me one Friday a few years ago, crying so hard she was coughing. She was expecting her second child and had just been informed by her obstetrician that her baby might have Down’s syndrome (the birth defect formerly referred to as mongolism). He recommended she come in the following Monday for an amniocentesis, in which a long needle would be passed into her belly to remove fluid for genetic testing of the baby. Debbie had already agreed to the test before calling me.

Her doctor had informed her that one of the blood tests, the alpha fetoprotein (AFP) was low, which suggested Down’s syndrome. Debbie was devastated, worried about her baby and facing an invasive procedure which itself causes a miscarriage as often as one out of 200 times—even in healthy pregnancies. I advised her to cancel the amniocentesis, until she had the AFP level repeated.

Debbie took my advice. She went in the following Monday and had the blood test re-done. The result was normal and she never needed an amniocentesis. The further result was my niece Jacqueline, now a shy young lady with break-your-heart brown eyes.

If a test gives an unexpected abnormal result, repeat it before proceeding to a more invasive procedure. This is the rule-of-thumb I applied when I advised Debbie to cancel the amnio. If you have any questions about the accuracy of the lab where the test was done, consider repeating it elsewhere. Of course, something can go wrong with the second test as well. If two results differ greatly and a lot is at stake, you can repeat the test again.

There was, by the way, another problem with the doctor’s recommendation of an amniocentesis: a safer, less invasive test should have been done first. AFP levels generally rise during pregnancy. If Debbie’s doctor had overestimated how many weeks pregnant she was—a common error—he might have expected a higher AFP level. A recent editorial in the New England Journal of Medicine recommended that women with low AFP levels undergo an ultrasound examination to be sure the doctor hasn’t inaccurately estimated the date of conception.i According to the authors, with the much-safer ultrasound exam, half of all women with low AFP levels can be spared an amniocentesis.

Many things lead to inaccurate results. You might ask What happened with the first AFP test? It’s hard to know for sure but there are several possibilities. The calibration of the machine could have been off at the time the first specimen was run, leading to the error. Perhaps another patient’s blood tube got switched with my sister’s. Maybe a technician copied a number incorrectly. All of these things occur, more often than most people are aware.

Other factors can affect the results of various tests: how the technician drew the blood, how long the tube sat on the counter before the test was run, when you last ate, if you just exercised, what medicines you have taken in the last few days. For most tests, no adjustment is made for your age, sex or size, even though they can affect results. A common error is to assume that the elderly will conform to the “normal range” determined for younger people.

A false positive result is not necessarily due to lab error. For many diseases, there is an overlap between the test results of the sick and the healthy. A line between abnormal and normal must be drawn somewhere and the decision where to place it is based on statistics. A cutoff is often chosen at a level where a normal person will get a normal result 95 percent of the time. In other words, five percent of normal people will get an abnormal result. For some, the result will be abnormal every time they take the test. This does not necessarily mean that they need more tests to figure out the reason. If the likelihood of a problem is remote, sometimes nothing further need be done.

Next: False Positive Tests

Return to Examining Your Doctor