I’ve been writing a lot about CT scans lately. Why are so many being done, and which children with head trauma really need one? CTs are really kind of neat—it’s amazing that we can peer into your body to see what’s going on in there. But like every other medical intervention and treatment, there are positives and negatives, pros and cons, a ying and a yang.

I’ve mentioned some downsides to CT scanning: they cost a lot, and often reveal incidental things that though technically “abnormal,” are meaningless. But they still cause anxiety and further costs and more CT scans for follow-up! Still, the most important problem with CT scans from a public health point of view is that they involve exposing the patient to ionizing radiation, with a resulting increased lifetime risk of cancer.

CTs use x-rays to peer through your body, the same x-rays that are used to make plain x-ray pictures. But with a CT, a whole series of x-rays are taken one after another in little “slices.” Then a computer algorithm stitches those plain films together to get the familiar CT pictures of your insides. A single CT scan exposes the body to as much radiation as 200-1000 plain x-ray films — the bigger the body part, the more net radiation is needed.

And the more radiation used, the more likely there is to be mischief. Every ionizing particle has a chance to knock into your DNA, causing damage that can lead to cancer. Now, your cells already have mechanisms in place to repair this sort of damage. But it doesn’t work 100% of the time, and the more radiation, the more damage, and the higher the chance that an important DNA change will slip through the cracks.

Your body, it should be said, is bathed in radiation every day. There are radioactive elements in the earth’s crust, and there are cosmic rays pelting your scalp even as you read this. Your body can repair the damage, almost 100% of the time. But if you add a lot of excess radiation — from occupational exposures for pilots or radiologists, or from medical diagnostic studies in a CT scanner, you increase your risks.

About 2% of the total 1.7 million cancers in the US are thought to have been caused by diagnostic radiation. Another way of looking at this: one case of extra cancer is caused by every 400 to 2,000 routine chest CT scans. That’s a broad estimate, and it illustrates how difficult it is to really pin down the risk of these studies.

The problem is that estimates of cancer rates are largely based on atomic bomb survivors, who were exposed to far more radiation and have experienced a large increased cancer rate. From them, researches have extrapolated backwards — smaller radiation doses leading to smaller, but real, increased cancer rates. But we don’t really know if this is a simple linear relationship. Do smaller radiation doses increase cancer rates proportionally, or higher or lower than relatively high doses? It is very difficult to know, because we’re talking about small exposures and small increases in population risk. But when you apply that risk to millions of people, you’ve got a significant quantity of cancers potentially triggered by well-intentioned medical testing.

As a pediatrician, I’m especially concerned. My patients have more years to live than adults, so more years to potentially develop cancer; also, their smaller bodies might provide less shielding from radiation. Because their cells are growing, their DNA may be more vulnerable to the damage caused by radiation as well.

So, what can patients and parents do to limit risk?

First, reduce studies, especially CT scans. Many are being done unnecessarily. Just a few minutes discussion may help doctors and patients understand that a CT scan does not need to be done.

Reduce doses. Newer CT technology relies on lower radiation doses. Special equipment can be used on children to further limit exposures.

When practical, choose imaging studies that do not use ionizing radiation. This includes MRI scans and ultrasounds—which do not involve any increased risk of cancer. However, these studies have their own limitations and may not always be practical or realistic substitutions.

Limit studies to the area of interest. If you have a lump in your wrist, you don’t need a CT scan of your entire arm; someone being evaluated for liver inflammation doesn’t need a CT scan that goes down to their pelvis. Also, keep in mind that some areas of the body seem inherently more at-risk for radiation damage—such as the thyroid and gonads. Those areas should be shielded or avoided if possible.

CT scanning is a life-saving, crucial medical tool. When used appropriately, it’s one of the most powerful ways we have of confirming diagnoses and evaluating potentially catastrophic problems. But they do have a downside. Doctors and patients need to discuss CT scans as they would any other medical treatment, their risks and benefits, options and alternatives, and whether they’re really needed, before the tests are done.

Roy Benaroch is a pediatrician who blogs at The Pediatric Insider. He is also the author of Solving Health and Behavioral Problems from Birth through Preschool: A Parent’s Guide and A Guide to Getting the Best Health Care for Your Child.