For decades, doctors have been recommending that children with scoliosis wear a back brace so that the sideways curve of the spine doesn't get worse.

But there was scant evidence to prove back braces really help.

And telling a teenage girl she's going to have to wear a brace 18 hours a day for 7 1/2 years, as pro golfer Stacy Lewis did, can be a tough sell for parent and child alike.

To avoid the brace, doctors have tried lots of other things — exercise, avoiding exercise, traction and electrical stimulation — but none of them have seemed to help.

The brace prevailed as the best way to avoid surgery. But were the doctors right?

To find out if a medical treatment works, researchers turn to randomized controlled trials to test one approach against an alternative, usually a sham treatment or placebo. But there's no easy placebo for a back brace, and researchers figured that many parents wouldn't want to leave their child's treatment choice up to chance.

This is a child's spine we're talking about. If not treated effectively, scoliosis can lead to permanent deformity, disc damage and neurological damage.

So researchers looking to answer the question about braces' effectiveness had a tough challenge. They screened 1,183 patients at 25 sites in the United States and Canada to find 116 families willing to participate in a randomized trial. Ultimately, the scientists enrolled another 126 children by letting their families decide whether to brace or not.

Then there's the matter of getting a teenager, most often a girl, to wear a back brace 18 hours a day.

The researchers put temperature sensors in the braces so they could tell if they were actually being worn. Almost one-third of the patients quit wearing the brace altogether, and only a few wore it for the recommended 18 hours a day.

The results were just published in the New England Journal of Medicine.

It turned out that the more hours a day a child wore a brace, the less the spine curvature got worse; 13 hours a day or more worked best.

Altogether, about three-quarters of the children who wore a brace avoided a spine curvature of 50 degrees or more, compared with about half those in the control group.

But that means half of the children who got no treatment turned out fine, as did 41 percent of the patients who didn't wear the brace much.

That, the authors conclude, may mean that the criteria for bracing may be too broad, "resulting in unnecessary treatment for many patients."

Or, as an accompanying editorial put it, finding out that so many children with scoliosis did well without bracing is important, too. "The challenge for the field going forward is to identify children who are most likely to benefit from bracing and those who are unlikely to benefit," the editorial said.

Stacy Lewis ended up having back surgery for scoliosis despite those long years in a brace. Her surgeons modified the procedure to reduce the side effects that could hurt the young golfer's prospects. They seemed to have made the right choice. Lewis is now ranked third in the world for women's golf.

Fortunately, most children will never have to face the brace. Only about 3 percent of children are diagnosed with scoliosis, and less than 1 percent end up needing treatment.

Copyright 2021 NPR. To see more, visit https://www.npr.org.