Overall, treatment barely differed, and three months later, both groups reported similar improvement in pain and functioning. Thirty-five percent in the group whose doctor didn't know the MRI results reported a positive outcome, compared with 41 percent of those whose doctors saw the MRI.
The reason may be because abnormal findings on the MRI and symptoms don't correlate much, Cohen said. Many middle-aged people who don't have back pain will have abnormal findings on the MRI.
While Cohen advocates against routine use of MRIs, he sees their value in certain cases. They should still be done, he said, when there are risk factors for serious problems. For example, an elderly woman who has fallen and may have a fracture of the vertebrae should have the MRI, he said.
Other research has found that doctors who have a financial interest in the imaging equipment are more likely than others to order the tests. Duke University Medical Center researchers looked at 500 lumbar spine MRI results ordered by two groups of doctors. One group had a financial interest in imaging equipment; the other did not.
They found 86 percent more negative scans -- indicating the test wasn't needed -- in the financial-interest group. They reported the findings at the 2011 Radiological Society of North America meeting.
Dr. Richard Deyo, a professor of family and internal medicine at Oregon Health and Science University in Portland and co-author of a commentary accompanying Cohen's study, said it's important to remember that "only a relatively small fraction of the people who get low back pain" will need the injection.
"It may be premature to suggest we stop doing MRIs for these patients," Deyo said. "Maybe the answer is to be more selective about who needs the injection." Better patient selection might naturally reduce the
All rights reserved