Year: 2001
Type: article
Abstract: The recent editorial by Bonnick (1) argues that “regression to the mean” does not apply to individual patients. She argues that physicians should monitor bone density in treated patients, realizing that only large changes that make precision error unlikely should be considered meaningful changes. In the Fracture Intervention Trial of Alendronate and the MORE Trial of Raloxifene, we studied women whose bone density values seemed to decline more than 4% during the first year of treatment (2). Given the 2% precision of the measurements, these changes in these women would have been classified as “significant” because they were beyond the 95% confidence interval for precision error, also called the “least significant change” (3). Of these women, 92% had an apparent gain in BMD during the following year while receiving identical treatment. If a clinician had seen any one of these women after the first year of treatment, she or he may have concluded that they were “not responding” to treatment and may have considered changing treatment. But these individual women each had a very good prognosis if they made no change in treatment: the vast majority would go on to have substantial apparent improvements in bone density during the next year. Regardless of the name one applies to this phenomenon, we think there is an important lesson for the care of individual patients: if a patient seems to lose bone density during the first year of monitoring, it may be unwise to stop or change treatment on this basis alone. The loss is probably due to random error in this otherwise excellent and precise measurement, not a biological failure to respond in that individual. This indicates that the collective results of a well-conducted trial are a more accurate guide to the treatment of an individual patient than the serial results in that patient.
Authors: Steven R. Cummings, A. Michael Parfitt
Language: English
Subfield: Orthopedics and Sports Medicine
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W2125546474