Results of Back Pain Trials: Significant or Important?Sometimes researchers come to conclusions that may not really be supported by the data. Results may be overstated in a positive way. This problem is called over-reporting.
Staff from the Institute for Research in Extramural Medicine in the Netherlands took a second look at 43 randomized controlled trials (RCTs). These studies involved exercise for chronic low back pain (LBP). Almost half reported very positive results for exercise therapy.
In fact, only six of the studies really showed that exercise was statistically and clinically important. And that's the problem. Statistics can show one thing on paper. But the clinical results (how the patients are doing) are a separate issue.
The authors point to one concept that researchers should pay more attention to. And that's the minimally clinically important difference (MCID). Measures of results such as pain, function, health status, or disability are often used to show results.
But how much change in each of these factors before and after the treatment is considered significant? That's what needs to be addressed for future studies. The authors say it would be good if a common base of information and MCID were used in all studies. This would help advance LBP research.
Exercise has been touted as one of the few treatment methods that actually works. But the results of this review suggest that these studies have overestimated the treatment effects of exercise. A closer look at studies reporting on other interventions for chronic LBP (advice, behavioral therapy, medications) may have the same problem of over-reporting.
To avoid exaggerating results of studies, the authors offer other researchers seven suggestions. First, choose one main measure of results. Make it clinically relevant. Do a sample size calculation. This tells the researcher how many people have to be in the study to get results that are statistically significant. They also review important statistics to use and how to analyze them.
Finally, the authors suggest that studies should not report success with their treatment unless the results have clinical importance.
Maurits van Tulder, PhD, et al. Statistical Significance Versus Clinical Importance. In Spine. July 15, 2007. Vol. 32. No. 16. Pp. 1785-1790.
|*Disclaimer:* The information contained herein is compiled from a variety of sources. It may not be complete or timely. It does not cover all diseases, physical conditions, ailments or treatments. The information should NOT be used in place of visit with your healthcare provider, nor should you disregard the advice of your health care provider because of any information you read in this topic.|
|All content provided by eORTHOPOD® is a registered trademark of Medical Multimedia Group, L.L.C.. Content is the sole property of Medical Multimedia Group, LLC and used herein by permission.|