Quantifying clinically significant change: a brief review of methods and presentation of a hybrid approach.

Overview

Treatment outcome researchers in orthopaedics frequently report only tests of statistical significance between group means to evaluate the effectiveness of a given intervention. Although important in establishing that mean differences are not caused by chance, these methods do not reflect the extent to which an intervention produces improvements that are meaningful and represent a return to health. This is an issue that is often of great interest to patients and clinicians. Other methods use a percentage change in an outcome measure (eg, 25% reduction in pain score) to classify treatment responders but often do not indicate whether the treatment restored a patient to normal. Researchers have developed several indices that provide a metric for statistically defining the amount of change that patients consider to be important. In this article, we focus on the concept of "clinical significance" and the different methods that have been developed to define clinically significant change using statistics. We then present a hybrid method that can classify whether a patient has returned to normal function. We apply this method to real patient data to illustrate its use with different outcome instruments commonly used in orthopaedic sports medicine. We advocate that the addition of these methods to reports from clinical outcome studies can deepen our understanding of the impact of interventions on patients' lives.