Is statistical significance clinically important?--A guide to judge the clinical relevance of study findings.

In this paper we describe several issues that influence the reporting of statistical significance in relation to clinical importance, since misinterpretation of p values is a common issue in orthopaedic literature. Orthopaedic research is tormented by the risks of false-positive (type I error) and false-negative (type II error) inferences, due to multiple testing and small sample sizes. Strict vigilance is required for interpretation of results and their accompanying p values to determine whether the results are of any clinical importance. To prevent type I and type II errors, primary and secondary outcome measures should be clearly defined and a sample size calculation should be performed on solely the primary outcome parameter. Analysis of multiple secondary outcome measures requires an adjusted significance level (e.g., Bonferroni correction). Prior to the sample size calculation, the minimal clinically important difference of the primary outcome measure has to be assessed in order to reveal a power that is required to achieve clinically important inferences. However, not only the treatment effect, but multiple factors such as the choice of the study population, follow-up duration, outcome measures, and the design of a study will eventually determine the clinical importance of the results.