How to Distinguish between Statistically Significant Results and Clinically Relevant Results.
Bennett DA.
BACKGROUND: A practicing clinician will often be confronted with the results of a new clinical trial in their relevant field and will be faced with the dilemma of determining whether these results are clinically relevant to their own work. This chapter aims to describe the concepts of statistical significance in randomized clinical trials from a mainly classical statistical inference perspective. This chapter describes approaches to assess clinical significance and illustrates these approaches with examples from the contemporary neurological literature. RESULTS: There are several approaches that have been described in the research literature to assess the clinical significance including the minimal important clinical difference, the fragility index, Bayesian approaches, and a graphical approach. Unfortunately none of these methods have been widely used in the neurological research literature. Examples are provided to illustrate how these methods can be applied to the contemporary neurological literature in order to provide the clinician with some guidance on their use. CONCLUSIONS: How the trial is designed can affect the external validity of the results and subsequently the clinical relevance of a randomized clinical trial. Large-scale streamlined clinical trials with inclusion criteria that are not too restrictive can improve the generalizability of trial results. Even highly statistically significant treatment effects can be unreliable if they are based on a small number of events. The approaches described in this chapter should provide the practicing clinician with a starting point in order to determine whether the reported statistically significant results are indeed clinically relevant.