Abstract
Background: There are multiple issues that arise when researchers focus on and only report "statistical significance" of study findings. An important element that is often not included in reports is a discussion of clinical relevance.
Objectives: The authors address issues related to significance, the use of effect sizes, confidence or credible intervals, and the inclusion of clinical relevance in reports of research findings.
Methods: Measures of magnitude, precision, and relevance such as effect sizes, confidence intervals (CIs), and clinically relevant effects are described in detail. In addition, recommendations for reporting and evaluating effect sizes and CIs are included. Example scenarios are presented to illustrate the interplay of statistical significance and clinical relevance.
Results: There are several issues that may arise when significance is the focus of clinical research reporting. One issue is the lack of attention to nonsignificant findings in published works although findings show clinical relevance. Another issue is that significance is interpreted as clinical relevance. As well, clinically relevant results from small-sample studies are often not considered for publication, and thus, findings might not be available for meta-analysis.
Discussion: Findings in research reports should address effect sizes and clinical relevance and significance. Failure to publish clinically relevant effects and CIs may preclude the inclusion of clinically relevant studies in systematic reviews and meta-analyses, thereby limiting the advancement of evidence-based practice. Several accessible resources for researchers to generate, report, and evaluate measures of magnitude, precision, and relevance are included in this article.