Clinical versus statistical significance as they relate to the efficacy of periodontal therapy

Abstract

The author discusses the shortcomings of using statistical significance testing as a method to infer that results of periodontal clinical trials are clinically meaningful. To compensate for these deficiencies, he also identifies criteria and periodontal parameters that can be used to reflect clinically significant outcomes. The author searched the medical and dental literature to identify commentaries that addressed the problems associated with interpreting statistical significance testing, or hypothesis testing, and defining clinical significance. The limitations of statistical significance testing related to identifying clinically significant changes include failure to indicate if the detected differences between variables in test and control groups are large or important. After reviewing various definitions of the term "clinical significance," the author reviews and proposes a comprehensive working definition of it. Regarding the efficacy of periodontal therapy, he delineates the advantages and limitations of specific criteria (such as absolute values, cut points) that can be used to define clinical significance. The author suggests that clinically significant results should be defined before initiating a study and statistical significance testing should be used to validate that findings did not occur by chance. This would help place the importance of clinical data into perspective, and it would enhance clinicians' ability to select the most appropriate therapies for particular sites in periodontal patients.