The Use of Soft Endpoints in Clinical Trials: The Search for Clinical Significance

Abstract

Measures of health-related quality of life and other “soft” endpoints have appeal to clinical trialists because of their direct relevance to the patient. Unfortunately, while one can define “statistical significance” precisely, what constitutes “clinical significance” remains elusive. A very small difference in a scale, while statistically significant, may have little relevance to the individual patient. Cardiologists have developed a number of soft endpoints, for example, the Killip Scale and the New York Heart Association Score, that define easily recognizable and distinguishable scenarios. Clinical trials in cardiology have used these scores as entry criteria or, occasionally, as primary endpoints and the field has been able to interpret results clinically. Many other fields rely on scales that lack verbal tags to clinical scenarios. The clinical community often cannot interpret data from such scales, even though the scales themselves are quite reliable, precise, and sensitive to change. Without meaningful approaches to defining clinical significance, such scales are unlikely to become acceptable in clinical trials except, perhaps, as exploratory endpoints. This paper discusses several approaches to defining clinical significance, such as attaching changes in scale to changes in risk as defined epidemiologically, matching levels of scale to objective levels of function, using expert panels and groups of patients to calibrate the scales, and adopting the increasingly popular metric “reliable change.”