Statistical Primer for Athletic Trainers: The Essentials of Understanding Measures of Reliability and Minimal Important Change

Abstract

Objective: To describe the concepts of measurement reliability and minimal important change.

Background: All measurements have some magnitude of error. Because clinical practice involves measurement, clinicians need to understand measurement reliability. The reliability of an instrument is integral in determining if a change in patient status is meaningful.

Description: Measurement reliability is the extent to which a test result is consistent and free of error. Three perspectives of reliability-relative reliability, systematic bias, and absolute reliability-are often reported. However, absolute reliability statistics, such as the minimal detectable difference, are most relevant to clinicians because they provide an expected error estimate. The minimal important difference is the smallest change in a treatment outcome that the patient would identify as important.

Recommendations: Clinicians should use absolute reliability characteristics, preferably the minimal detectable difference, to determine the extent of error around a patient's measurement. The minimal detectable difference, coupled with an appropriately estimated minimal important difference, can assist the practitioner in identifying clinically meaningful changes in patients.

Keywords: minimal detectable difference; outcomes; reporting statistical findings.

Figure.

Scatterplots for 2 scenarios of relative reliability between 2 testers. A, Good relative reliability (intraclass correlation coefficient = 0.95). B, Poor relative reliability (intraclass correlation coefficient = 0.36).