Test–retest reliability of wrist joint position sense in healthy adults in a clinical setting

Abstract

Introduction Proprioceptive assessments of the wrist inform clinical decision making. In wrist rehabilitation, joint position sense has emerged as one way of assessing conscious proprioception with varying methods and minimal psychometric analysis reported. The purpose of this study was to standardise the wrist joint position sense test method for clinical use and to determine its test–retest reliability in a healthy population. Methods Four wrist positions (20° and 45° flexion, 20° and 45° extension) were measured twice in a random order, by a single rater, using a universal goniometer on the same day. The absolute error in degrees between each position and reposition was calculated. For relative reliability analysis, the intraclass correlation coefficient (3,1) was calculated. For absolute reliability the standard error of the measurement was calculated and Bland–Altman plots visually inspected. Results Fifty-five healthy volunteers (mean age 31.1 SD±10.25 years) were assessed. The mean absolute error, summarised for all positions for test and retest, was 3.98°. The intraclass correlation coefficients were poor to fair (0.07–0.47), and standard error of the measurement was 2° (rounded) for all positions. The limits of agreement were fairly narrow, and the Bland–Altman plots showed random distribution of errors for each position, therefore the measurement error was clinically acceptable. Conclusions The active wrist joint position sense test using goniometry demonstrated poor to fair test–retest reliability and acceptable measurement error in healthy volunteers. The wrist joint position sense angle of 20° flexion was the most reliable.