Test–Retest Reliability of Postural Sway Measures Using a Portable Low‐Cost Force Plate in Healthy Adults

BACKGROUND AND PURPOSE: Center of pressure (CoP) metrics derived from force plates are widely used to quantify postural control, but laboratory-grade systems limit routine clinical and field implementation. Portable low-cost force plates could enable physiotherapists to monitor balance longitudinally, provided that their measurements are sufficiently reliable and clinically interpretable. This study addresses a gap in the literature by systematically comparing the test-retest reliability of CoP outcomes across bipedal and single-leg stance conditions using a portable low-cost force plate, providing device-specific and task-specific validation data relevant to clinical and field-based implementation. METHODS: This test-retest reliability study included 27 healthy young adults who completed two laboratory sessions 7-10 days apart. In each session, participants performed three 30-s trials of bipedal quiet stance and three 30-s trials of single-leg stance on each leg barefoot with standardized arm position and visual fixation. Mean values across repetitions were analyzed. Relative reliability was assessed using two-way mixed-effects ICC for absolute agreement (3,1), with 95% confidence intervals. Absolute reliability was quantified using typical error (TE) and coefficient of variation (CV). Paired t-tests evaluated systematic between-session differences. RESULTS: Across all outcomes, single-leg stance demonstrated substantially higher inter-session reliability compared to bipedal stance. Measures of total CoP displacement and velocity during single-leg stance exhibited good to excellent reliability (ICC = 0.85-0.90) and low absolute error (CV ≈ 8%), with direction-specific displacements showing similarly consistent results (ICC = 0.83-0.90; CV generally < 10%). In contrast, the CoP ellipse area and standard deviation measures were considerably less stable (ICC = 0.57-0.79) and displayed markedly higher variability (CV ≈ 16%-23%). The contrast was even more pronounced in the bipedal stance, where reliability was poor to moderate across all parameters (ICC = 0.39-0.65), with the ellipse area exhibiting excessive variability (CV > 60%). Notably, no systematic between-session differences were observed for most outcomes, further supporting the consistency of the measurements. DISCUSSION: values should be interpreted strictly as measurement-error thresholds and not as indicators of clinical responsiveness or meaningful clinical change because responsiveness was not evaluated.