Assessing the Validity and Reliability of a Single Lumbar-Mounted IMU System for Gait Analysis

Abstract

Wearable inertial sensors offer a practical alternative to gold-standard optoelectronic systems for gait analysis, though their validity remains uncertain due to sensor placement. This study examined the intra-rater and inter-rater reliability and the concurrent validity of a single lumbar-mounted inertial measurement unit (Baiobit, BTS Bioengineering, Garbagnate Milanese, Italy) compared with a 3D optoelectronic system (SMART-DX 6000, BTS Bioengineering). Thirty healthy adults walked along an 8 m walkway at a self-selected speed, and seven spatiotemporal gait parameters (cadence, velocity, stride length, step length, stance, swing, and single-support phases) were computed by both systems. Reliability and validity were assessed using intraclass correlation coefficients (ICC), standard error of measurement, minimum detectable change, paired tests, Spearman correlation, Passing–Bablok regression, and Bland–Altman analysis. Baiobit showed intra- and inter-rater ICCs of 0.53–0.90 and 0.66–0.88, respectively. Bland–Altman results indicated non-significant biases for global parameters (velocity: −0.06 m/s; cadence: 1.11 steps/min), whereas spatial measures showed significant biases (stride length: 0.11 m; step length: 0.06 m). Gait phase parameters demonstrated low correlations (r = 0.08–0.11) and proportional systematic errors. Overall, the Baiobit system provided reliable and valid estimates of global spatiotemporal parameters but lacked precision for gait phase metrics, underscoring limitations that currently prevent it from fully replacing optoelectronic systems.