Construct Validity and Test–Retest Reliability of the Al Balance Stabilometric System in Healthy Young Adults

Abstract

Each instrumented balance platform requires its own measurement-property evidence. We examined the test–retest reliability of the Al Balance Stabilometric System and its hypothesis-based construct validity relative to the Biodex Balance System SD. Forty healthy adults aged 18–25 years completed four Al Balance modules: four-condition sensory assessment (mCTSIB), Limits of Stability (LoS), unilateral stance (Bilateral Comparison Test), and Postural Stability Test. The matched sensory conditions, unilateral stance, and LoS were also performed on Biodex in randomized device order; Al Balance was repeated after 7 days. Validity was assessed with repeated-measures ANOVAs, repeated-measures correlations (rrm), within-participant profiles and standardized interaction models; reliability was assessed with intraclass correlation coefficients (ICC), standard error of measurement and minimal detectable change. Both systems captured the expected visual and surface effects, and cross-device convergence, rather than numerical agreement, was supported for the sensory conditions (rrm = 0.513–0.598). However, individual-level agreement was weak: between-device correlations of participant-level sensory response magnitudes were non-significant after Holm correction, and individual asymmetry scores in unilateral stance converged weakly. Device × condition interactions indicated differing relative sensitivity. Test–retest reliability was outcome-dependent, moderate to excellent (ICC(A,1) = 0.678–0.946). In healthy young adults, Al Balance provided repeatable scores that are device-specific rather than interchangeable with Biodex.