Discriminant validity and test–retest reliability of a digital game-based rehabilitation platform to assess balance and cognition in elderly Indians with diabetes: a cross-sectional study

Abstract

Ageing impairs balance and cognitive function, exacerbated by type 2 diabetes mellitus (T2DM), thereby increasing fall risk. Dual-task assessments combining motor and cognitive demands via feasible clinical tools are needed. This cross-sectional study assessed the test–retest reliability and discriminant validity of a computer game-based rehabilitation protocol (CGRP) incorporating balance and cognitive tasks in older adults with T2DM. Thirty older adults with T2DM (mean age: 68.83 ± 2.90 years) underwent CGRP evaluations of centre of pressure (COP) excursions under 10 conditions, repeated after one week. For discriminant validity, an additional 30 participants (15 with T2DM and 15 age-matched non-diabetics) were recruited. Reliability was analysed via intraclass correlation coefficients (ICC), standard error of measurement, and minimal detectable change. Discriminant validity was examined through correlations with traditional balance scales consisting of the Berg balance scale (BBS), dynamic gait index (DGI), activities-specific balance confidence (ABC) and group comparisons using baseline scores. The CGRP showed excellent test–retest reliability (ICC = 0.82–0.98; 95% confidence intervals: 0.69–0.99), with no significant systematic errors (p > 0.05). COP excursions increased under DT (dual task) conditions, particularly on sponge surfaces. Weak correlations (Spearman’s rho: −0.372 to 0.455; p > 0.05) with BBS, DGI, and ABC supported discriminant validity (weaker in T2DM). Significant group differences were observed in ABC scores (p = 0.011) and sponge surface eyes-closed COP excursions (p = 0.037). The CGRP is a reliable and valid tool for assessing DT balance and cognition in older adults with T2DM.