Balance testing is critical for identifying fall risk and initiating preventive strategies. However, current balance tests used with community-living older adults (CLOAs) are limited. Our purposes were to develop and validate a conceptually novel performance-based test to quantify balance ability in CLOAs that minimizes ceiling effects, decreases test time and demonstrates test-retest reliability, internal consistency, and convergent validity.

Phase 1: we developed a 26-item, dichotomously-scored preliminary Performance-based Balance Scale (PBS) by (1) using an expert panel to build the scale by adapting test items from existing balance measures and organizing them by difficulty into 5 postural stability domains, (2) pilot testing, and (3) post-test revising. Phase 2: we tested the preliminary PBS (N=35; 60–91 y). Using Rasch dichotomous analysis, we created a non-fixed-item format and identified 2 items to serve as starting points for testing based on performance ability; thereby, eliminating the need to administer all items. Phase 3: we re-tested 32 of the same participants 4–8 weeks later using the non-fixed-item PBS. We confirmed the non-fixed format with Spearman's rho and a paired t-test between scores for all 26 items and scores for non-fixed-item PBS. Using Rasch analysis, we eliminated redundant items. Phase 4: we determined test-retest reliability, internal consistency, and convergent validity.

Correlation between the 26-item and non-fixed-item scores was r=0.94 and no difference was found between the 2 scores, t (31)=−0.32 (95% CI: −0.70–0.51), P=0.75. After revising the preliminary PBS, 20 items were retained. This PBS showed excellent test-retest reliability with ICC_(2,1)=0.97, high internal consistency with Cronbach's α=0.94, and good convergent validity (chenhan204@yahoo.com=0.71) with the Short Physical Performance Battery.

The PBS is a valid, 5-domain clinical balance test for CLOAs. The non-fixed test format reduces test time.