OBJECTIVE

To determine whether the measurement properties of an instrument that combines items from the Berg Balance Scale (BBS) and the Functional Gait Assessment (FGA) called the supports measuring balance across the functional mobility spectrum

DESIGN

Retrospective cohort.

SETTING

Item-level data were from an archival research database.

PARTICIPANTS

Ambulatory individuals (N=93, BBS=50 [29-56], FGA=16 [0-30], Fugl-Meyer Assessment of Lower Extremities=27 [14-34], self-selected walking speed=0.4±0.2 m/s, mean age ± SD, 61.7±11.3y; 30.1% female) with chronic stroke (≥6 months).

INTERVENTIONS

Not applicable.

MAIN OUTCOME MEASURES

Unidimensionality was evaluated with a principal components analysis (PCA) of residuals. FBAM rating-scale characteristics, item hierarchy, item and person fit, and person separation were investigated using the Andrich Rating Scale Model.

RESULTS

PCA findings indicate the FBAM is sufficiently unidimensional. Rating scale structure was appropriate without modifying the original BBS and FGA scoring systems. Item hierarchy aligned with clinical and theoretical predictions (hardest item: FGA-gait with narrow base of support, easiest item: BBS-sitting unsupported). One item (BBS-standing on 1 foot) misfit, however, removal marginally affected person measures and model statistics. The FBAM demonstrated high person reliability (0.9) and 6 people (∼6%) misfit the expected response pattern. The FBAM separated participants into 4 statistically distinct strata, without a floor or ceiling effect.

CONCLUSIONS

The FBAM is a unidimensional measure for balance ability across a continuum of functional tasks. Rating-scale characteristics, item hierarchy, item and person fit, and person separation support the FBAM's measurement properties in persons with chronic stroke. Future work should investigate measurement with fewer items and whether the FBAM addresses barriers to adoption of standardized balance measures in clinical practice.