BACKGROUND AND PURPOSE: Reactive balance training (RBT) has shown promise for enhancing reactive balance control and reducing falls poststroke. However, the optimal training parameters (e.g., intensity, duration) are unknown. This secondary analysis examined dose-response relationships between specific reactive balance training characteristics and improvements in reactive balance control and fall rates poststroke. METHODS: 30 people with chronic stroke (mean age: 66.6 ± 8.8 years; time after stroke: 3.3 ± 2.3 years) completed up to 12 one-hour reactive balance training sessions, twice per week. Training included experiencing losses of balance due to internal or external perturbations while performing voluntary tasks. The tasks were of four types: stable, quasi-mobile, mobile, and unpredictable, each with choice of three difficulty levels (normal, increased, or reduced). Multiple linear regression (reactive balance outcome) and negative binomial regression (fall rates) were used to model associations between training characteristics (total number of perturbations, task difficulty, perceived challenge, and success rate) and post-training fall rates and changes in the reactive balance control sub-score of the mini-Balance Evaluation Systems Test (mini-BESTest). RESULTS: A higher number of perturbations were significantly associated with better post-intervention reactive balance scores on the mini-BEST (p = 0.010). There were no significant associations with any other training characteristics and post-intervention mini-BEST Scores. For fall incidence rates, there were no significant associations with any training characteristic (all p > 0.05). DISCUSSION: Greater exposure to RBT was associated with improvements in reactive balance control among individuals with chronic stroke. Participants who completed a higher number of perturbations within the standardized 12-session protocol showed greater improvements in reactive balance control, consistent with a dose-response association. These results indicate that training volume-reflected in the number of perturbations experienced-may be an important factor to consider when refining RBT protocols.
Faria et al. (Wed,) studied this question.