Predictors of improved balance performance in persons with Parkinson's disease following a training intervention: analysis of data from an effectiveness-implementation trial.

Objective: (1) To determine associated factors of improved balance performance after a 10-week HiBalance intervention period, and (2) to determine effects of the programme on modifiable factors found above, considering both groups.
Design: Pre-posttest substudy founded on the outcomes evaluation of an effectiveness-implementation trial.
Participants: Sixty-one participants were allocated the HiBalance training, while 56 were controls.
Intervention: Participants received a 10-week, two times weekly, progressive balance training, that is, HiBalance intervention, led by physical therapists. The intervention was group based and gradually incorporated dual-tasking over the training period. Participants also performed, unsupervised, a 1×/week home exercise programme.
Main Outcome: The Mini-Balance Evaluation Systems Test (Mini-BESTest) assessed balance performance, and those having improved by ⩾2 points were classified as positive responders. Balance confidence was the secondary outcome.
Results: Fifty-three (87%) participants completed the intervention and 32 (60%) improved their balance scores by ⩾2 points in the intervention group, with 11 (24%) in the control group. The multivariable logistic regression analysis revealed two independently associated factors of improved balance, which included balance confidence (odds ratio (OR) = 0.95; 95% confidence interval (CI) = 0.90-0.99) and attendance of ⩾80% of training sessions (OR = 10.10; 95% CI = 1.71-59.60). The final model demonstrated good fit and acceptable discrimination (area under the curve = 0.84). Secondary analysis revealed a fair relationship (Rho = 0.30; P  = 0.044) between improvements in balance confidence and balance performance in the intervention but not control group.
Conclusion: Two personal factors were significantly associated with a higher likelihood of improvement in clinically measured balance performance. The HiBalance intervention appears to benefit those with lower balance confidence.