The Mini-BESTest is an independent predictor of falls in Parkinson Disease

Background: Falls in Parkinson Disease (PD) are a complex health problem, with multidimensional causes and consequences. Objectives: To identify the fall predictors in individuals with PD and compare falters and nonfalters considering their socio-demographic, anthropometric, clinical and functional status. Methods: A multicenter cross-sectional design was employed. Variables included: age, sex, body mass index, PD progression, levodopa dosage, activities limitation and motor impairments (UPDRS ADL/Motor), level of physical activity (human activity profile - HAP), fear of falls (Falls Efficacy Scale-International-FES-I), freezing of gait (Freezing of Gait Questionnaire - FOG-Q), gait speed (10 meters walk test - 10-MWT), lower limb functional strength (Five Times Sit-toStand Test - FTSST), balance (Mini-BESTest), mobility (Timed Up & Go - TUG) and dual-task dynamic (TUG-DT). Seventeen potential predictors were identified. Logistic regression and ROC curve were applied. Results: Three-hundred and seventy individuals (44.87% falters and 55.13% non-falters) completed the study. Falters presented worse performance in UPDRS motor/ADL/Total, FES-I, FOG-Q, Mini-BESTest, HAP, TUG and TUG-DT and the majority were inactive. The Mini-BESTest Total was the main independent predictor of falls (OR =0.92; p < 0.001; 95% CI = 0.89, 0.95). For each one-unit increase in the Mini-BESTest, there was an average reduction of 8% in the probability of being a faller. A cut-off point of 21.5/28 (AUC = 0.669, sensitivity 70.7% and specificity 55.1%) was determined. Conclusion: Besides characterizing and comparing falters and non-falters, this study showed that the Mini-BESTest was the strongest individual predictor of falls in individuals with PD, highlighting the importance of evaluating dynamic balance ability during fall risk assessment. (C) 2019 Associacao Brasileira de Pesquisa e Pos-Graduacao em Fisioterapia. Published by Elsevier Editora Ltda. All rights reserved.