Background and Purpose: Poor reactive steps may lead to falls in people with Parkinson disease (PwPD). However, whether reactive steps can be improved in PwPD at risk for falls or whether step training reduces falls remains unclear. This study aimed to determine whether 2 weeks of reactive step training result in (1) immediate and retained improvements in stepping and (2) fewer prospective falls in PwPD at fall risk. Methods: Twenty-five PwPD (70.52 years ± 7.15; Hoehn & Yahr range 1-3) at risk for falls completed a multiple baseline, open-label, uncontrolled pre-/postintervention study. Stepping performance was assessed at 2 baseline assessments (B1 and B2) followed by a 2-week, 6-session training protocol. Stepping was assessed immediately (P1) and 2 months after training (P2). Primary outcomes were anteriorposterior margin of stability (MOS), step length, and step latency during backward stepping. Fall frequency was measured for 2 months before and after training. Results: MOS during backward steps was significantly larger (better) after training (P < 0.001, d = 0.83), and improvements were retained for 2 months (P = 0.04, d = 0.66). Step length was not statistically significant different after training (P = 0.13, d = 0.46) or at follow-up (P = 0.08, d = 0.62), although effect sizes were medium and large, respectively. Step latency improved after initial exposure (P = 0.01, d = 0.60) but not following training (P = 0.43, d = 0.35). Twelve participants experienced fewer falls after training than before (10 = no change, 5 = increase; P = 0.12). Greater improvements in MOS were related to fewer falls (P = 0.04).

No abstract