Poster 236: Predictive Factors for Ambulation in Stroke Patients Using a Multivariate Analysis

Objective: To investigate the predictive factors for the recovery of ambulation in stroke patients in the rehabilitation setting. Design: A prospective repeated-measures design. Setting: Rehabilitation hospital. Participants:150 hemiplegic patients at their first stroke. Intervention: Intensive rehabilitative training. Main Outcome Measures: FIM instrument, FIM motor score, upper and lower Motricity Index, Trunk Control Test (TCT), and Functional Ambulation Classification (FAC). Results: The mean age was 68.9±11.52 years. The mean time between stroke and start of rehabilitation was 25.3±16.9 days (median, 20), while the mean time of hospital care was 52.3±29.4 days (median, 45). Upper and lower Motricity Index, TCT, FIM, and FIM motor, and age at admission were significantly linearly related to FAC at discharge, while side of hemorrhage, type of stroke (hemorrhagic or ischemic), hypertension, diabetes, dyslipidemia, and sex were not. Logistic regression analysis showed that the independent variables related to FAC were age, TCT and total FIM: the model correctly allocated 124 of 150 cases (83%). Considering an output value below .05 on logistic regression to be a risk factor of FAC=0, we found that 21 of 35 patients with FAC=0 had an output value below .05 and that only 3 patients with an output below .05 had FAC>0. Conclusions: We used FAC as a measure of outcome in walking because of its reliability and sensitivity in patients with stroke. Logistic regression showed that dichotomized FAC was explained by TCT, total FIM, and age, with an 83% correct allocation rate. Focusing on the group with FAC=0 (35 patients), output below .05 on logistic analysis was a risk factor with an odds ratio of 41.6 (95% CI, 12.5–139; sensitivity, 60%; specificity, 96.5%). These findings seem to be interesting for optimizing rehabilitation goals.