Predictive factors for ambulation in 150 stroke patients using the multiple regression model

Objective: To investigate the predictive factors to the recover of the ambulation in stroke patients in rehabilitation setting. Methods: Side of hemorrhage, type of stroke, hypertension, diabetes, dyslipidemia were record from 150 (80 male, 70 female) patients. At admission and discharge we measured the Functional Independence Measure (FIM), motor FIM (motFIM), upper and lower Motricity Index (up-MI and low-MI) and Trunk Control Test (TCT). The outcome variable was the Functional Ambulation Classification (FAC), assessed at the discharge from the rehabilitation setting. The forward stepwise multiple regression was used with FAC as dependent variable. Results: The up- and low-MI, the TCT, the FIM and motFIM, and the age at entry were found to be significantly linearly related to the FAC at discharge, while side of hemorrhage, type of stroke, hypertension, diabetes, dyslipidemia and sex were not. The stepwise multiple regression analysis showed that the independent variables related with FAC were age, TCT and FIM, with a high coefficient of correlation (r=0.82). The patients with higher score of FAC had higher values of both TCT and FIM (Kruskal-Wallis, p<0.01) and a shorter time of hospitalization (Kuskal-Wallis, p=0.004). Discussion: We used the FAC as measure of outcome in walking because of its reliability and sensitivity in patients with stroke. Multiple linear regression showed that FAC was explained by TCT, FIM and age, with a high degree of correlation. Interestingly, FAC was not found to be related with side of hemorrhage, type of stroke and hypertension.