Prediction of Balance After Inpatient Rehabilitation in Stroke Subjects with Severe Balance Alterations at the Admission.

Background: Patients with severe motor alterations would be those on who the prediction of the expected motor response after inpatient rehabilitation programs is most required.

Objectives: To analyze if the balance progress measured by the Berg Balance Scale and the time of hospitalization could be independent predictors of the Berg Balance at the end of a post stroke rehabilitation program in patients with severe balance alteration at the admission. Secondly, to compare a Berg Balance prediction model at the time of discharge based on the Berg Balance at the time of admission (model 1) to a Berg Balance prediction model at the time of discharge based on Berg Balance progress and the time of hospitalization (model 2).

Methods: Subjects suffering a first subacute supratentorial stroke admitted for inpatient rehabilitation between 2010 through 2018 were included to develop two linear regression models of predicted Berg Balance at discharge (n=149).

Results: According to model 1 (p < 0.0001, R= 0.166), the Berg Balance at the admission would be a predictor of the Berg Balance at discharge from hospitalization. According to model 2 (p < 0.0001, R= 0.993) the Berg Balance progress (β= 1.026; p < 0.0001) and the hospitalization time (β=-0.006; p < 0.0001) would be independent predictors of the Berg Balance at discharge.

Conclusions: The motor response to the rehabilitation programs in subacute patients with severe motor alterations could be explained on the basis of balance condition at the admission, but this explanation may be improved considering the progress on the balance the patients achieve during inpatient rehabilitation irrespective the time of hospitalization.