New model may help predict patient outcome after stroke rehab

The model has potential application in rehabilitation research and stroke management, according to the study authors, and a calculator for it is available free of charge online.


A recently developed model may help predict whether rehabilitation will significantly improve stroke patients' physical function and independence.

Researchers in Italy derived two models from 717 patients, one for the study's primary outcome and one for the secondary outcome. All patients had been admitted to the hospital for rehab after stroke. The primary outcome was achievement of a motor Functional Independence Measure score of more than 61 points at discharge, indicating mild stroke impairment. The secondary outcome was recovery to a physical independence grade of 5 or above in the Functional Independence System, which indicates that a patient needs no more than supervision to groom, dress the upper body, or transfer from a chair to a bed or transfer onto a toilet; needs no more than minimal assistance to bathe, dress the lower body, or use a wheelchair or walk; functions at least at modified independence in eating and in bowel and bladder movements; and may need total assistance to climb stairs.

Model 1 included age, time from stroke to rehab admission, motor and cognitive Functional Independence Measure scores at admission, and unilateral neglect. Model 2 included age, male sex, time since stroke onset, and motor and cognitive Functional Independence Measure scores at admission. Each model was prospectively validated after derivation in 875 patients. The study results were published online by Stroke on Oct. 19.

Mean patient age was 72 years in the validation cohort and 70 years in the derivation cohort. Patients in the validation cohort were more likely to have been admitted earlier to inpatient rehab units (mean time from stroke to admission, 15 days vs. 25 days). Overall, 28.7% of patients achieved the primary outcome in the derivation cohort and 31.3% achieved it in the validation cohort, while the secondary outcome was achieved by 13.9% and 18.3%, respectively. Excellent discrimination was seen in both models. The area under the curve was 0.883 and 0.913 for model 1 and model 2, respectively, in the derivation cohort. In the validation cohort, the area under the curve for each model was 0.866 and 0.850, respectively. An independent association was seen between better four-year survival and both improved physical functioning (hazard ratio, 0.43; 95% CI, 0.25 to 0.71; P=0.001) and level of independence needing only supervision (hazard ratio, 0.32; 95% CI, 0.14 to 0.68; P=0.004).

The authors noted that their study was retrospective, that unmeasured or undocumented factors may have affected the outcome, and that median time from stroke to rehab admission was longer than in the U.S. for comparable populations, among other limitations. However, they concluded that the model they developed provides “an easy-to-use, accurate, and validated predictive tool for potential application in rehabilitation research and stroke management.” A calculator is available to download free of charge online.