Back

Early recovery during stroke rehabilitation predicts the functional outcome at discharge: a retrospective clinical study.

Anziano, M.; Cusinato, R.; Tzovara, A.; Spierer, L.; Chabwine, J.

2025-02-04 rehabilitation medicine and physical therapy
10.1101/2025.02.03.25321578 medRxiv
Show abstract

BackgroundModels predicting functional recovery upon stroke rehabilitation (SR) are not implemented in clinical practice due to their complexity and limited power in patients with moderate/severe deficits. ObjectivesTo obtain preliminary evidence supporting the predictive value of a single measure of early recovery on SR discharge functional status. MethodsWe tested simple predictive models based on the functional independence measure (FIM) changes during the first three weeks of SR (3wFIMgain) in 81 patients with moderate/severe impairment. ResultsPatients with poor or good 3wFIMgain featured similar admission impairment, rate of complications, presence of aphasia or dominant-sided hemisyndrome, while poor early recovery was associated with lesions to the left sensory-motor cortex, temporo-parietal junction, and superior longitudinal fasciculus. The 3wFIMgain effectively predicted the discharge FIM gain in patients with [≥] 6 weeks of SR stay. Combined with the admission FIM, the 3wFIMgain improved a predictive model for the discharge FIM. ConclusionsThe 3wFIMgain bears strong discharge prognostic value and synthetizes complex information about the early recovery of stroke patients that cannot be described by individual clinical predictors. Measures of early recovery might improve the power of models and simplify them, thus favoring application in clinical practice. Future validation in confirmatory studies is required.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.