Reliability of stroke-related disability assessment at discharge from a stroke unit

The modified Rankin Scale (mRS) score at discharge from the stroke unit (SU) is often used as a surrogate for the 3-month outcome, but the reliability of such an approach remains uncertain. Our aim was to investigate how mRS assessment at SU discharge compares with early post-discharge and 3-month post-discharge evaluation, considering discharge destination.

We prospectively enrolled adult acute stroke patients discharged from a single tertiary SU with residual neurological deficit and mRS 1-4. Functional assessment was made by a single mRS-certified neurologist using the Rankin Focused Assessment (RFA) at SU discharge, and at 7-21 and 83-97 days post-discharge.

Of 116 enrolled patients, 109 completed the early post-discharge assessment and 104 the 3-month follow-up. The agreement between mRS at SU discharge and in the early follow-up was moderate overall (Krippendorff's alpha 0.71) and in patients discharged from the hospital (alpha 0.68) but poor in those transferred to the rehabilitation ward (RW) (alpha 0.41). Somewhat worse reliability was observed for the 3-month mRS (alpha 0.63, 0.53, and 0.26, respectively). Discharge assessments tended to overestimate dependency, particularly in RW-transferred patients (28.3% vs. 8.6% in discharges from the hospital).

The modified Rankin Scale assessment reliability performed by an experienced neurologist at SU discharge is moderate and significantly poorer in RW-transferred patients. However, on the population level, it seems satisfactory for patients discharged from the hospital, even as a surrogate for the 3-month post-discharge mRS.