Abstract
Purpose: Final global disability 3 months post-stroke is the cardinal endpoint in acute stroke clinical trials. The most similar variable available in administrative datasets is discharge destination at end of the acute hospitalization. We investigated the predictive value of discharge destination for final global disability.
Methods: In the public dataset of the two NINDS-TPA trials, we characterized discharge destination as a 4 level ordinal variable. Correlation coefficients and logistic models probed the relation with the modified Rankin Scale (mRS) of global disability at 3 months.
Results: Among the 624 ischemic stroke patients, discharge destination was home in 42.1% , rehabilitation in 33.0% , skilled nursing facility in 9.8% , and in hospital-death in 12.4% . A strong correlation was noted between hospital discharge destination and 3 month mRS, r = 0.71, P < 0.001. Length of stay showed a weaker correlation with 3 month mRS, r = 0.29, p < 0.0001. A multiple logistic regression model identified 4 categories of independent predictors of 3 month global disability outcome, with discharge destination as the strongest independent variable.
Conclusions: Discharge destination is a powerful predictor of final 3 month global disability outcomes and a valid outcome measure for use in local and national quality improvement programs.
