In:
Stroke, Ovid Technologies (Wolters Kluwer Health), Vol. 50, No. Suppl_1 ( 2019-02)
Kurzfassung:
Introduction: Prediction of outcome after stroke may help clinicians provide effective stroke management and plan long-term care for patients. Hypothesis: We aimed to develop and validate a risk score for predicting functional outcome available at hospital after ischemic stroke using linked big data. Methods: A total of 24,907 cases with acute ischemic stroke (within 7 days after onset) were enrolled, who were registered in the Clinical Research Center for Stroke (CRCS) registry from 2006 to 2014 and successfully linked the Health Insurance Review and Assessment Service (HIRA) database. Among them, 22,005 cases were derivation and internal validation group, and 2,902 cases were external validation group. We assessed the functional outcome using a modified Rankin Scale (mRS) at 3-month after onset ischemic stroke. We evaluated the predictive value of variables associated with the favorable 3-month outcome (mRS ≤ 2) and developed the score using the logistic regression coefficients. The prediction model performance was assessed by the area under the receiver-operating characteristic curve (AUC). Results: Proportion of cases with favorable outcome were 67.0% (n=14,748) in the derivation and internal validation group, and 68.4% (n=1,986) in the external validation group. Multivariable predictors of functional outcome at 3-month included age, gender, stroke severity, body mass index, stroke subtype, previous functional status, recanalization treatment, pre-stroke antiplatelet or anticoagulation therapy, fasting glucose level, history of stroke, hyperlipidemia, diabetes mellitus, hypertension, congestive heart failure, kidney disease on dialysis, cancer before the stroke. The AUC of the prediction model was 0.855 (0.850 - 0.860) in the derivation group and 0.854 (0.849 - 0.860) in the internal validation group for 3-month functional outcome. In addition, the AUC of the model was 0.860 (0.846 - 0.874) in the external validation group for outcome at 3-month. Conclusions: This prediction of functional outcome model is a valid clinical tool to predict functional outcome following ischemic stroke and reliable externally. This prediction model may assist to estimate functional outcome after stroke and to decide the care plan after stroke.
Materialart:
Online-Ressource
ISSN:
0039-2499
,
1524-4628
DOI:
10.1161/str.50.suppl_1.TP386
Sprache:
Englisch
Verlag:
Ovid Technologies (Wolters Kluwer Health)
Publikationsdatum:
2019
ZDB Id:
1467823-8
