Guideline-Guided Prognostic Models of Mortality Following First-Ever Ischemic Stroke

This study aimed to develop and validate guideline-guided prognostic models for predicting mortality in patients with first-ever ischemic stroke. The study included 899 adult patients (≥ 18 years) with confirmed diagnosis of first-ever ischemic stroke from the National Stroke Registry (NSR) of Malaysia (2009–2020). The primary outcome assessed was post stroke mortality within 90 days. Multivariable regression was used to develop (75%, n = 674) and internally validated (25%, n = 225). Model performance was evaluated using discrimination [area under the receiver operating characteristic curve (AUROC) and calibration [Hosmer-Lemeshow test (HLT)]. The final prognostic model included age ≥ 60 years [adjusted odds ratio, aOR 2.39, 95% confidence interval, CI (1.37–4.18)], Glasgow Coma Scale (GCS) score of ≤ 8 [2.66, (1.31–5.40)], diabetes [2.42 (1.41–4.16)], and non-adherence to antiplatelet therapy within 48 hours [2.30 (1.26–4.20)], dysphagia screening [2.90 (1.63–5.17)], antiplatelet therapy at discharge [3.63 (2.07–6.36)], lipid-lowering therapy [2.09 (1.10-4.00)], stroke education [39.61 (21.92–71.57)] and rehabilitation [10.75 (6.00-19.25)]. The prognostic model demonstrated good validation performance, with an AUROC of 0.94 (HL p = 0.630). The study demonstrates that the guideline-guided prognostic models developed are effective in predicting mortality after a first-ever ischemic stroke. The model holds considerable promise for enhancing clinical decision-making and patient care. Additionally, risk scores generated from these models could be valuable for creating risk calculators, aiding healthcare providers, patients, and other stakeholders in making well-informed clinical decisions in managing patients with stroke.