Related Experiment Video
Updated: Mar 29, 2026
Establishing a Competing Risk Regression Nomogram Model for Survival Data
Published on: October 23, 2020
A Comparison of Mortality Prognostic Scores in Ischemic Stroke Patients
Jie Xu1, Yongli Tao2, Xuewei Xie1
1Department of Neurology, Beijing Tiantan Hospital, Capital Medical University, Beijing, China; China National Clinical Research Center for Neurological Diseases, Beijing, China; Center of Stroke, Beijing Institute for Brain Disorders, Beijing, China; Beijing Key Laboratory of Translational Medicine for Cerebrovascular Disease, Beijing, China.
Insights
Four stroke mortality scores accurately predict outcomes in Chinese patients. However, their predictive accuracy decreases for patients receiving thrombolysis therapy.
Area of Science:
- Neurology
- Public Health
- Biostatistics
Background:
- Stroke mortality prediction is crucial for patient management.
- Several new stroke mortality scores have been developed.
- Validation in diverse populations is essential.
Purpose of the Study:
- To compare the predictive accuracy of four recent stroke mortality scores in the Chinese population.
- To evaluate the Get With The Guidelines-Stroke risk score, iScore (30-day and 1-year), and PLAN score.
- To assess performance for in-hospital, 30-day, and 1-year mortality.
Main Methods:
- Analysis of 9698 hospitalized acute ischemic stroke patients from the China National Stroke Registry.
- Calculation of Area Under the Receiver Operating Characteristic Curve (AUC) for each score.
- Comparison of AUC values using the z-test.
Main Results:
- All four scores demonstrated predictive accuracy (AUC > 0.75) for short-term and long-term mortality.
- No significant differences were found among the AUC values of the four scores.
- A notable reduction in AUC values was observed for patients receiving thrombolysis therapy.
Conclusions:
- The four evaluated stroke mortality scores are valuable for predicting mortality in Chinese ischemic stroke patients.
- Caution is advised when applying these scores to patients who have received thrombolysis therapy.
- Further research may be needed to refine scores for specific treatment groups.
Objective:
Our objective is to compare the predictive accuracy of 4 recently established stroke mortality scores in the Chinese population, including the Get With The Guidelines-Stroke risk score, the iScore prediction rule for 30-day mortality, the iScore prediction rule for 1-year mortality, and the PLAN score.
Methods:
We analyzed data from 9698 hospitalized patients with acute ischemic stroke in the China National Stroke Registry (September 2007-August 2008). Outcome measures were in-hospital mortality, 30-day mortality, and 1-year mortality. To evaluate the performance of each model, we calculated the area under the receiver operating characteristic curve. The areas under the curve (AUCs) were compared using the z-test.
Results:
Among 9698 patients, the mean age was 65.5 years and 38.6% were female. Overall in-hospital mortality, 30-day mortality, and 1-year mortality were 3.0%, 5.9%, and 14.6%, respectively. The AUC values of 4 scores for each outcome were all more than .75. The z-test did not show significant difference among the AUC values of these 4 scores. For patients who received thrombolysis therapy, these 4 scores had apparent reductions in the AUC values.
Conclusions:
We could confirm the predictive value of all these 4 scores for short-term and long-term mortalities in Chinese IS patients. However, these 4 scores need to be interpreted with caution in the patients who received thrombolysis therapy.
