A novel nomogram to predict mortality in patients with stroke: a survival analysis based on the MIMIC-III clinical database

Background: Stroke is a disease characterized by sudden cerebral ischemia and is the second leading cause of death worldwide. We aimed to develop and validate a nomogram model to predict mortality in intensive care unit patients with stroke.

Methods: All data involved in this study were extracted from the Medical Information Mart for Intensive Care III database (MIMIC-III). The data were analyzed using multivariate Cox regression, and the performance of the novel nomogram, which assessed the patient's overall survival at 30, 180, and 360 days after stroke, was evaluated using Harrell's concordance index (C-index) and the area under the receiver operating characteristic curve. A calibration curve and decision curve were introduced to test the clinical value and effectiveness of our prediction model.

Results: A total of 767 patients with stroke were randomly divided into derivation (n = 536) and validation (n = 231) cohorts at a 7:3 ratio. Multivariate Cox regression showed that 12 independent predictors, including age, weight, ventilation, cardiac arrhythmia, metastatic cancer, explicit sepsis, Oxford Acute Severity of Illness Score or OASIS score, diastolic blood pressure, bicarbonate, chloride, red blood cell and white blood cell counts, played a significant role in the survival of individuals with stroke. The nomogram model was validated based on the C-indices, calibration plots, and decision curve analysis results.

Conclusions: The plotted nomogram accurately predicted stroke outcomes and, thus may contribute to clinical decision-making and treatment as well as consultation services for patients.