Defining the Age of Young Ischemic Stroke Using Data-Driven Approaches.

Area of Science:

• Neurology
• Data Science
• Epidemiology

Background:

• The definition of young ischemic stroke (IS) age is inconsistent across studies.
• Establishing a precise age cut-point is crucial for clinical and epidemiological research.

Purpose of the Study:

• To leverage electronic health records (EHRs) and data science to determine an optimal age cut-point for defining young ischemic stroke (IS).

Main Methods:

• Two data science approaches were applied to patient-level EHRs from 13 Pennsylvania hospitals.
• Approach 1: Grouped comorbidities using ICD-9/10 codes and calculated patient similarity scores to identify age trends.
• Approach 2: Utilized machine learning models (GLM, RF, XGBoost) to classify patients across 17 age groups, analyzing feature importance for comorbidity patterns.

Main Results:

• The study analyzed 7,555 IS cases and 31,067 controls.
• Approach 1 identified optimal ages of 53.7 years for females and 51.0 years for males.
• Approach 2 yielded optimal ages of 53 (GLM), 52 (RF), and 54 (XGB) for females, and 52 (GLM/RF) and 53 (RF) for males, highlighting sex-based differences in comorbidity patterns.

Conclusions:

• Data-driven analysis suggests an optimal age of 54 for women and 52 for men to define young stroke in a central Pennsylvania population.
• Further validation with diverse populations is recommended.