Machine learning-driven health profiling and multidimensional trajectory analysis in first-ever ischaemic stroke:...

Area of Science:

• Neurology
• Public Health
• Data Science

Background:

• Ischemic stroke is the most common stroke type, causing significant long-term health burdens.
• Patients with first-ever ischemic stroke show considerable variability in post-stroke health and clinical presentations.
• Understanding this heterogeneity is crucial for improving patient outcomes.

Purpose of the Study:

• To identify distinct patient clusters with heterogeneous multidimensional health profiles using the k-prototype clustering algorithm.
• To characterize synergistic trajectories of core health attributes within the largest cluster through parallel process latent class growth modeling.
• To elucidate static heterogeneity and dynamic recovery patterns following a first-ever ischemic stroke.

Main Methods:

• A multicentre, cross-sectional, and longitudinal study design.
• Utilizing patient self-reported outcomes and objective measurements for comprehensive health evaluation.
• Employing k-prototype clustering and parallel process latent class growth modeling for data analysis.

Main Results:

• The study aims to identify distinct patient clusters based on demographic, clinical, and behavioral factors.
• It will characterize shared longitudinal trajectories in the development of adverse health outcomes.
• Analysis will combine cross-sectional and longitudinal data to reveal static and dynamic recovery patterns.

Conclusions:

• This research will delineate heterogeneous patient clusters following ischemic stroke.
• It will elucidate shared longitudinal recovery trajectories.
• Findings are expected to improve understanding and management of post-stroke health heterogeneity.