Predictive models for delay in medical decision-making among older patients with acute ischemic stroke: a comparative...

Area of Science:

• Gerontology
• Neurology
• Data Science

Background:

• Delayed medical decision-making in older adults with acute ischemic stroke (AIS) poses significant health risks.
• Understanding the factors contributing to these delays is crucial for timely intervention.

Purpose of the Study:

• To identify factors influencing delayed medical decision-making in older AIS patients.
• To compare the predictive performance of logistic regression and the Light Gradient Boosting Machine (LightGBM) algorithm for these delays.

Main Methods:

• A cross-sectional study involving 309 older patients (≥60 years) with AIS.
• Data collected included demographics, stroke characteristics, stroke knowledge, health literacy, and social network.
• Logistic regression and LightGBM models were developed and compared using Accuracy, Recall, F1 Score, AUC, and Precision.

Main Results:

• A high delay rate of 74.76% was observed in older AIS patients.
• Key influencing factors identified by both models included stroke severity, stroke recognition, previous stroke knowledge, health literacy, and social network.
• The LightGBM model demonstrated superior performance with an AUC of 0.909.

Conclusions:

• The LightGBM algorithm is optimal for early identification of delayed medical decision-making in older AIS patients.
• Identified factors provide insights for developing targeted prevention and intervention strategies.
• Reducing decision-making delays can promote better health outcomes for older AIS patients.