AI in predicting COPD in the Canadian population

Area of Science:

• Health Informatics
• Computational Medicine
• Pulmonology

Background:

• Chronic Obstructive Pulmonary Disease (COPD) is a progressive lung disease affecting 10% of Canadians over 35.
• Primary care settings are crucial for managing chronic diseases like COPD and maintaining patient health.
• Accurate identification of COPD is vital for timely intervention and effective disease management.

Purpose of the Study:

• To develop and evaluate machine learning models for identifying COPD patients within Canadian primary care.
• To leverage comprehensive Electronic Medical Record (EMR) data for COPD prediction.
• To identify key clinical features indicative of COPD from EMR data.

Main Methods:

• Utilized structured and unstructured EMR data from primary care clinics across seven Canadian provinces.
• Applied two supervised machine learning models: Multilayer Neural Networks (MLNN) and Extreme Gradient Boosting (XGB).
• Evaluated model performance using accuracy metrics on a test dataset.

Main Results:

• The Extreme Gradient Boosting (XGB) model achieved 86% accuracy in identifying COPD patients.
• The Multilayer Neural Networks (MLNN) model achieved 83% accuracy.
• Key predictive features included patient age, medications, health conditions, and risk factors.

Conclusions:

• Supervised machine learning, particularly XGB, can effectively identify COPD in primary care using EMR data.
• This approach aids in disease surveillance and improves evidence-based care delivery for COPD patients.
• The identified key symptoms provide valuable insights for COPD diagnosis in primary care settings.