Which Patients with COPD Would Benefit from Cough Monitoring?

Area of Science:

• Pulmonary Medicine
• Biomedical Engineering
• Data Science

Background:

• Effective exacerbation detection in chronic obstructive pulmonary disease (COPD) relies on identifying patients with increased cough during exacerbations.
• Cough monitoring requires careful patient selection for optimal performance.

Purpose of the Study:

• To determine if patient data can predict increased cough during exacerbations in COPD patients.
• To assess the impact of patient stratification on the performance of a cough-based exacerbation alert system.

Main Methods:

• Post hoc analysis of data from a validation study of a nighttime cough-based alert system.
• Investigated correlations between patient data and cough statistics (median, interquartile range).
• Developed a logistic regression classifier using patient age and symptom questionnaire data to predict increased cough during exacerbation.

Main Results:

• Patient data showed no statistically relevant correlation with nighttime cough count median or interquartile range.
• Patients with and without increased cough during exacerbation exhibited distinct characteristics.
• A classifier achieved 85% accuracy in predicting increased cough during exacerbation.
• Patient stratification using the classifier improved alert system sensitivity from 59% to 76%.

Conclusions:

• Patient data can be utilized to stratify COPD patients for targeted cough monitoring.
• Stratification enhances the performance of exacerbation alert systems, leading to improved sensitivity.