Back

Using Hourly Aggregated Respiratory Rate and Expiratory Time with Machine Learning to Identify Remote COPD Exacerbations

Smunyahirun, R.; Filhol, T.; Trakoolwilaiwan, T.; Al-Ani, R. I.; Lynn, M.; Liang, L.; Balasundaram, G.; Lombay, J.; Singh, G.; Wiens, M. O.; Moschovis, P. P.; Olivo, M.

2025-10-25 respiratory medicine
10.1101/2025.10.24.25338693 medRxiv
Show abstract

Exacerbations of chronic obstructive pulmonary disease (COPD) are a major cause of morbidity and mortality. Various models for identifying exacerbations have been proposed, but few models are based on continuous and passively recorded respiratory physiology data. We enrolled 17 subjects with COPD with at least 2 exacerbations in the prior 12 months. We developed machine learning models using aggregated respiratory rates and expiratory times collected passively in a home-setting to identify COPD exacerbations. Models achieved areas under the ROC curve of 0.7989 and 0.8720 respectively, for exacerbations defined by self-reported outcomes and COPD Assessment Test (CAT) score changes. This pilot study demonstrates the feasibility of using passively recorded respiratory physiological data to identify COPD exacerbations.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.