Previously submitted to: JMIR mHealth and uHealth (no longer under consideration since Nov 04, 2022)
Date Submitted: Oct 14, 2022
Warning: This is an author submission that is not peer-reviewed or edited. Preprints - unless they show as "accepted" - should not be relied on to guide clinical practice or health-related behavior and should not be reported in news media as established information.
Screening the Risk of Obstructive Sleep Apnea by Utilizing Supervised Learning Techniques Based on Anthropometric Features and Snoring Events
ABSTRACT
Background:
Obstructive sleep apnea (OSA) is sleep-disordered breathing and is typically diagnosed by polysomnography (PSG). However, PSG is time-consuming and has some clinical limitations.
Objective:
This study thus aimed to establish machine learning models to screen for the risk of having moderate-to-severe and severe OSA based on easily acquired features (e.g., symptoms or risk factors of OSA).
Methods:
This retrospective study collected data on 3629 patients from Taiwan, who had undergone PSG for symptoms of OSA. Their baseline characteristics, anthropometric measures, and PSG data were obtained. The number of snoring events of PSG was further derived, and correlations among the collected variables were investigated. Next, this study utilized six common supervised machine learning techniques to establish OSA risk screening models, including random forest (RF), XGBoost, k-nearest neighbors, support vector machine, logistic regression, and naïve Bayes. First, data were independently separated into a training and validation dataset (80%) and a test dataset (20%). The approach which had the highest accuracy in the training and validation phase was employed to perform the classification for the test dataset. Moreover, the feature importance of employed models was determined by calculating the Shapley value of every factor, which represented the impact on OSA risk screening.
Results:
RF models manifested the highest accuracy and area under the receiver operating characteristic curve (AUC) in the training and validation phase in screening for both OSA severities (over 70% of accuracy and over 80% of AUC). Hence, we employed the RF to perform the classification of the independent test dataset, and results showed 79.32% accuracy for moderate-to-severe OSA and 74.37% accuracy for severe OSA. Next, snoring events and the visceral fat level were the most and second-most essential features of screening for OSA risk.
Conclusions:
Snoring events and the visceral fat level were essential features of screening for OSA risk. Based on these easily assessed variables, this study established models that can be considered to apply for screening OSA risk in populations with similar craniofacial features.
Citation
Request queued. Please wait while the file is being generated. It may take some time.
Copyright
© The authors. All rights reserved. This is a privileged document currently under peer-review/community review (or an accepted/rejected manuscript). Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review and ahead-of-print citation purposes only. While the final peer-reviewed paper may be licensed under a cc-by license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes.