Identifying individual sleep apnea/hypoapnea epochs using smartphone-based pulse oximetry

Ainara Garde, Parastoo Dekhordi, John Mark Ansermino, Guy A. Dumont

Research output: Chapter in Book/Report/Conference proceedingChapterAcademicpeer-review

4 Citations (Scopus)

Abstract

Sleep apnea, characterized by frequent pauses in breathing during sleep, poses a serious threat to the healthy growth and development of children. Polysomnography (PSG), the gold standard for sleep apnea diagnosis, is resource intensive and confined to sleep laboratories, thus reducing its accessibility. Pulse oximetry alone, providing blood oxygen saturation (SpO2) and blood volume changes in tissue (PPG), has the potential to identify children with sleep apnea. Thus, we aim to develop a tool for at-home sleep apnea screening that provides a detailed and automated 30 sec epoch-by-epoch sleep apnea analysis. We propose to extract features characterizing pulse oximetry (SpO2 and pulse rate variability [PRV], a surrogate measure of heart rate variability) to create a multivariate logistic regression model that identifies epochs containing apnea/hypoapnea events. Overnight pulse oximetry was collected using a smartphone-based pulse oximeter, simultaneously with standard PSG from 160 children at the British Columbia Children's hospital. The sleep technician manually scored all apnea/hypoapnea events during the PSG study. Based on these scores we labeled each epoch as containing or not containing apnea/hypoapnea. We randomly divided the subjects into training data (40%), used to develop the model applying the LASSO method, and testing data (60%), used to validate the model. The developed model was assessed epoch-by-epoch for each subject. The test dataset had a median area under the receiver operating characteristic (ROC) curve of 81%; the model provided a median accuracy of 74% sensitivity of 75%, and specificity of 73% when using a risk threshold similar to the percentage of apnea/hypopnea epochs. Thus, providing a detailed epoch-by-epoch analysis with at-home pulse oximetry alone is feasible with accuracy, sensitivity and specificity values above 73% However, the performance might decrease when analyzing subjects with a low number of apnea/hypoapnea events.
Original languageEnglish
Title of host publication38th Annual IEEE International Conference of the Engineering in Medicine and Biology Society, EMBC 2016
Subtitle of host publication16-20 Aug. 2016, Orlando, FL, USA
PublisherIEEE
Pages3195-3198
Number of pages4
ISBN (Electronic)978-1-4577-0220-4
ISBN (Print)978-1-4577-0219-8
DOIs
Publication statusPublished - 13 Oct 2016
Externally publishedYes
Event38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2016 - Disney’s Contemporary Resort at Walt Disney World® Resort, Orlando, United States
Duration: 16 Aug 201620 Aug 2016
Conference number: 38
https://embc.embs.org/2016/

Conference

Conference38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2016
Abbreviated titleEMBC
CountryUnited States
CityOrlando
Period16/08/1620/08/16
Internet address

Fingerprint

Oximetry
Sleep Apnea Syndromes
Apnea
Polysomnography
Sleep
Heart Rate
Logistic Models
Sensitivity and Specificity
British Columbia
Blood Volume
Growth and Development
ROC Curve
Pulse
Respiration
Smartphone
Oxygen

Cite this

Garde, A., Dekhordi, P., Ansermino, J. M., & Dumont, G. A. (2016). Identifying individual sleep apnea/hypoapnea epochs using smartphone-based pulse oximetry. In 38th Annual IEEE International Conference of the Engineering in Medicine and Biology Society, EMBC 2016: 16-20 Aug. 2016, Orlando, FL, USA (pp. 3195-3198). IEEE. https://doi.org/10.1109/EMBC.2016.7591408
Garde, Ainara ; Dekhordi, Parastoo ; Ansermino, John Mark ; Dumont, Guy A. / Identifying individual sleep apnea/hypoapnea epochs using smartphone-based pulse oximetry. 38th Annual IEEE International Conference of the Engineering in Medicine and Biology Society, EMBC 2016: 16-20 Aug. 2016, Orlando, FL, USA. IEEE, 2016. pp. 3195-3198
@inbook{a1232d890cae49959976ba5b6d6eb910,
title = "Identifying individual sleep apnea/hypoapnea epochs using smartphone-based pulse oximetry",
abstract = "Sleep apnea, characterized by frequent pauses in breathing during sleep, poses a serious threat to the healthy growth and development of children. Polysomnography (PSG), the gold standard for sleep apnea diagnosis, is resource intensive and confined to sleep laboratories, thus reducing its accessibility. Pulse oximetry alone, providing blood oxygen saturation (SpO2) and blood volume changes in tissue (PPG), has the potential to identify children with sleep apnea. Thus, we aim to develop a tool for at-home sleep apnea screening that provides a detailed and automated 30 sec epoch-by-epoch sleep apnea analysis. We propose to extract features characterizing pulse oximetry (SpO2 and pulse rate variability [PRV], a surrogate measure of heart rate variability) to create a multivariate logistic regression model that identifies epochs containing apnea/hypoapnea events. Overnight pulse oximetry was collected using a smartphone-based pulse oximeter, simultaneously with standard PSG from 160 children at the British Columbia Children's hospital. The sleep technician manually scored all apnea/hypoapnea events during the PSG study. Based on these scores we labeled each epoch as containing or not containing apnea/hypoapnea. We randomly divided the subjects into training data (40{\%}), used to develop the model applying the LASSO method, and testing data (60{\%}), used to validate the model. The developed model was assessed epoch-by-epoch for each subject. The test dataset had a median area under the receiver operating characteristic (ROC) curve of 81{\%}; the model provided a median accuracy of 74{\%} sensitivity of 75{\%}, and specificity of 73{\%} when using a risk threshold similar to the percentage of apnea/hypopnea epochs. Thus, providing a detailed epoch-by-epoch analysis with at-home pulse oximetry alone is feasible with accuracy, sensitivity and specificity values above 73{\%} However, the performance might decrease when analyzing subjects with a low number of apnea/hypoapnea events.",
author = "Ainara Garde and Parastoo Dekhordi and Ansermino, {John Mark} and Dumont, {Guy A.}",
year = "2016",
month = "10",
day = "13",
doi = "10.1109/EMBC.2016.7591408",
language = "English",
isbn = "978-1-4577-0219-8",
pages = "3195--3198",
booktitle = "38th Annual IEEE International Conference of the Engineering in Medicine and Biology Society, EMBC 2016",
publisher = "IEEE",
address = "United States",

}

Garde, A, Dekhordi, P, Ansermino, JM & Dumont, GA 2016, Identifying individual sleep apnea/hypoapnea epochs using smartphone-based pulse oximetry. in 38th Annual IEEE International Conference of the Engineering in Medicine and Biology Society, EMBC 2016: 16-20 Aug. 2016, Orlando, FL, USA. IEEE, pp. 3195-3198, 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2016, Orlando, United States, 16/08/16. https://doi.org/10.1109/EMBC.2016.7591408

Identifying individual sleep apnea/hypoapnea epochs using smartphone-based pulse oximetry. / Garde, Ainara; Dekhordi, Parastoo; Ansermino, John Mark; Dumont, Guy A.

38th Annual IEEE International Conference of the Engineering in Medicine and Biology Society, EMBC 2016: 16-20 Aug. 2016, Orlando, FL, USA. IEEE, 2016. p. 3195-3198.

Research output: Chapter in Book/Report/Conference proceedingChapterAcademicpeer-review

TY - CHAP

T1 - Identifying individual sleep apnea/hypoapnea epochs using smartphone-based pulse oximetry

AU - Garde, Ainara

AU - Dekhordi, Parastoo

AU - Ansermino, John Mark

AU - Dumont, Guy A.

PY - 2016/10/13

Y1 - 2016/10/13

N2 - Sleep apnea, characterized by frequent pauses in breathing during sleep, poses a serious threat to the healthy growth and development of children. Polysomnography (PSG), the gold standard for sleep apnea diagnosis, is resource intensive and confined to sleep laboratories, thus reducing its accessibility. Pulse oximetry alone, providing blood oxygen saturation (SpO2) and blood volume changes in tissue (PPG), has the potential to identify children with sleep apnea. Thus, we aim to develop a tool for at-home sleep apnea screening that provides a detailed and automated 30 sec epoch-by-epoch sleep apnea analysis. We propose to extract features characterizing pulse oximetry (SpO2 and pulse rate variability [PRV], a surrogate measure of heart rate variability) to create a multivariate logistic regression model that identifies epochs containing apnea/hypoapnea events. Overnight pulse oximetry was collected using a smartphone-based pulse oximeter, simultaneously with standard PSG from 160 children at the British Columbia Children's hospital. The sleep technician manually scored all apnea/hypoapnea events during the PSG study. Based on these scores we labeled each epoch as containing or not containing apnea/hypoapnea. We randomly divided the subjects into training data (40%), used to develop the model applying the LASSO method, and testing data (60%), used to validate the model. The developed model was assessed epoch-by-epoch for each subject. The test dataset had a median area under the receiver operating characteristic (ROC) curve of 81%; the model provided a median accuracy of 74% sensitivity of 75%, and specificity of 73% when using a risk threshold similar to the percentage of apnea/hypopnea epochs. Thus, providing a detailed epoch-by-epoch analysis with at-home pulse oximetry alone is feasible with accuracy, sensitivity and specificity values above 73% However, the performance might decrease when analyzing subjects with a low number of apnea/hypoapnea events.

AB - Sleep apnea, characterized by frequent pauses in breathing during sleep, poses a serious threat to the healthy growth and development of children. Polysomnography (PSG), the gold standard for sleep apnea diagnosis, is resource intensive and confined to sleep laboratories, thus reducing its accessibility. Pulse oximetry alone, providing blood oxygen saturation (SpO2) and blood volume changes in tissue (PPG), has the potential to identify children with sleep apnea. Thus, we aim to develop a tool for at-home sleep apnea screening that provides a detailed and automated 30 sec epoch-by-epoch sleep apnea analysis. We propose to extract features characterizing pulse oximetry (SpO2 and pulse rate variability [PRV], a surrogate measure of heart rate variability) to create a multivariate logistic regression model that identifies epochs containing apnea/hypoapnea events. Overnight pulse oximetry was collected using a smartphone-based pulse oximeter, simultaneously with standard PSG from 160 children at the British Columbia Children's hospital. The sleep technician manually scored all apnea/hypoapnea events during the PSG study. Based on these scores we labeled each epoch as containing or not containing apnea/hypoapnea. We randomly divided the subjects into training data (40%), used to develop the model applying the LASSO method, and testing data (60%), used to validate the model. The developed model was assessed epoch-by-epoch for each subject. The test dataset had a median area under the receiver operating characteristic (ROC) curve of 81%; the model provided a median accuracy of 74% sensitivity of 75%, and specificity of 73% when using a risk threshold similar to the percentage of apnea/hypopnea epochs. Thus, providing a detailed epoch-by-epoch analysis with at-home pulse oximetry alone is feasible with accuracy, sensitivity and specificity values above 73% However, the performance might decrease when analyzing subjects with a low number of apnea/hypoapnea events.

U2 - 10.1109/EMBC.2016.7591408

DO - 10.1109/EMBC.2016.7591408

M3 - Chapter

SN - 978-1-4577-0219-8

SP - 3195

EP - 3198

BT - 38th Annual IEEE International Conference of the Engineering in Medicine and Biology Society, EMBC 2016

PB - IEEE

ER -

Garde A, Dekhordi P, Ansermino JM, Dumont GA. Identifying individual sleep apnea/hypoapnea epochs using smartphone-based pulse oximetry. In 38th Annual IEEE International Conference of the Engineering in Medicine and Biology Society, EMBC 2016: 16-20 Aug. 2016, Orlando, FL, USA. IEEE. 2016. p. 3195-3198 https://doi.org/10.1109/EMBC.2016.7591408