21 September 2023 in Journal of Medical Internet Research
Background: Sleep disorders, such as obstructive sleep apnea (OSA), comorbid insomnia and sleep apnea (COMISA), and insomnia are common and can have serious health consequences. However, accurately diagnosing these conditions can be challenging as a result of the underrecognition of these diseases, the time-intensive nature of sleep monitoring necessary for a proper diagnosis, and patients’ hesitancy to undergo demanding and costly overnight polysomnography tests. Objective: We aim to develop a machine learning algorithm that can accurately predict the risk of OSA, COMISA, and insomnia with a simple set of questions, without the need for a polysomnography test. Methods: We applied extreme gradient boosting to the data from 2 medical centers (n=4257 from Samsung Medical Center and n=365 from Ewha Womans University Medical Center Seoul Hospital). Features were selected based on feature importance calculated by the Shapley additive explanations (SHAP) method. We applied extreme gradient boosting using selected features to develop a simple questionnaire predicting sleep disorders (SLEEPS). The accuracy of the algorithm was evaluated using the area under the receiver operating characteristics curve. Results: In total, 9 features were selected to construct SLEEPS. SLEEPS showed high accuracy, with an area under the receiver operating characteristics curve of greater than 0.897 for all 3 sleep disorders, and consistent performance across both sets of data. We found that the distinction between COMISA and OSA was critical for accurate prediction. A publicly accessible website was created based on the algorithm that provides predictions for the risk of the 3 sleep disorders and shows how the risk changes with changes in weight or age. Conclusions: SLEEPS has the potential to improve the diagnosis and treatment of sleep disorders by providing more accessibility and convenience. The creation of a publicly accessible website based on the algorithm provides a user-friendly tool for assessing the risk of OSA, COMISA, and insomnia.
1 January 2024 in Journal of Clinical Neurology
Shin S, et al. J Clin Neurol. 2024 Jul;20:e56. https://doi.org/10.3988/jcn.2024.0042
1 December 2022 in Journal of Sleep Medicine
Objectives: Nicotine stimulates release of neurotransmitters that regulate the sleep-wake cycle and thereby leads to insomnia. Smoking is associated with upper airway distress; however, its role in severe sleep-related breathing disorders remains controversial. In this study, we investigated the effects of smoking on obstructive sleep apnea (OSA).Methods: We investigated 1,163 patients diagnosed with OSA who underwent polysomnography between March 2020 and July 2022. We recorded details including smoking status (current, former, and non-smoker), demographics, questionnaire-related data, and polysomnography findings and performed univariate analysis to compare these variables between smokers and non-smokers. We also analyzed the correlation between smoking status and OSA severity. The risk of smoking on the severity of OSA was determined using logistic regression analysis.Results: Current and former smokers included 461 male (49.1%) and 10 female (4.4%) (p=0.001). Smokers had a high apnea-hypopnea index (AHI) and oxygen desaturation index (ODI) (p<0.001), high arousal index (p=0.001), and severe daytime sleepiness (indicated by the Epworth Sleepiness Scale, p15, odds ratio 1.33, p=0.04), and AHI was independent of smoking.Conclusions: Controversy regarding the severity of OSA with smoking has currently not been definitively determined. However, our results provide new evidence to support the association between smoking and the ODI, which few studies have investigated to date.
14 February 2023 in JMIR Preprints
BACKGROUND Sleep disorders, such as obstructive sleep apnea (OSA), comorbid insomnia and sleep apnea (COMISA), and insomnia are common and can have serious health consequences. However, accurately diagnosing these conditions can be challenging as a result of the underrecognition of these diseases, the time-intensive nature of sleep monitoring necessary for a proper diagnosis, and patients’ hesitancy to undergo demanding and costly overnight polysomnography tests. OBJECTIVE We aim to develop a machine learning algorithm that can accurately predict the risk of OSA, COMISA, and insomnia with a simple set of questions, without the need for a polysomnography test. METHODS We applied extreme gradient boosting to the data from 2 medical centers (n=4257 from Samsung Medical Center and n=365 from Ewha Womans University Medical Center Seoul Hospital). Features were selected based on feature importance calculated by the Shapley additive explanations (SHAP) method. We applied extreme gradient boosting using selected features to develop a simple questionnaire predicting sleep disorders (SLEEPS). The accuracy of the algorithm was evaluated using the area under the receiver operating characteristics curve. RESULTS In total, 9 features were selected to construct SLEEPS. SLEEPS showed high accuracy, with an area under the receiver operating characteristics curve of greater than 0.897 for all 3 sleep disorders, and consistent performance across both sets of data. We found that the distinction between COMISA and OSA was critical for accurate prediction. A publicly accessible website was created based on the algorithm that provides predictions for the risk of the 3 sleep disorders and shows how the risk changes with changes in weight or age. CONCLUSIONS SLEEPS has the potential to improve the diagnosis and treatment of sleep disorders by providing more accessibility and convenience. The creation of a publicly accessible website based on the algorithm provides a user-friendly tool for assessing the risk of OSA, COMISA, and insomnia.
1 December 2023 in Journal of Sleep Medicine
Objectives: Narcolepsy diagnosis is often complicated, using relatively simple criteria that rely on sleep onset rapid eye movement periods (SOREMP) and average sleep latency in the multiple sleep latency test (MSLT). Therefore, the main goal of this study was to gain a new perspective on narcolepsy diagnosis and explore the role of rapid eye movement (REM) dysregulation by examining the progression of sleep stage transitions leading to SOREMP.Methods: From 2008 to 2022, 109 patients diagnosed with narcolepsy who underwent polysomnography and MSLT were included. Comprehensive data on clinical features, polysomnography, and MSLT findings were collected. For the MSLT, we not only considered SOREMP and mean sleep latency but also the sleep stage transition preceding SOREMP, the presence of N3, and the average SOREMP duration.Results: Among the 109 patients enrolled in the study, 59 were diagnosed with narcolepsy type 1 (NT1) and 50 with narcolepsy type 2 (NT2). We observed a high frequency of transitions in the MSLT of waking to REM (W-R) in patients with NT1. Moreover, the occurrence of W-R was associated with a higher number of SOREMPs. The temporal distribution of the MSLT results revealed a significant correlation; W-R occurrences were more frequent during the third nap.Conclusions: W-R occurrence was frequent in patients with NT1, indicating more pronounced REM dysregulation than in patients with NT2.
30 September 2026 in Sleep Medicine Research
Background and Objective Sleep disorders are common and frequently under-recognized in clinical practice. Existing screening instruments are often disorder-specific or require lengthy assessments. This study aimed to develop and validate the Seoul Sleep Questionnaire (SSQ), a brief screening instrument for common sleep disorders.Methods We retrospectively reviewed data from 1,726 adults who underwent polysomnography at a tertiary sleep clinic between November 2018 and February 2020. Participants completed the SSQ and established sleep questionnaires. Item reduction, exploratory factor analysis (EFA), reliability testing, validity analyses, and receiver operating characteristic (ROC) analyses were performed. To address diagnostic imbalance, sensitivity analyses were conducted using a diagnosis-balanced subsample.Results Item reduction yielded a final 7-item questionnaire. EFA supported a three-factor structure representing obstructive sleep apnea (OSA), insomnia, and parasomnia domains, with good model fit (comparative fit index=0.999, Tucker–Lewis index=0.997, root mean square error of approximation=0.013). Internal consistency was acceptable (McDonald’s ω=0.83). SSQ scores showed significant correlations with subjective and polysomnography-derived sleep measures. The OSA subscale was associated with apnea-hypopnea index (r=0.429, p<0.01), while the insomnia subscale correlated with sleep onset latency and wake after sleep onset. ROC analyses demonstrated acceptable diagnostic performance for OSA (area under the curve [AUC]=0.787) and insomnia (AUC=0.741), and high discriminative ability for parasomnia (AUC=0.919). Comparable results were observed in the diagnosis-balanced subsample.Conclusions The SSQ demonstrated acceptable psychometric properties and diagnostic performance as a brief screening questionnaire in an OSA-dominant tertiary sleep clinic population. Further validation in primary care, community-based, and independent external populations is warranted before broader implementation.