Predicting the Risk of Sleep Disorders Using a Machine Learning-Based Simple Questionnaire: Development and Validation Study.

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.

A general model-based causal inference method overcomes the curse of synchrony and indirect effect.

To identify causation, model-free inference methods, such as Granger Causality, have been widely used due to their flexibility. However, they have difficulty distinguishing synchrony and indirect effects from direct causation, leading to false predictions. To overcome this, model-based inference methods that test the reproducibility of data with a specific mechanistic model to infer causality were developed. However, they can only be applied to systems described by a specific model, greatly limiting their applicability. Here, we address this limitation by deriving an easily testable condition for a general monotonic ODE model to reproduce time-series data. We built a user-friendly computational package, General ODE-Based Inference (GOBI), which is applicable to nearly any monotonic system with positive and negative regulations described by ODE. GOBI successfully inferred positive and negative regulations in various networks at both the molecular and population levels, unlike existing model-free methods. Thus, this accurate and broadly applicable inference method is a powerful tool for understanding complex dynamical systems.

Validation of the Korean version of the Metacognitions Questionnaire-Insomnia (MCQ-I) scale and development of shortened versions using the random forest approach.

We aimed to validate a Korean version of the Metacognitions Questionnaire-Insomnia (MCQ-I) and develop two shortened versions of the MCQ-I by applying the Random Forest (RF) algorithm. A total of 310 participants responded through an online survey, during April 3-6, 2021, which included rating scales such as the Insomnia Severity Index (ISI), Pittsburgh Sleep Quality Index (PSQI), Patient Health Questionnaire-9 (PHQ-9), and the Hospital Anxiety and Depression Scale (HADS), as well as the MCQ-I. After validating the scale, we developed two shortened versions by applying the RF. Finally, we explored the psychometric properties of the shortened versions. The Korean version of the MCQ-I showed good internal consistency based on a Cronbach's alpha of 0.96. Factor analyses showed good model fits for the single structure of the MCQ-I. From the results of the RF, 6 of the 60 items of the MCQ-I were sufficient to distinguish between people with MCQ-I scores above the cut-off value and the rest with high accuracy (AUC>0.97), leading to the 6-item (MCQI-6) version of the MCQ-I. Furthermore, we have also developed a 14-item (MCQI-14) version of the MCQ-I with higher accuracy (AUC>0.98). Both versions were reliable based on their internal consistency (alpha = 0.843 and 0.912), and confirmatory factor analysis showed good model fits for both shortened versions. In addition, good convergent validity of both shortened versions with insomnia, sleep quality, depression, and anxiety were observed. The Korean version of the MCQ-I and two shortened versions (MCQI-6, and MCQI-14) were useful, reliable, and valid tools to evaluate the role of metacognitive beliefs in sleep problems among the Korean population.