Development and assessment of a risk prediction model for moderate-to-severe obstructive sleep apnea

• DOI: 10.3389/fnins.2022.936946
• Publication Date: 2022-08-05T22:05:35Z
• Authors: Xiangru Yan; Liying Wang; Chunguang Liang; Huiying Zhang; Ying Zhao; Hui Zhang; Haitao Yu; Jinna Di
• Keywords: Sleep

OSA is an independent risk factor for several systemic diseases. Compared with mild OSA, patients moderate-to-severe have more severe impairment in the function of all organs body. Due to current limited medical condition, not every patient can be diagnosed and treated time. To enable timely screening we selected easily accessible variables establish a prediction model.We collected 492 who had polysomnography (PSG), divided them into disease-free group (control group), according PSG results. Variables entering model were identified by random forest plots, univariate analysis, multicollinearity test, binary logistic regression method. Nomogram created based on results, area under ROC curve was used evaluate discriminative properties nomogram model. Bootstrap method internally validate model, calibration curves plotted after 1,000 replicate sampling original data, accuracy evaluated using Hosmer-Lemeshow goodness-of-fit test. Finally, performed decision analysis (DCA) STOP-Bang questionnaire (SBQ), NoSAS score assess clinical utility.There are 6 final namely BMI, Hypertension, Morning dry mouth, Suffocating awake at night, Witnessed apnea, ESS total score. The AUC this 0.976 (95% CI: 0.962-0.990). test χ2 = 3.3222 (P 0.1899 > 0.05), general agreement ideal curve. has good consistency predicting actual occurrence risk, accuracy. DCA shows that net benefit higher than SBQ NoSAS, utility.The obtained study predictive power superior other models questionnaires. It applied community population clinic prioritization treatment.