Objective:To establish a simple and efficient clinical prediction model of moderate and severe obstructive sleep apnea hypopnea (OSAHS) in snoring patients based on the clinical data and morphological measurement data in order to increase the early diagnosis and then early intervention of OSAHS. The prediction model is evaluated by external validation.Methods:A total of 299 subjects from January 2015 to December 2018 were selected to perform polysomngraphy (PSG) in Yangpu Hospital, Tongji University School of Medicine. According to the PSG results, they were divided into moderate and severe OSAHS groups (143 cases) and control groups (156 cases). Clinical complications data and morphological measurement data were collected. The regression equation and ROC curve were established according to the Logistic regression method. Then, another 110 subjects from January 2019 to October 2019 were chosen as verified data group, and used to verify the accuracy of the prediction model. The data of 110 subjects were put into the equation according to risk factors and assignment. The ROC curve was drawn and the area under the curve was calculated. The sensitivity, specificity, accuracy, positive predictive value and negative predictive value were calculated.Results:The predicted equation was: y = -10.707 86+0.589 60 × sex+ 0.141 61 × BMI+ 1.281 62 × tonsil size degree+ 1.807 43 × modified Mallampati degree′tongue position. The AUC of the ROC curve of prediction model in training set was 0.851(95% CI 0.807-0.895), the sensitivity was 83.9%, the specificity was 79.5%, and the cut-off value was 0.634.The AUC of the ROC curve in validation set was 0.827(95% CI 0.751-0.904) with a sensitivity of 73.3% and a specificity of 86.0%, and an accuracy of 79.1%. Its positive predictive value was 5.238, and negative predictive value was 0.310. Conclusions:The predictive model constructed by the combination of clinically accessible data (sex) and morphological measurement (BMI, tonsil size degree, modifiedMallampatidegree) has a relatively high predictive efficiency for screening snoring patients with moderate and severe OSAHS. The predictive model is proved with good forecast accuracy by the external verification method.