【Abstract】 Background Alcohol use disorders are common chronic relapsing psychiatric disorders, and severe alcohol use disorders require early and rapid recognition and timely and appropriate management to avoid irreversible harm. Currently, the assessment of severity is based on psychiatric examination by the clinician, and there is limited research on the factors affecting AUD severity and prediction models. Objective Analyzing the factors influencing the severity of alcohol use disorder patients in a psychiatric hospital in Nanning, China, and constructing a risk prediction model to provide a reference for assessing the disease progression of alcohol use disorder patients. Methods Retrospectively selected 1 358 first-time hospitalized patients admitted to the Fifth People's Hospital of Nanning City from January 1, 2017 to December 31, 2022 who met the diagnostic criteria of Diagnostic and Statistical Manual of Mental Disorders (5th Edition) (DSM-5) for alcohol use disorders as the study subjects, collected their basic data, and were divided into the mild-moderate group according to the severity of the disease (n = 330) and severe group (n=1 028), and compared the differences in basic data between the two groups. The patients were divided into training and test sets according to 7:3, and a logistic regression model was constructed in the training set samples, and the predictive value of the model for disease severity was analyzed in the test set samples by using the subject's work characteristics (ROC) curve. Results Compared to patients in the mild-moderate group, patients in the severe group had a higher proportion of patients living in urban areas (χ2=7.804), farmers (χ2=17.991), a higher frequency of alcohol consumption higher than 1-2 drinks/day (χ2=35.267), a greater age of first drinking (t=-3.858), a greater greater number of comorbid somatic disorders (Z=-22.782), and higher proportions of gamma-glutamate aminotransferase (χ2=259.940) and total bilirubin abnormalities (χ2=148.552) (P<0.05 or 0.01). The results of the logistic analysis conducted in the training set showed that farmers (OR=2.024, 95% CI: 1.352~3.029), older age at first drinking (OR=1.075, 95% CI: 1.025~1.129), drinking also outside of mealtimes (OR=3.988, 95% CI: 2.408~6.606), total bilirubin abnormalities (OR=1.034, 95% CI: 1.000~1.069), and more comorbid somatic diseases (OR=4.386, 95% CI: 2.636~7.298) were risk factors for disease severity in AUD patients. The area under the curve (AUC) for this model in the test set sample was 0.906. Conclusion In psychiatric hospitals, farmers, older age at first drinking, drinking also outside meal times, abnormal total bilirubin, and comorbidities with more somatic illnesses may be risk factors for severe AUD. |