| Objective To construct and validate the prediction model for critical COVID-19 combined with bacterial or fungal infection. Methods?A total of 186 patients diagnosed with critical COVID-19 in our hospital from December 1, 2022 to January 11, 2023 were retrospectively selected as subjects. One hundred and twenty-seven cases from December 1 to 28, 2022 were used as the modeling group, and 59 cases from December 29, 2022 to January 11, 2023 were used as the validation group. The subjects were divided into infected and non-infected groups according to whether they had bacterial or fungal infections. General and clinical data of the subjects before obtaining positive culture results were recorded, independent risk factors were screened by logistic regression analysis, a column graph was constructed, and the accuracy of the model was verified. Results?Logistic regression analysis showed that age, APACHEⅡ, number of underlying diseases, disturbance of consciousness, and use of ventilator were independent risk factors for bacterial or fungal infection in patients with severe COVID-19. Based on this, a line graph model is established. Hosmer-Lemeshouw test results showed that P=0.459 in the modeling group and P=0.982 in the verification group, indicating a good fit. Receiver operating characteristic (ROC) showed that this line graph model had good differentiation, and the areas under the curve predicted by the line graph models of the modeling group and the validation group were 0.941 (95%CI: 0.899～0.983) and 0.843 (95%CI: 0.728～0.959). Decision curve analysis shows that the prediction model is of high value. Conclusion?The graph model for predicting the risk of bacterial or fungal infection based on the risk factors of patients with severe COVID-19 has certain practical value, and can provide reference for the early identification and management of high-risk patients for health care workers. |