关键词: 化疗；恶心；呕吐；预测模型；Logistic回归

Abstract: Objective: To develop a risk prediction model of Chemotherapy-Induced Nausea and Vomiting (CINV), thus to provide a basis for the development of the best antiemetic regimen. Methods: High-risk factors of CINV were screened based on systematic literature review. Chemotherapy patients were investigated in the oncology department of 12 tertiary grade A hospitals in Chongqing from September 2020 to May 2021. The Chinese version of The Multinational Association of Supportive Care in Cancer (MASCC) Antiemetic Tool (MAT) was used to evaluate the occurrence of CINV. The vomiting that not occurrence or nausea NRS score <3 was treated as outcome indicators. The independent risk factors of CINV were analyzed by multi-factor Logistic regression, and the predictive efficacy, sensitivity and specificity of the CINV risk prediction model were evaluated by Receiver Operating Characteristic (ROC) curve analysis. Results: Among 2215 patients, 639 patients (28.8%) developed vomiting or nausea with NRS score≥3. Multivariate Logistic regression analysis screened out 11 independent risk factors for CINV, the Area Under the ROC Curve (AUC) of the constructed CINV risk prediction model was 0.843, the sensitivity was 81.7%, the specificity was 73.3% (P<0.001). The Hosmer-Lemeshow test was used for model fit test with χ2=8.652, P=0.372. Conclusion: The Logistic regression-based CINV risk prediction model shows good discrimination, precision and model performance. It may provide a basis for medical personnel's development of personalized and reasonable antiemetic regimen for the patients.

Key words: chemotherapy; nausea; vomiting; prediction model; Logistic regression