Objective: To filter the risk factors of breast cancer. To study the risk evaluation model. Then compare it with the Gail model, which has been already widely used in USA.Methods: Apply logistic regression model to filter the risk factors of breast cancer. Then use these important factors to establish the new risk evaluation model. Evaluate the model by discriminant analysis. Furthermore ROC was used to compare this model with the Gail model, and to find the appropriate cutpoints.Results: The potential risk factors of breast caner included menopause, abortion, medicine-contraception, breast cancer-family history, bad motion, undesirable life, the absence of motion-adjusting ability, excess working, and current age. In addition, the breast caner risk assessment model was developed. Then the cutpoints were set up according to the screening indexes and the demand of clinical practice. We suggested, regarding the predicated probability≤0.29 as the low risk, the probability≥0.58 as the high risk and the probability between 0.29 and 0.58 as the moderate risk.Conclusions: The Gail and the new model established were just preliminary self-assessment tools. When these tools were used to evaluate the risk of breast cancer, some other critical factors should be considered at the same time.
|