Analysis And Prediction Of Patient And Treatment Delay Of Breast Cancer Based On Statistical Learning Methods

In recent years,the incidence of breast cancer in China has shown an increasing trend,and it is a serious threat to women's health and survival.Patient and treatment delays are prevalent in breast cancer and can lead to serious clinical characterization and prognosis in breast cancer patients.Studying the delay of breast cancer and taking targeted measures to influence the delay are important for shortening the time of treatment and treatment of breast cancer patients and improving the survival of breast cancer patients.In this paper,the clinical data of 298 breast cancer patients in Dalian were analyzed.The retrospective cohort study was used as a research method.Based on the statistical learning method,the analysis and prediction of breast cancer diagnosis and treatment delay were studied,including the following three aspects:(1)Patient and treatment delays were defined based on survival analysis.Most of the existing research relies entirely on expert experience to define the delay time and does not establish a suitable definition method.The study performed survival analysis based on local clinical data and defined delays based on the p-value of the log-rank test.The Kaplan-Meier survival curves were then grouped according to the delay classification,and the results were further verified using the Cox model.(2)Based on chi-squared test,ANOVA and logistic regression analysis,Correlation analysis was carried out with delays in patients and their clinical manifestations and sociodemographic factors.This paper first analyzed the clinical characterization and delay of patients to discuss the impact of delay on the patient's specific cancer characterization,and further proved the rationality of the delay definition;Then found social demographic factors that affect delays and derive independent influencing factors based on multiple logistic regressions;finally,contacted the local economic and social conditions,analyzed the results and made recommendations for patients,governments or medical institutions to reduce delays.(3)Based on ensemble learning,a delay classification model was established to predict the patient's tendency to delay.In this study,based on the clinical characterization and social demographic factors that were most closely related to delays,the prediction algorithm based on ensemble learning was proposed.The algorithm used logistic regression as the base learner and the gradient boosting was used to integrate the base learner to predict delays.In this paper,survival analysis was used to define the delay of breast cancer for patients who have been diagnosed for more than one month.The results were statistically significant.This conclusion has certain reference significance for patients' self-examination of breast cancer and clinical treatment of the hospital;The association analysis showed that the patient's delay had a significant effect on the clinical characterization,and the patient's marital relationship,income level,hospital grade and health care measures were the social influencing factors of the delay.Medical institutions and governments have implemented guiding measures to reduce breast cancer delays.This paper compared the proposed model with common classification models and ensemble learning methods.The results showed that the proposed algorithm had better prediction results.