Influencing Factors And Construction Of Risk Prediction Model For Fear Of Cancer Recurrence In Convalescent Breast Cancer Patients

Background and PurposeBreast cancer has become the most common malignant tumor in China,and its5-year survival rate can reach 72.7%.For cancer survivors,the recurrence of cancer fear is ubiquitous psychological problem,research has shown that in cancer patients,49% of patients had higher levels of cancer recurrence of fear,a high level of fear will affect the quality of life of patients with cancer recurrence and increase the risk of cancer recurrence,excessive use of health resources,and so on and so forth.Meanwhile,breast cancer patients scored highest among all cancer patients for fear of cancer recurrence.At present,most studies at home and abroad focus on newly diagnosed cancer patients,while there is no in-depth study on patients in the stage of cancer recovery.At the same time,the fear of cancer recurrence significantly affects the quality of life of patients in the rehabilitation period.The common sense model of disease self-regulation points out that external factors,such as existing cognition,emotion and social support,will affect patients’ psychology when facing the threat of disease.Therefore,this study based on self-adjusting knowledge model,and rehabilitation of breast cancer patients through the research of the influence factors of recurrence of fear and use Logistic,random forests in machine learning and deep learning methods of artificial neural network of three fear recovery period in patients with breast cancer recurrence prediction model was constructed,so as to provide reference for related research to reduce the fear of recurrence and improve the quality of life.MethodsBased on the "common sense model of disease self-regulation",the influential factors affecting the recurrence fear of breast cancer patients in the recovery stage were preliminarily screened through literature review,and the effective influential factors were determined through single factor analysis.Based on the results of univariate analysis,the factors obtained from multivariate analysis were taken as independent variables,and the occurrence of high recurrence fear was taken as dependent variables to establish a Logistic model.The 17 influencing factors obtained from literature studies were taken as independent variables,and the occurrence of high recurrence fear was taken as the dependent variable to establish the artificial neural network model and random forest model,and 10×10 cross-validation was used to verify the model.At the same time,the sensitivity specificity and accuracy of AUC values of the three prediction models were compared to determine the most accurate and reliable prediction model for the fear of recurrence of breast cancer in the convalescent stage.ResultsThe literature to determine fear factors including social demographic factors of recurrence,clinical factors and cognitive,emotional factors,such as a total of 17 after multiariable analysis select 6 factors influencing the recurrence of fear,age,occupation state respectively,chemotherapy,self-image,cancer-related fatigue and illness cognition,three prediction models were established with high fear of recurrence as the dependent variable.The accuracy,sensitivity,specificity,positive predictive value and negative predictive value of Logistic regression model were 83.7%,79.3%,86.9%,85.1% and 81.7%,respectively,and the AUC value was 0.746.The accuracy,sensitivity,specificity,positive predictive value and negative predictive value of the artificial neural network model were 78.2%,80.0%,75.8%,81.9% and 73.4%,respectively,and the AUC value was 0.818.The accuracy,sensitivity,specificity,positive predictive value and negative predictive value of the random forest model were 84.4%,89.4%,77.4%,84.4% and 84.2%,respectively,and the AUC value was0.892.ConclusionAge,occupational status,chemotherapy,self-image,cancer-related fatigue and disease perception were the factors that affected the fear of cancer recurrence in convalescent breast cancer patients.The three kinds of prediction models of cancer recurrence fear in convalescent breast cancer patients have good differentiation,but the performance of the random forest model is better than that of the neural network model and the Logistic regression model,and can be used to screen the high-risk population of cancer recurrence fear in convalescent breast cancer patients.