Prediction of high-level fear of cancer recurrence in breast cancer survivors: An integrative approach utilizing random forest algorithm and visual nomogram.

PURPOSE: This study is the first attempt to use a combination of regression analysis and random forest algorithm to predict the risk factors for high-level fear of cancer recurrence and develop a predictive nomogram to guide clinicians and nurses in identifying high-risk populations for high-level fear of cancer recurrence.

METHODS: After receiving various recruitment strategies, a total of 781 survivors who had undergone breast cancer resection within 5 years in four Grade-A hospitals in China were included. Besides demographic and clinical characteristics, variables were also selected from the perspectives of somatic, cognitive, psychological, social and economic factors, all of which were measured using a scale with high reliability and validity. This study established univariate regression analysis and random forest model to screen for risk factors for high-level fear of cancer recurrence. Based on the results of the multi-variable regression model, a nomogram was constructed to visualize risk prediction.

RESULTS: Fatigue, social constraints, maladaptive cognitive emotion regulation strategies, meta-cognition and age were identified as risk factors. Based on the predictive model, a nomogram was constructed, and the area under the curve was 0.949, indicating strong discrimination and calibration.

CONCLUSIONS: The integration of two models enhances the credibility of the prediction outcomes. The nomogram effectively transformed intricate regression equations into a visual representation, enhancing the readability and accessibility of the prediction model's results. It aids clinicians and nurses in swiftly and precisely identifying high-risk individuals for high-level fear of cancer recurrence, enabling the development of timely, predictable, and personalized intervention programs for high-risk patients.