Cox Proportional Hazard Regression Interaction Model and Its Application to Determine The Risk of Death in Breast Cancer Patients after Chemotherapy

Abstract

Introduction: This research is based on medical record data of breast cancer patients who seek treatment at the Central General Hospital, dr. Sardjito Yogyakarta, from 2018-2021 has as many as 105 patients. Several risk factors for cancer include demographic factors, clinical factors, tumor factors, and therapy. These factors lead to different psychological states of patients, resulting in the rate of recovery and death of patients.
 Objective: To determine the risk of death in breast cancer patients after chemotherapy.
 Methods: The method used in this study is Cox Proportional Hazard survival analysis with an interaction model. The variables studied were age, marital status, profession, insurance, BMI, comorbidities, duration of chemotherapy, chemotherapy agent, chemotherapy type, and tumor size.
 Results: The analysis results using SPSS software obtained the best hazard and survival model with four significant variables, namely the duration of chemotherapy, chemotherapy agents, chemotherapy types, and the interaction between BMI and chemotherapy types.
 Conclusions: The most significant risk factor for death was palliative chemotherapy type with HR 27.195 and 3-5 chemotherapy agents with HR 4.997. Meanwhile, the long duration of chemotherapy and the interaction between lean BMI and palliative chemotherapy reduced the risk of death by HR 0.967 and 0.128, respectively.