PERBANDINGAN MODEL REGRESI STRATIFIED COX DAN EXTENDED COX PADA ANALISIS SURVIVAL PENDERITA KANKER PAYUDARA

Breast cancer is a tumor that develops in the breast where cells in the mammary gland divide and develop uncontrollably. Breast cancer is the most common cancer that causes death in women among other cancers. This study aims to determine the factors that affect the survival of breast cancer patients from the METABRIC database. This study was analyzed using survival analysis method. The method that is often used is the Cox proportional hazard model where the proportional hazard assumption must be met. There is variable that do not meet the assumption so that the methods used are stratified Cox and extended Cox. The stratified Cox model overcomes variables that do not meet the assumption by stratifying variables that do not meet the assumption. The extended Cox model overcomes variables that do not meet the assumption by interacting the variables with a time function. The time functions used in this study are linear time functions and logarithmic time functions. Based on the smallest AIC value, the best model is the stratified Cox regression model without interaction. Factors that affect the survival of breast cancer patients from the METABRIC database are tumor size, chemotherapy, stage 1, stage 2, and type of surgery.