Background and Objective: Breast cancer is one of the most common cancers in women worldwide. This study was conducted to analyze the factors related to survival of patients with breast cancer using two models of Cox proportional specific-cause and sub-distribution models (direct modeling of cumulative incidence). These patients were at the competitive risk of death from breast cancer and also death due to other causes that the occurrence of any of these events prevented another from happening.
Materials and Methods: This historical cohort study comprised 573 breast cancer patients who had referred to the Mahdieh medical center of Hamedan during 2004 to 2011 and were followed until 2015. To determine the risk factors among the competitive risks, the Cox specific-cause and cumulative incidence models were fitted. The data were analyzed using the SPSS and R softwares.
Results: The findings showed that among the causes of breast cancer death in the cumulative incidence model, only the tumor size had a significant effect (p = 0.0054) and for death due to other causes in the Cox model, the tumor size was also significant (P = 0.033).  However, the other variables in two models had no significant effect on death.
Conclusion: Considering the significance of the tumor size in the survival of breast cancer patients, the Cox model takes no account of the other risk information with censorship, the cumulative incidence function is recommended to be used in modeling the risk factors in breast cancer.