Sardar Jahani - MSc, Department of Biostatistics, School of Medicine, Arak University of Medical Sciences, Arak, Iran- Department of Biostatistics, School of Public Health, Hamadan University of Medical Sciences, Hamadan, Iran
Mina Hoseini - MSc, Department of Biostatistics and Epidemiology, School of Public Health, Kerman University of Medical Sciences, Kerman, Iran
Rashed Pourhamidi - MSc, Non Communicable Diseases Research Center, Bam University of Medical Sciences, Bam, Iran
Mahshid Askari - MSc, Department of Biostatistics, School of Medicine, Arak University of Medical Sciences, Arak, Iran
Azam Moslemi - PhD, Department of Biostatistics, School of Medicine, Arak University of Medical Sciences, Arak, Iran

Background: Breast cancer is one of the most common causes of death among women worldwideand the second leading cause of death among Iranian women. The incidence of this malignancy inIran is ۲۲ per ۱۰۰,۰۰۰ women. These patients have long-term survival time with advances inmedical sciences. The present study aimed to identify the risk factors of breast cancer using Coxproportional hazard and Cox mixture cure models.Study design: It is a retrospective cohort study.Methods: In this cohort study, we recorded the survival time of ۱۴۰ breast cancer patients referredto Ali Ibn Abitaleb Hospital in Rafsanjan, Iran, from ۲۰۰۱ to ۲۰۱۵. The Kaplan-Meier curve wasplotted; moreover, two Cox proportional hazards and the Cox mixture cure models were fitted forthe patients. Data analysis was performed using SAS ۹.۴ M۵ software.Results: The mean age of patients was reported as ۴۷.۱۲ ±۱۲.۴۸ years at the commencement ofthe study. Moreover, ۸۳.۵۷% of patients were censored. The stage of disease was a significantvariable in Cox and the survival portion of Cox mixture cure models (P=۰.۰۰۱). The consumptionof herbal tea, tumor size, duration of the last lactation, family history of cancer, and the type oftreatment were significant variables in the cured proportion of the Cox mixture cure model(P=۰.۰۰۱).Conclusion: The Cox mixture cure model is a flexible model which is able to distinguish betweenthe long-term and short-term survival of breast cancer patients. For breast cancer patients, cureeffective factors were the stage of the disease, consumption of herbal tea, tumor size, duration ofthe last lactation, family history, and the type of treatment.

Breast cancer, Survival analysis, Cox model, Long-term survivors