ABSTRACT
OBJECTIVES: A survival analysis of breast cancer patients in southern Iran according to age has yet to be conducted. This study aimed to quantify the factors contributing to a poor prognosis, using Cox and empirical Bayesian additive hazard (EBAH) models, among young (20-39 years), middle-aged (40-64 years), and elderly (≥ 65 years) women.METHODS: Data from 1,574 breast cancer patients diagnosed from 2002 to 2012 in the cancer registry of Fars Province (southern Iran) were stratified into 3 age groups. The Kaplan-Meier method was used to estimate the overall survival rates. Cox and EBAH models were applied to each age category, and the Akaike information criterion was used to assess the goodness-of-fit of the 2 hazard models.RESULTS: As of December 2012, 212 women (13.5%) in our study population had died, of whom 43 were young (15.3%), 134 middle-aged (11.8%), and 35 elderly (22.3%). The 5-year survival probability by age category was 0.83 (standard error [SE], 0.03), 0.88 (SE, 0.01), and 0.75 (SE, 0.04), respectively.CONCLUSIONS: The Nottingham Prognostic Index was the most effective prognostic factor. The model based on Bayesian methodology performed better with various sample sizes than the Cox model, which is the most widely used method of survival analysis.
Subject(s)
Aged , Female , Humans , Breast Neoplasms , Breast , Iran , Methods , Prognosis , Proportional Hazards Models , Sample Size , Survival RateABSTRACT
OBJECTIVES: A survival analysis of breast cancer patients in southern Iran according to age has yet to be conducted. This study aimed to quantify the factors contributing to a poor prognosis, using Cox and empirical Bayesian additive hazard (EBAH) models, among young (20-39 years), middle-aged (40-64 years), and elderly (≥ 65 years) women. METHODS: Data from 1,574 breast cancer patients diagnosed from 2002 to 2012 in the cancer registry of Fars Province (southern Iran) were stratified into 3 age groups. The Kaplan-Meier method was used to estimate the overall survival rates. Cox and EBAH models were applied to each age category, and the Akaike information criterion was used to assess the goodness-of-fit of the 2 hazard models. RESULTS: As of December 2012, 212 women (13.5%) in our study population had died, of whom 43 were young (15.3%), 134 middle-aged (11.8%), and 35 elderly (22.3%). The 5-year survival probability by age category was 0.83 (standard error [SE], 0.03), 0.88 (SE, 0.01), and 0.75 (SE, 0.04), respectively. CONCLUSIONS: The Nottingham Prognostic Index was the most effective prognostic factor. The model based on Bayesian methodology performed better with various sample sizes than the Cox model, which is the most widely used method of survival analysis.