Determining the Effective Factors on Gastric Cancer Using Frailty Model in South-East and North of Iran

Abstract

Background and Purpose: Gastric cancer is the third leading cause of mortality in Iran after cardiovascular diseases and accidents. The aim of the present study was to assess survival and it’s affecting factors in gastric cancer patients through using Cox and parametric models along with frailty.

Materials and Methods: In this study, the medical records of gastric cancer patients treated from 2008 to late 2010 were collected in Afzalipour and Bahonar Hospitals in Kerman and Imam Khomeini Hospital in Sari. 383 patients entered the study and were followed up for at least five years. The survival of patients was assessed by using Cox proportional hazard, log-normal and log-logistic models under gamma and inverse-Gaussian distributions, as two special models for frailty. Models efficiency comparison criteria were Akaike information criterion and Cox-Snell residuals.

Results: Out of 196 patients in Kerman, 132(67.3%) were males and 64(32.7%) were females. The average age of the patient was 61yr and 59 yr for the males and females, respectively. Also, the survival rates after 1, 3, and 5 years of the diagnosis were 62%, 50%, and 45%, respectively. In the city of Sari, 69% (129 people) of the patients were male and 31% were female. The mean ages of male and female were 66 and 62 yr, respectively. At the same time, 1, 3, and 5 year survival rates of patients were 58%, 36%, and 30%, respectively. Based on Akaike information criterion, Cox-Snell residuals, and non-monotonic failure rate, log-logistic model along with gamma frailty was more fitted in comparison with other models. Using this model, radiotherapy, heartburn, and tumor grade were found as significant predictors.

Conclusion: Radiotherapy, heartburn, and tumor grade could be considered as more affected factors. According to rejection of the proportional hazard assumption, assessments of residual figures, and according to non-significant frailty effect by log-normal model, log-logistic model along with gamma frailty was found to be the best fitted model.