Mortality risk factors in patients with gastric cancer using Bayesian and ordinary Lasso logistic models: A study in the Southeast of Iran

Abstract

Aim: The aim of this study was to apply two types of statistical models to determine the factors that influence the mortality rate in patients with gastric cancer. Background: In Iran, gastric cancer ranks the first and second most prevalent among men and women, respectively. It is the first cause of death in Iran in both gendersival. Methods: In this retrospective study, data were obtained from 339 (216 male) patients diagnosed with gastric cancer in the city of Kerman (South-East of Iran) during 2001-2015. In this study, ordinary and Bayesian Lasso (least absolute shrinkage and selection operator) logistic regression models, with goodness-of-fit indices, were compared and the models' risk factors were also determined. Results: The mean age of the participants was 62.84-14.53 years, and 12.4% of them were younger than 45 years. Also, the mortality rate was 57.7%. Gender, morphology of the tumor, and time of diagnosis were found to be significant factors in the mortality of the patients in both models. This study found that the Bayesian Lasso model had better fitness. Conclusion: The high mortality rate of gastric cancer and its high prevalence at age below 45 years are alarming. Thus, great attention should be paid to prevention, early diagnosis, especially in females, and adenocarcinoma to improve the survival of patients with gastric cancer.