Statistical Analysis Of Breast Cancer Types and Associated Factors: Application of Bayesian approach.

Abstract

Introduction: Breast cancer is a form of the disease characterized by cells in the breast becoming ab- normal and multiplying uncontrollably, resulting in a tumor. Breast cancer(BC) is the most common public health problem and the main cause of cancer-related death worldwide. In 2020, there were about 2.3 million new cases of breast cancer globally and about 685,000 deaths from this disease, with large geographical variations observed between countries and world regions. Objectives: This study has aimed to determine significant predictors of breast cancer for women by applying Bayesian approaches to multinomial logistic regression analyses. It is also designed to de- termine the risk factors and see the diagnosis of breast cancer patients at Jimma University Medical Center. Methods: The study is conducted at Jimma University Medical center and the data is secondary, which is obtained from Jimma University Medical center, Oncology Department. The response vari- able is the histologic type of breast cancer disease, whether it is ductal carcinoma, lobular carcinoma or it is other types of carcinoma. The multinomial unordered category of breast cancer cases has been analyzed with factors like age of patients, residence area, marital status, stages, treatments that a patient received, family history, HRT(hormone replacement therapy) and oral contraceptives. Bayesian Multinomial logistic regression is employed for posterior model fitting by using the MCMC Metropolis algorithm made for parameters. Results:Approximately 63.48%, 17.06%, and 19.46% of breast cancer women were attacked by duc- tal, lobular, and other histologic types of carcinoma respectively, at Jimma medical center. Women fromurbanareaswere(RRR=0.036; 95%CrI(0.005,0.162))and(RRR=0.138;95%CrI(0.0444, 0.432)) less likely to have lobular and others histologic type of carcinoma respectively compared to women from rural areas. The patients who received treatment surgery and radiotherapy were (RRR= 6.706; 95%CrI(1.887,24.532)) and (RRR=7.924;95%CrI(1.896,36.162))respectively, more likely to have lobular carcinoma compared to the patients who received chemotherapy. Conclussion:Based on the study results we can conclude that age, residence area, treatments,oral contraceptives and releated factors had a significant effect on breast cancer patients