M. Nadjib Bustan, M. Arif Tiro, Adiatma
Accuracy of breast cancer diagnosis is required to provide appropriate action or treatment. Classification of breast cancer diagnosis can be divided into various stadiums from stages 0 through IV. The classification of the breast cancer stadiums is probably related to risk factors, clinical complaints and outcome of the given treatment. The aim of the study was to determine the most appropriate model of breast cancer classification in diagnosis based on risk factors, clinical symptoms and treatment received by the patients. Therefore, predictor factors in this study was consisting of the location of cancer, the presence of metastasis, chemotherapy, age, weight, marital status and parity. The data came from medical records of breast cancer patients treated at the hospital Dr. Wahidin Sudirohusodo, Makassar, Indonesia during the year 2016 as many as 261 where drawn 181 samples. The purpose of analysis was to determine the magnitude of the association between predictors and the classification of breast cancer diagnoses in logistic regression models. The most appropriate logistic regression model test was multinomial logistics where the values of Pearson and Deviance Chi-square Goodness of Fit showed statistically accepted, with pseudo r-square Nagelkerke 0.658. In addition, McFadden value 0.421 means this model is able to explain the variation of breast cancer stage classification by 42%. The analysis showed that the value of simultaneous test of logistic regression -2log-likelihood was 209.914 (p <0,001), where the main predictors were age and chemotherapy with g 0(x) = -42,903 - 0,069Age - 1,987 Chemotheraphy; g 2(x) = -42,646 - 0,069Age - 1,987 Chemotheraphy and. g 3(x) = -24,254 - 0,069Age - 1,987 Chemotheraphy Moreover, without chemotherapy, breast cancer patients had an increased chance of getting into a higher stage with age, proportional to the stage of breast cancer diagnosis. The model explain the chances of an increased likelihood of clinical stages was related to the age and chemotherapy treatment. © Published under licence by IOP Publishing Ltd.
Department of Statistics, Universitas Negeri Makassar, Makassar, 90222, Indonesia
