Modeling the Survival of Retrospective Clinical Data from Prostate Cancer Patients in Komfo Anokye Teaching Hospital, Ghana
Asiedu-Addo, Samuel KKwesi; Nwi-Mozu, Isaac; Ali, Mohammed

Abstract
In this study, we model and estimate how long a patient at Komfo Anokye Hospital can live with prostate cancer diseases knowing the status of the cancer. A total of 199 patients (2005-2013) were retrieved. Variables considered were age, treatment and stage of the cancer. AIC was used to select the best parametric model. By comparing different parametric models using AIC, the result indicated that Weibull model fit best. A straight line graph of K-M survival against time {log(-log S(t)) vs. log t)} produced confirmed Weibull assumption. A plot of Kaplan Meier hazard with the Weibull hazard further confirmed the Weibull assumption.

Full Text: PDF