|
Abstract:
|
Understanding factors affecting survival outcomes and quality of life is important in advising elderly patients on treatment options for bladder cancer. It has been shown that the mental health plays as important role as physical health in patient outcomes. In this study, we implemented clustering technique to empirically identify subgroups of bladder cancer (BC) patients, with similar level of mental and physical scores. The data was obtained from the Surveillance, Epidemiology, and End Results-Medicare Health Outcomes Survey linked dataset. The gap statistics value was used to identify the optimal number of clusters. The k-mean clustering with input variables age, and the mental and physical scores measured by the SF-36 health related quality of life instruments identified 5 distinct clusters. We analyzed the association between the identified clusters and the patient characteristics. We observed a significant difference in patient reported activities of daily living, number of comorbidities, fall outcome and survival time among the identified clusters. This classification can be used as effective proxies for survival prediction, and will allow for improved decision making.
|
