Citas bibligráficas
Rebaza, D., (2017). Modelos semiparamétricos de eventos recurrentes: caso aplicación a pacientes con cáncer de mama [Tesis, Universidad Nacional Agraria La Molina]. https://hdl.handle.net/20.500.12996/2884
Rebaza, D., Modelos semiparamétricos de eventos recurrentes: caso aplicación a pacientes con cáncer de mama [Tesis]. : Universidad Nacional Agraria La Molina; 2017. https://hdl.handle.net/20.500.12996/2884
@mastersthesis{renati/246150,
title = "Modelos semiparamétricos de eventos recurrentes: caso aplicación a pacientes con cáncer de mama",
author = "Rebaza Fernández, Diana del Rocío",
publisher = "Universidad Nacional Agraria La Molina",
year = "2017"
}
The recurrence of an event in a patient is the observed frequency of this event over a period of time during follow-up, e.g. successive hospitalizations of pneumonia, episodes of epilepsy, relapses of cancer, among others. Recurrent event models are very useful for application in these phenomena, and the present research is intended to illustrate and compare particular models for recurrent event data without random effect: Andersen and Gill (A-G); Wei, Lin and Weissfeld (WLW); and Prentice, Williams and Peterson (PWP), which are models based on the Cox extension of proportional hazards, in these models assume independence of events. Another studied model is the Gamma Shared Fragility model that considers a term of fragility and assumes that this term influences the recurrence of the events of the same subject. For the estimation of the parameters in the models without random effect, the maximum likelihood method was used, while for the fragility model was the penalized maximum likelihood method, which penalizes the function of base risk. The data used for the application of these methodologies was provided by the physician Gynecologist Oncologist Dr. Vladimir Villoslada Terrones of the National Institute of Neoplastic Diseases (INEN, in its Spanish acronym). These data describe a set of variables related to breast cancer in a prospective cohort of 68 patients with positive diagnosis undergoing mastectomy surgery. When processing and analyzing the obtained results, we found that the model Andersen and Gill (A-G) and Prentice, Williams and Peterson (PWP) are the best fit to this data set. Besides, we found that the factors associated with risk of recurrence of breast cancer are the age of onset of the study, the age of first menstruation (menarche) and lobular carcinoma type. These models present similar results due to the statistical significance in the variables and compliance with the proportional risk assumption.
Este ítem está sujeto a una licencia Creative Commons Licencia Creative Commons