Parametric models for risk modeling in patients with breast cancer

Postgraduate Thesis uoadl:2776620 394 Read counter

Unit:
Postgraduate Programme Biostatistics & Health Science Data
Library of the School of Health Sciences
Deposit date:
2018-07-05
Year:
2016
Author:
Chantzioti Maria
Supervisors info:
Καρλής Δημήτριος, Καθηγητής, Τμήμα Στατιστικής, Οικονομικό Πανεπιστήμιο Αθηνών
Μπουρνέτας Απόστολος, Καθηγητής, Τμήμα Μαθηματικών, Εθνικό και Καποδιστριακό Πανεπιστήμιο Αθηνών
Δαφνή Ουρανία, Καθηγήτρια, Τμήμα Νοσηλευτικής, Εθνικό και Καποδιστριακό Πανεπιστήμιο Αθηνών
Original Title:
Παραμετρικά μοντέλα για τη μοντελοποίηση του κινδύνου σε ασθενείς με καρκίνο του μαστού
Languages:
Greek
Translated title:
Parametric models for risk modeling in patients with breast cancer
Summary:
Scope of the present study was to investigate the hazard function in patients with breast cancer. The hazard function was observed having a special pattern and aim of this study was primarily to modelize this function using parametric models achieving highly flexible functional forms and furthermore to give a provision of the disease progress. Such a result is not feasible with the widely used non-parametric or semi-parametric models at hand.
In contrast to these inadequate models, the distribution that was proposed to be used was the Generalized Gamma distribution and additionally other parametric models with low number of parameters such as Weibull, log Normal and log Logistic distributions were also tested. Generalized Gamma distribution was proved to be more beneficial, followed by log-Normal.
Examining all our data with the distributions that were stood out, the results were very satisfying regarding the flexibility of the Generalized Gamma distribution. Nevertheless, the study has its limitations due to the low number of events in some subsets, as the data weren’t enough to prove the superiority of these distributions, something that can be reassessed later in other clinical studies with a larger number of events.
Main subject category:
Health Sciences
Keywords:
Parametric models, Breast cancer, Risk function
Index:
No
Number of index pages:
0
Contains images:
Yes
Number of references:
14
Number of pages:
58
File:
File access is restricted only to the intranet of UoA.

ΧΑΝΤΖΙΩΤΗ ΜΑΡΙΑ_ΔΙΠΛΩΜΑΤΙΚΗ_MSC.pdf
1 MB
File access is restricted only to the intranet of UoA.