Unit:
Postgraduate Programme Biostatistics & Health Science DataLibrary of the School of Health Sciences
Author:
Konidari Konstantina
Supervisors info:
Φώτιος Σιάννης, Επίκουρος Καθηγητής, Τμήμα Μαθηματικών, ΕΚΠΑ
Απόστολος Μπουρνέτας, Καθηγητής, Τμήμα Μαθηματικών, ΕΚΠΑ
Ευαγγελία Σαμόλη, Επίκουρη Καθηγήτρια, Ιατρική, ΕΚΠΑ
Original Title:
Παλινδρόμηση ποσοστημορίων στην μετά- ανάλυση με δεδομένα επιβίωσης
Translated title:
Quantile regression in meta- analysis with time to event data
Summary:
In the present paper an alternative way to resolving survival data through the quota regression is proposed an alternative way to resolving survival data using Quantile regression. Cox model is a well-known model for modeling survival data based on the proportionality of hazard. Proportionality does not always hold and this creates limitations. Quantile regression is a method of data analysis that can provide solution to the above problem as you do not need to assume a distribution for random errors and estimation of quantiles is influenced only by the behavior of the distribution near the particular quantile.
For the purposes of the study, we have simulated 10 studies which are initially analyzed by the Cox model and the censored quantile regression method. Then we compare the two models and study what similarities and differences they have. Finally, we proceed in meta analysis with the estimates from the Cox model and with the estimates from the censored quantile model. In the specific simulated data, the results of the analysis of studies analyzed with the Cox model and the studies analyzed with the censored quantile regression model showed that treatment reduces the risk of the event and increases survival.
Quantile regression may be useful in the modeling of survival data, especially when we are interested in survival or when the the proportionality of hazards does not hold. Even in the case it holds it would be better to use both methods to get more complete and accurate conclusions.
Main subject category:
Health Sciences
Keywords:
Quantile regression, Time to event data , Meta- analysis, Censored data, Cox model
File:
File access is restricted only to the intranet of UoA.
Konstantina Konidari-master.pdf
2 MB
File access is restricted only to the intranet of UoA.
