|
" Using survival analysis to investigate breast cancer in the Kurdistan region of Iraq "
Raza, Mahdi
| Document Type
|
:
|
Latin Dissertation
|
| Record Number
|
:
|
832270
|
| Doc. No
|
:
|
TLets699419
|
| Main Entry
|
:
|
Raza, Mahdi
|
| Title & Author
|
:
|
Using survival analysis to investigate breast cancer in the Kurdistan region of Iraq\ Raza, Mahdi
|
| College
|
:
|
City, University of London
|
| Date
|
:
|
2016
|
| student score
|
:
|
2016
|
| Degree
|
:
|
Thesis (Ph.D.)
|
| Abstract
|
:
|
The objective of this thesis is to carry out a survival analysis for patients with breast cancer. Using data from the Nanakaly and Hewa hospitals in the cities of Erbil and Suleimaniah, respectively, cases where there is hidden censoring on survival time were investigated. The aim of this study was to identify the main risk factors and quantify the overall risk for breast cancer. We developed a new Markov chain-based method for generating survival curves and hazard functions. In particular we adjusted the Kaplan Meier analysis to find a survival curve with hidden censoring of the data, and also estimated a survival function from the biased one obtained directly from the data by generating new models in two cases; with and without censoring. To ensure the validity of the suggested model we considered different simulation techniques applied to the Nanakaly data. Because of the availability of a good survival function, we chose to work with a German data set. As a result we conclude that our model performs well in many circumstances, and its predictions, even when less accurate, are always an improvement on considering the apparent survival curves from the unadjusted data. For the data from Nanakaly hospital, the only variable we had to consider was age at diagnosis and the survival results showed that this was a significant variable. With far more detailed reports available for Hewa hospital, we were able to identify estrogen abundance, smoking habits and tumour grade, as having a statistically significant impact on the incidence of breast cancer. On the other hand, when analysing the Nanakaly and Hewa data for comparison with German data, in all three cases the survival curve is greater among younger patients. The suggested models may be verified using cross validation or by using new data. Finally, we note that it would be preferable to have accurate data to applying our methods to imperfect data. Therefore we established both a general and a specific flowchart to collect the data in the future. Encouraging the implementation of the recommended procedures might serve to obtain the data needed to develop a more comprehensive understanding of breast cancer in Kurdistan.
|
| Subject
|
:
|
QA Mathematics
|
| Added Entry
|
:
|
City, University of London
|
| |