Chen, Shangzhen
(2019)
Regression modeling for competing risks based on pseudo-observations, with application to breast cancer study.
Master's Thesis, University of Pittsburgh.
(Unpublished)
Abstract
In medical research, a patient might experience a failure due to different causes, where each cause would be considered a competing risk. The standard method to analyze the type of data that is used most frequently in medical research is the cause-specific hazard regression analysis based on the Cox proportional hazards model or the cumulative incidence function-based method. The Cox proportional hazards model is used to model the cause-specific event rates and treat the other type of events as independent censoring. The new methods which have been proposed directly assess the effect of covariates on the cumulative incidence functions, which is the pseudo-observations approach proposed by Andersen and Klein in 2007. The scheme of pseudo-observations approach is to (1) choose some fixed time points that are equally spaced on the event scale, (2) calculate pseudo-observations for each individual at those fixed time points using the jackknife technique, and (3) fit a generalized linear model with GEE method based on the conditional cumulative incidence function using the pseudo-observations.
We applied the pseudo-observation approach to a breast cancer study. The goal of the study was to assess the effect of the covariates on the cumulative incidence function. The results showed that nodal status and tumor size are positively related to cumulative incidence of death following breast cancer recurrence and that age has a negative relationship with the cumulative incidence of death following breast cancer recurrence. In addition, nodal status and tumor size are not significantly associated with death due to causes other than breast cancer. Age is positively related to death not due to breast cancer.
Public health significance: Because the method explored and applied here is a readily accessible procedure for censored time-to-event data, providing straightforward interpretation of the effect of the predictors on the cumulative incidence function, dissemination of the method through a real world example would help the researchers, stakeholders,and patients in public health understand the usefulness of the method and grasp the interpretation of the analysis for competing risks data.
Share
Citation/Export: |
|
Social Networking: |
|
Details
Item Type: |
University of Pittsburgh ETD
|
Status: |
Unpublished |
Creators/Authors: |
|
ETD Committee: |
|
Date: |
16 July 2019 |
Date Type: |
Publication |
Defense Date: |
5 June 2019 |
Approval Date: |
16 July 2019 |
Submission Date: |
3 June 2019 |
Access Restriction: |
No restriction; Release the ETD for access worldwide immediately. |
Number of Pages: |
46 |
Institution: |
University of Pittsburgh |
Schools and Programs: |
School of Public Health > Biostatistics |
Degree: |
MS - Master of Science |
Thesis Type: |
Master's Thesis |
Refereed: |
Yes |
Uncontrolled Keywords: |
pseudo-observations, generalized linear model with GEE |
Date Deposited: |
16 Jul 2019 17:20 |
Last Modified: |
16 Jul 2019 17:20 |
URI: |
http://d-scholarship.pitt.edu/id/eprint/36864 |
Metrics
Monthly Views for the past 3 years
Plum Analytics
Actions (login required)
![View Item View Item](/style/images/action_view.png) |
View Item |