Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form

Parametric Regression Analysis and Discrimination Analysis of Competing Risks Data

Shi, Haiwen (2013) Parametric Regression Analysis and Discrimination Analysis of Competing Risks Data. Doctoral Dissertation, University of Pittsburgh. (Unpublished)

Primary Text

Download (1MB) | Preview


This thesis contains two parts focusing on regression analysis and diagnostic accuracy analysis of competing risks data. In the first part, we propose a parametric regression model for the cumulative incidence functions (CIFs) commonly used for competing risks data. The model adopts a modified logistic model as the baseline CIF and a generalized odds-rate model for covariate effects, and it explicitly takes into account the constraint that a subject with any given prognostic factors should eventually fail from one of the causes such that the asymptotes of the CIFs should add up to one. We hence model the CIF from the primary cause assuming the generalized odds-rate transformation and the modified logistic function as the baseline CIF. Under the additivity constraint, the covariate effects on the competing cause are modeled by a function of the asymptote of the baseline distribution and the covariate effects on the primary cause. The inference procedure is straightforward by using standard maximum likelihood theory. We demonstrate desirable finite-sample performance of our model by simulation studies in comparison with existing methods. Its practical utility is illustrated in an analysis of a breast cancer data set to assess the treatment effect of tamoxifen on breast cancer recurrence that is subject to dependent censoring by second primary cancers and deaths.
Diagnostic accuracy studies progressed in the past decade to involve complicated survival outcomes beyond the traditional dichotomous outcome. Another recent advance in diagnostic medicine is the appearance of novel measures for accuracy improvement due to the addition of new markers. In the second part of this thesis, we intend to integrate these two evolving areas and contribute a discussion on assessing accuracy improvement for censored survival outcomes. Furthermore, we consider competing-risk censoring in addition to the usual independent censoring and provide statistical procedures with inference details. In particular, we consider fitting regression models based on the CIF for the primary event. Parallel estimators are proposed using inverse probability weighting or based on the bivariate CIF. Both estimators perform very well in simulation studies and in an application to the second breast cancer study.


Social Networking:
Share |


Item Type: University of Pittsburgh ETD
Status: Unpublished
CreatorsEmailPitt UsernameORCID
Shi, Haiwenhas9@pitt.eduHAS9
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Committee ChairCheng, Yuyucheng@pitt.eduYUCHENG
Committee MemberSampson, Allanasampson@pitt.eduASAMPSON
Committee MemberGleser, Leongleser@pitt.eduGLESER
Committee MemberJeong, Jong-Hyeonjeong@nsabp.pitt.eduJJEONG
Date: 30 January 2013
Date Type: Publication
Defense Date: 12 November 2012
Approval Date: 30 January 2013
Submission Date: 1 November 2012
Access Restriction: 1 year -- Restrict access to University of Pittsburgh for a period of 1 year.
Number of Pages: 64
Institution: University of Pittsburgh
Schools and Programs: Dietrich School of Arts and Sciences > Statistics
Degree: PhD - Doctor of Philosophy
Thesis Type: Doctoral Dissertation
Refereed: Yes
Uncontrolled Keywords: Area under the receiver operating characteristic curve, Cause-specific hazard function, Competing-risk censoring, Cumulative incidence function, Diagnostic and prognostic accuracy improvement, Integrated discrimination improvement, Long-term incidence, Modified three-parameter logistic model, Net reclassification improvement, Parametric modeling.
Date Deposited: 30 Jan 2013 16:55
Last Modified: 15 Nov 2016 14:06


Monthly Views for the past 3 years

Plum Analytics

Actions (login required)

View Item View Item