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

The effects of misspecification of submodels in joint modeling of repeated measures and time-to-event outcomes

Mao, Jason (2020) The effects of misspecification of submodels in joint modeling of repeated measures and time-to-event outcomes. Master's Thesis, University of Pittsburgh. (Unpublished)

[img]
Preview
PDF
Submitted Version

Download (597kB) | Preview

Abstract

Glioblastoma (GBM) is the most common form of primary brain tumor in the US. It is highly aggressive and has a median survival rate of 12 to 14 months with treatment. It has significant effects on a patient’s neurocognitive functions, so consequently, patient reported outcomes (PROs) are useful for evaluating patients’ physical and mental state in a way that biomarkers cannot fully capture.
Joint models, commonly used in biomedical research, combine traditional mixed models and survival analysis models, associating the longitudinal outcome with the time-to-event outcome. These models improve inferences on both types of outcomes by accounting for their underlying relationship, where events times are associated with the longitudinal outcomes.
Using data from a net-clinical benefits (NCB) sub-study of RTOG 0825, which evaluated the effects Bevacizumab on newly diagnosed GBM patients, we fit joint models to longitudinal PRO measures of symptom severity and interference with daily life and time-to-event data of GBM progression-free survival. We use these scenarios to simulate joint models where we misspecify the underlying survival and longitudinal submodels to investigate the effect of model misspecification on the association parameter that ties together the submodels.
We found that estimates of the association parameter are relatively robust to misspecification of the underlying survival distribution but not to misspecification of the assumed trajectory of the longitudinal submodel. Individual simulations were prone to extremely biased estimates, unstable estimates, and programming errors, so further investigation is suggested.
Public Health Significance: Limited research has been done regarding the impact of misspecifying joint models. This thesis can inform methods to improve the analysis of biomarker and time-to-event data. These models, in turn, would have a public health impact when biomarkers can be used as surrogates for intervention in major health related events and thus facilitate early intervention of those events when necessary. Here, we illustrate an example confirming a result from RTOG 0825 that Bevacizumab has a negative effect on PROs in addition to investigating the association of these PROs on GBM progression.


Share

Citation/Export:
Social Networking:
Share |

Details

Item Type: University of Pittsburgh ETD
Status: Unpublished
Creators/Authors:
CreatorsEmailPitt UsernameORCID
Mao, Jasonjmm411@pitt.edujmm411
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Committee ChairAnderson, Stewartsja@pitt.edu
Committee MemberYouk, Adaayouk@pitt.edu
Committee MemberMiller, Rachelmillerr@edc.pitt.edu
Date: 29 January 2020
Date Type: Publication
Defense Date: 2 December 2019
Approval Date: 29 January 2020
Submission Date: 19 November 2019
Access Restriction: No restriction; Release the ETD for access worldwide immediately.
Number of Pages: 43
Institution: University of Pittsburgh
Schools and Programs: Graduate School of Public Health > Biostatistics
Degree: MS - Master of Science
Thesis Type: Master's Thesis
Refereed: Yes
Uncontrolled Keywords: joint models, misspecification
Date Deposited: 29 Jan 2020 20:27
Last Modified: 29 Jan 2020 20:27
URI: http://d-scholarship.pitt.edu/id/eprint/37827

Metrics

Monthly Views for the past 3 years

Plum Analytics


Actions (login required)

View Item View Item