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

A Comparison of Regression Methods in Data Subject to Detection Limits: An Application to Lung Fiber Analysis Among Brake Workers

Liu, Yimeng (2013) A Comparison of Regression Methods in Data Subject to Detection Limits: An Application to Lung Fiber Analysis Among Brake Workers. Master's Thesis, University of Pittsburgh. (Unpublished)

[img]
Preview
PDF
Primary Text

Download (684kB) | Preview

Abstract

Objective: This thesis aims to apply and compare selected regression methods with a lung fiber analysis dataset. Final results based on 19 cases will be compared to 2011 Marsh et al.’s analysis based on the first 15 cases.
Methods: Two research questions for the lung fiber dataset are: (1) is there a relationship between the lung fiber concentration of TAA and lung fiber concentration of AC? and (2) is there a relationship between the lung fiber concentration of TAA and duration of employment as a brake worker? Besides the substitution method, bivariate normal regression was used in the doubly left-censored situation in question 1, while the censored normal regression and regression modeling with count data were used in the situation with only the dependent variable subject to detection limits in question 2.
Result: (1) The estimate of the slopes between the log-scale of two lung concentrations (TAA vs AC) were 0.59, 0.57, 0.59 and 0.54 in the simple linear regression with substitution (DL, 0.5DL, DL/√2) and the bivariate normal regression, respectively. All of the slope estimates were statistically significant different from zero (p-value = 0.001, 0.003, 0.002 and 0.003). (2) The estimate of the slopes between the log-scale of the TAA lung fiber concentrations and DOE were 0.001, 0.014, 0.008, 0.020 and 0.030 in the simple linear regression with substitution (DL, 0.5DL, and DL/√2), censored normal regression and the negative binomial regression, respectively. All of the slope estimates were not statistically significant different from zero (p-value = 0.933, 0.486, 0.675, 0.390 and 0.439).
Conclusions: The consistent results from the substitution and other methods provide support for both a positive relationship between the lung concentration of TAA and AC and for no relationship between the lung concentration of TAA and DOE among 19 brake workers with mesothelioma. These findings are consistent with Marsh et al.’s findings in 2011 based on the first 15 cases. The public health significance is that the study results provide additional support for the conclusion that exposure to non-commercial amphibole asbestos, and not chrysotile, is related to the observed mesothelioma in brake workers. However, these conclusions need to be verified with a larger sample size.


Share

Citation/Export:
Social Networking:
Share |

Details

Item Type: University of Pittsburgh ETD
Status: Unpublished
Creators/Authors:
CreatorsEmailPitt UsernameORCID
Liu, Yimengyil103@pitt.eduYIL103
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Thesis AdvisorMarsh, Gary M.gmarsh@pitt.eduGMARSH
Committee MemberYork, Ada O.ayouk@pitt.eduAYOUK
Committee MemberTang, Gonggot1@pitt.eduGOT1
Committee MemberDing, Yingyingding@pitt.eduYINGDING
Committee MemberSharma, Ravi K.rks1946@pitt.eduRKS1946
Date: 27 September 2013
Date Type: Publication
Defense Date: 29 August 2013
Approval Date: 27 September 2013
Submission Date: 20 September 2013
Access Restriction: 5 year -- Restrict access to University of Pittsburgh for a period of 5 years.
Number of Pages: 66
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: detection limit; non-detect; left-censoring; regression
Date Deposited: 27 Sep 2013 14:37
Last Modified: 01 Sep 2018 05:15
URI: http://d-scholarship.pitt.edu/id/eprint/19810

Metrics

Monthly Views for the past 3 years

Plum Analytics


Actions (login required)

View Item View Item