After the <italic>Deepwater Horizon</italic> oil release in the Gulf of Mexico, the National Institute of Environmental Health Sciences initiated an epidemiological study (the GuLF STUDY) to investigate the potential adverse health effects associated with the oil spill response and clean-up work. The exposure assessment is a critical component of the GuLF STUDY because it allows the investigation of the exposure-disease relationship. This involves the analysis of thousands of personal inhalation monitoring measurements that were collected by BP and its contractors during the entire remediation effort. A substantial portion of these data, however, has values below the limits of detection (LOD). This dissertation investigates various statistical methods for handling data with detection limits and presents the methodology and assessment of the inhalation exposures for workers on the four main rig vessels that were responsible for stopping the leak.The first section of this dissertation evaluates three established classical (or `frequentist') methods for analyzing data with censored observations to estimate the arithmetic mean (AM), geometric mean (GM), geometric standard deviation (GSD), and the 95th percentile (X0.95) of the exposure distribution: the Maximum Likelihood (ML) Estimation, the &beta-substitution, and the Kaplan-Meier (K-M) methods. Each method was challenged with computer-generated exposure datasets drawn from lognormal and mixed lognormal distributions with sample sizes (N) varying from 5 to 100, GSDs ranging from 2 to 5, and censoring levels ranging from 10% to 90%, with single and multiple LODs. Using relative bias and relative root mean squared error (rMSE) as the evaluation metrics, the &beta-substitution method was found to generally perform as well or better than the ML and K-M methods in most simulated conditions. The ML method was suitable for large sample sizes (N &ge 30) up to 80% censoring for lognormal distributions with small variability (GSD=2-3). The K-M method generally provided accurate estimates of the AM when the censoring was <50% for lognormal and mixed distributions. The second section describes a Bayesian framework for analyzing censored data. Similar computer simulation was conducted to compare the &beta-substitution method with a Bayesian method. The Bayesian method using non-informative priors and the &beta-substitution method were generally comparable in bias and rMSE when estimating the AM and GM. For the GSD and the 95th percentile, the Bayesian method with non-informative priors was more biased, and had a higher rMSE than the &beta-substitution method but the use of more informative priors generally improved the Bayesian method's performance, making both the bias and the rMSE more comparable to the &beta-substitution method. The advantage of the Bayesian method is that it allows the use of prior information and also provides estimates of uncertainty for all parameters (GM, GSD, and 95th percentile) whereas the &beta-substitution method only provides estimates of uncertainty for the AM. The third chapter presents a methodology for assessing the occupational exposures and estimates of inhalation exposures for workers on the four main rig vessels (Enterprise, DD2, DD3, and Q4000) that were responsible for stopping the leak in the hot zone closest to the well site. Exposure groups (EGs) were created on based on chemicals, locations, vessels, time periods, and job titles/tasks. Bayesian method were used to analyzed exposures for total hydrocarbons (THCs), benzene, toluene, ethylbenzene, xylene (BTEX chemicals) and hexane. THC measurements were least censored compared other chemicals evaluated. THC exposures changed over time and varied by vessels and exposure groups. Highest exposures were generally observed in the time period before the well was successfully top capped. Exposures gradually decreased over time after top capping in most exposure groups except a few that might be involved in the decontamination effort. BTEX chemicals and hexane exposures were substantially lower than THC. The variability of the EGs for the GuLF STUDY were generally high, reflecting the non-routine, time-dependent nature of spill response efforts as well as the challenges of retrospectively constructing exposures for oil spill study.
University of Minnesota Ph.D. dissertation. July 2014. Major:Environmental Health. Advisor: Gurumurthy Ramachandran, Sudipto Banerjee, Ph.D.,. 1 computer file (PDF); xi, 142 pages, appendix p. 131-142.
Huynh, Tran Bich.
An assessment of occupational inhalation exposures to volatile oil components on four rig vessels for the GuLF STUDY.
Retrieved from the University of Minnesota Digital Conservancy,
Content distributed via the University of Minnesota's Digital Conservancy may be subject to additional license and use restrictions applied by the depositor.