When the Deepwater Horizon drilling rig exploded and sank in the Gulf of Mexico in 2010, more than 50,000 clean-up workers responded to the disaster. Many were exposed to chemicals released into the air from crude oil, burning of the oil, and dispersants used to break up the oil. Until recently, estimating their exposure has been difficult, mainly because the air samplers worn by many workers lacked the sensitivity to detect some chemicals and because measurements for particular chemicals were not available for all groups of workers. To address this knowledge gap, researchers at the University of Minnesota School of Public Health have developed a new statistical method for calculating the workers’ chemical exposures.
A study of the new method was published in the Annals of Work Exposures and Health. The work is a component of the GuLF STUDY being conducted by the National Institute of Environmental Health Sciences investigating potential adverse health effects among the oil spill response and clean-up workers.
“It’s really important for us to know what the exposures were for workers who carried out specific jobs or tasks during the response and clean-up effort,” says lead author and biostatistics doctoral student Carrie Groth. “Our method allows us to estimate exposures in scenarios where our measurements indicated that the exposures were below the lowest level the laboratory could determine (known as a limit of detection (LOD)) but where we have measurements on related chemicals that can be used to inform the exposure of the chemical of interest. Furthermore, we can also predict exposures when measurements for particular chemicals are unavailable for some workers.”
During the Deepwater Horizon disaster, the incident area spanned several states — including 1,100 miles of coastline — so it wasn’t possible to monitor all workers or environments and every kind of chemical that the workers may have encountered.
In many instances, however, pieces of the exposure puzzle are known, such as measurements for some chemicals and the presence of similar chemicals. This model uses linear relationships between two chemicals among some workers to better inform exposure to the chemical of interest while accounting for measurements below the LOD. This relationship can then be used to predict exposure for cases where chemical measurements were not taken for other workers.
The method is particularly valuable for estimating “censored chemicals,” which are the chemicals that were likely present, but at levels too low to be detected by the laboratory method (below the LOD). Quantifying censored chemicals is important because low levels could still have as yet unknown or synergistic health effects.
“Environmental health scientists and industrial hygienists can use this new approach because it allows them to work with multiple-chemical scenarios in a new way to more accurately account for levels that are too low for detection with current technology,” says Groth.
As part of the GuLF STUDY, Groth and her co-researchers now plan to use the method to provide exposure estimates for workers enrolled in the cohort study whose exposures varied based on the time period that they were involved in the response and clean-up effort, the geographical location of their work, their tasks, and other work characteristics.