Interpreting sampling data from Gulf Coast to assess workers' exposure

Both OSHA and BP have set up webpages that offer their air monitoring procedures and results related to the Gulf oil disaster. Several worker health and safety experts have examined the data and offered interpretations of the results, including Eileen Senn, a former OSHA inspector. She reviewed BP's June 9, 2010 Personal Exposure Monitoring Results Summary, which includes a graphical analysis of 2,100 personal samples collected to-date for benzene, total hydrocarbons, and 2-butoxy ethanol, an ingredient in the dispersant Corexit EC9527A. Senn notes:

"The document provides no information on sampling and analytical methods, detection limits, the locations the samples were taken, the tasks workers were performing, the duration or time of sampling, or the raw data behind the graphs. BP reports 2,324 of 2,843 (82%) results of "none detected."

They also report 519 samples with detectable levels of the following contaminants: benzene, total hydrocarbons,and 2-butoxy ethanol. Senn notes:

"The BP data shows that the worst exposures are taking place offshore. This is to be expected since that is where the crude oil is leaking and releasing volatile components, oil is burning, and where dispersants are being sprayed. However, the data also shows significant nearshore exposures. And on the beach, the number of samples showing 2-butoxy ethanol levels is especially alarming since this is where the majority of clean-up workers are working."

BP explains the data this way:

"The air sampling data indicates that the exposures of these workers to airborne concentrations of crude oil or dispersant chemicals of interest are below the occupational exposure limits."

Eileen Senn, MS correctly takes exception to BP's characterization of the data. She writes:

"This is not technically true since there is no OSHA PEL for Total Hydrocarbons. Using the outdated OSHA Permissible Exposure Limits (PELs) for benzene (1 ppm) and 2-butoxy ethanol (50 ppm) are not protective of workers. Instead BP should be using NIOSH Recommended Exposure Levels (RELs) of 0.1 ppm for benzene and 5 ppm for 2-butoxy ethanol or other more protective Occupational Exposure Limits (OELs). The criteria of 100 ppm BP has chosen for Total Hydrocarbons is extremely high, especially given we don't know the constituent chemicals in this mixture."

The data on OSHA's website indicates their industrial hygienists have collected air and bulk samples on six days between May 24-June 5. The samples were collected near workers doing beach cleanup and during nearshore activities, such as boat deck hands tying up boom and monitoring in-situ fires. (OSHA reports that it only has jurisdiction up to 3 miles offshore, they have not taken offshore samples.) In their June 10, 2010 OSHA Laboratory Analysis Results by Site and OSHA Direct Reading Results by Site OSHA describes their involvement as follows:

"OSHA industrial hygienists are monitoring worker exposure and looking for health effects from chemical exposure to oil, weathered oil, oil dispersants, cleaning agents and other materials as identified. When evaluating worker exposure, direct reading instruments and shift-long sampling is performed in the actual breathing zone of workers. OSHA has also sampled areas that are periodically frequented by workers but the samples are not taken from the breathing zone of workers. Finally, OSHA has sampled directly over tar balls, inside bags containing contaminated materials, and other locations that do not represent worker exposure but provide information on the types of chemicals that may be coming off contaminated materials."

OSHA provides more detail for its sampling results than BP, and OSHA offers several occupational exposure limit (OEL) values (e.g., NIOSH recommended exposure limits, ACGIH TLVs) to allow readers to compare the levels identified in the disaster zone to levels established by expert groups. However, almost all OSHA's results listed as "ND," the acronym used by OSHA's lab to mean "Non-detected: Materials sampled were not found at levels high enough to positively identify." It would be helpful if OSHA amended its webpage with information clarifying these ND's, in particular, if the analyte was not detected during laboratory analyses, the limit of detection should be listed. As one industrial hygienist informed me, NDs can exceed occupational exposure limits (OEL) depending on what OEL an employer or regulatory authority is using for decision-making. That makes more complete explanations of what ND really means (for each analyte) even more important.

Here's just a sample of what you see on the OSHA's oil spill response monitoring pages:

On May 24, 2010: Volatile organic compounds (VOCs) of 19 ppm and 2 ppm on a vessel during and after use of a citrus cleaner to clean decks.
June 2, 2010: levels of carbon monoxide up to 84 ppm on an ignitor boat placing flares in fire booms, and a photo ionization detector (PID) direct reading of 1.0 to 1.5 (but OSHA reports no units or what the result might represent.)

Eileen Senn, MS offers her hypotheses of what this data might tell us:

• The volatiles have all evaporated and there is no hazard on the beaches: Given the heat, this is at least part of the explanation.
• Bad planning: Maybe sampling isn't taking place in the right places at the right times.
• Not enough sampling: There are 25,000 workers and less than 2,500 air samples collected in the 52 days since April 20. It is safe to say that not even one worker's exposures to all the chemicals in the air have been fully characterized.
• Not sampling for the right things: What exactly is in that Total Hydrocarbons or VOC mixture? This benign-sounding label hides the fact that carcinogens, mutagens, and teratogens are no doubt included. Maybe there is exposure to ultrafine particles which can carry toxic components deep into the lungs. But OSHA took no samples to evaluate this.

OSHA notes that "The main concern along the Gulf coast and on the water would be salt spray interfering with the hydrogen chloride analysis. A large part of this could be mitigated through the use of a pre-filter which would capture any spray before the air enters the sampling tube." This footnote appears on OSHA's Direct Reading Results by Site "Sample results which are reported with a "<" in front of a value indicate that no reading or result was obtained at the limit of detection for the instrument or method being used." For many direct reading results, OSHA is reporting < 20 ppm for Total Petroleum Hydrocarbons and < .02 ppm for Benzene. This is too high except for screening.

As an expert in industrial hygiene (and its limitations) and a former OSHA inspector, Ms. Senn is known in our occupational health community for the mantra "Sampling is not the answer." She explains:

If the question is "What are workers breathing and how should they be protected?" sampling may not be the answer or at least not the whole answer. Substance by substance air sampling is not an adequate approach to assessing airborne exposures to complex mixtures of chemicals in the Gulf. This question should be answered by a team of experts with access to narrative exposure assessments would describe typical exposures for typical cleanup tasks being performed. Exposure assessments should be stories informed by numbers, not just numbers. The stories would include the chemicals and their hazards; typical activities performed by workers and the typical machinery, equipment, and tools used; activities that may cause chemicals to become airborne near employees; activities that may cause chemicals to contact employees' eyes, skin, hair, gloves, shoes, or clothing; and measures being taken to reduce exposures, including personal protective equipment.

OSHA's permissible exposure limits (PELs) perpetuate an impractical and flawed approach to controlling chemical exposures whereby OSHA cannot require engineering and administrative controls, respirators or medical surveillance controls unless the agency collects air samples that prove that employee exposures exceed a PEL. Opportunities to prevent worker exposures are being missed because of reliance on and misinterpretation of BP's air sampling data as a "Clean Bill of Health." The prime example of this is OSHA basing their decision not to require respirators on their own and BP's sampling results. OSHA has not vigorously rejected BP's reliance on PELs in interpreting their data.

In addition, OSHA is using many resources to conduct sampling that might be better used to visually assess worker's exposures and recommend control measures. Meanwhile, workers and community members are sick and can smell chemicals in the air. EPA describes three types of odors being reported and implies they are not harmful. But for at least two chemicals, benzene and ethyl benzene, the odor threshold is well above the hazardous level. OSHA has created an OSHA Personal Protective Equipment Sampling Matrix, with recommended personal protective equipment (PPE) for sixteen job tasks. Why not extend this sensible approach to worker protection to respiratory protective equipment?

Many of those who responded to the 1989 Exxon Valdez spill are now ill. Are we about to repeat that awful history due to air sampling? OSHA needs to change its approach. And soon.