“We’re not stupid”: OSHA Silica Hearings Week #1

Yep. “We’re not stupid” was just one of the many memorable moments at last week’s public hearing on OSHA’s proposed rule on respirable crystalline silica. The remark came from epidemiologist Robert Park of CDC’s National Institute for Occupational Safety and Health (NIOSH).  He was compelled to respond to a comment made by Tony Cox, a consultant retained by the American Chemistry Council.

Cox, who was expounding on the OSHA’s peer-reviewed risk assessment, asserted that the agency has not demonstrated a causal link between silica and lung disease at OSHA’s proposed permissible exposure limit of 50 ug/m3.  The industry consultant’s opinion is grossly out-of-step with the abundance of literature concluding otherwise. It also ignores the history of silica-related disease dating back centuries involving silica-exposed workers. That history was more than enough for NIOSH to recommend to OSHA in 1974 that it adopt the 50 ug/m3 exposure limit.  Forty years later, when OSHA is proposing that exact 50 ug/m3 limit, the evidence is stronger than ever.

Many in the audience and the OSHA staff likely expected some of the witnesses to make some fringe assertions about the science, but that one about causation was laughable.

Other opponents of the rule have provided a litany of uncertainties that they insist OSHA must assess before moving forward with its rule. Just on the subject of exposure monitoring and analysis, the long list of potential biases offered by them include:

cyclone particle distribution, dust weighings, balance calibration, particle size in calibration standards, linear range of calibration standards, low limit of quantitation, calculations, errors in pump calibration, errors in cyclone calibration, cyclone leakage, filter cassette leakage, improper flow rates, cyclone particle distribution, cyclone contamination, and contamination during shipping.

The long litany of criticisms about the scientific basis for the rule have—and will continue— to be made throughout the rulemaking.  These criticism will persist years down the road when the rule is challenged in court. This is just part of the typical life cycle for developing a new OSHA standard.

A bright spot for some at last week’s hearing, but a frustration for others, was the testimony of the National Industrial Sand Association (NISA). The trade association represents 31 firms that mine and process 100 percent pure silica. NISA representatives testified primary about their 30-year old Silicosis Prevention Program (SPP). They credit the SPP with preventing the disabling and deadly disease silicosis among workers of the member companies that have implemented the program. (NISA is in the midst of an epidemiological study to determine whether that claim is true.  Study protocol reviewed by NIOSH. Result expected in late 2015.)

When NISA kicked off its testimony saying this:

“NISA members strongly believe that the exposure monitoring and medical surveillance that they implement …must be included in a comprehensive OSHA standard addressing crystalline silica.”

Opponents of the rule sunk into their seats, and were dreading what NISA would say next.

Mr. Andrew O’Brien of Unimin Corp—the largest producer of industrial sand in the world and a firm with 2,400 employees—asserted that the costs of conducting exposure monitoring and medical surveillance are “quite minimal.”  [Cringing in the audience by opponents of the OSHA proposal.]

O’Brien added:

“Our experience shows that companies don’t need to hire outside contractors to competently conduct exposure monitoring. Company staff can be trained to do it competently.” [Groaning in the audience by the opponents.]

In NISA’s written comments to OSHA, the group estimates the cost to conduct exposure monitoring is between $106 and $127 per sample. That’s significantly more than OSHA’s projection. Moreover, for construction employers who want to avoid doing exposure monitoring at all, OSHA is offering something easy. For 11 of the dustiest tasks in construction, such as using a masonry saw or drywall sanding, if the employer uses a prescribed set of simple controls, they will be deemed to be in compliance with the rule. They would incur no cost for exposure monitoring.

The NISA endorsement of exposure monitoring and medical surveillance caused some positive head nodding in the audience from proponents of the rule. Their reaction was not surprising given what they’d heard the day before from witnesses retained by the Chamber of Commerce. But some proponents of a rule were not ready to concur with all of NISA’s pitch to OSHA.  The trade association urges OSHA to maintain the current PEL of 100 ug/m3, but adopt the other provisions of its proposal. That is one of the alternatives offered by OSHA in its proposal.

This week’s line-up of witnesses includes those representing the Building and Construction Trades Department, AFL-CIO, the United Autoworkers, the National Association of Manufacturers, the American Chemistry Council (again) and the National Asphalt Pavement Association. On Thursday, a small delegation of non-union workers from Houston and New Jersey will describe their experiences working in silica dust.  They will be a reality check for those who are paid to manufacture uncertainty when it comes to health protective regulations.


  1. #1 John Newquist
    March 26, 2014

    Thanks for the notes and analysis.

  2. #2 Kevin
    March 27, 2014

    Out of curiosity, what is the exposure when participating in a day at the beach? Including the building of sandcastles?

    On a more serious note, has a distinction been made between amorphous and crystaline silica?

  3. #3 Celeste Monforton
    March 27, 2014

    Yes, witnesses have discussed the difference between crystalline and amorphous silica. Playing in sand at the beach will generate negligible exposure to respirable silica. The sand used to build those fun (and sometimes elaborate) structures are created using wet sand. Regrettably, some opponents of improved protections for workers who are exposed to respirable silica, use the beach sand example to minimize the risk.

  4. #4 Frank Hearl
    Washington, D.C.
    March 31, 2014

    Kevin, aside from being wet (i.e., not able to become airborne) as Celeste notes, the grain size of beach sand is large. You wouldn’t find “respirable” size dust to inhale. It is the respirable size particles that cause silica-related lung diseases. Respirable dust has a mass median aerodynamic diameter of 4 micrometers, is mostly smaller than 10 micrometers and can penetrate to the alveolar regions of the lung.

  5. #5 Carol Rice
    University of Cincinnati
    March 31, 2014

    Applause to the continued pro-active work of Unimin health and safety personnel, speaking for the industry. Unimin engineers have developed innovative engineering controls to limit silica exposure in processing and during particulate removal from workers’ clothing. These advances demonstrated that innovation in health and safety is good business–for workers and business.

  6. Carol,
    Also was great to hear from you. Thanks for weighing in on Unimin’s OHS program.

  7. #7 Tony Cox
    April 1, 2014

    Celeste writes, “Cox, who was expounding on the OSHA’s peer-reviewed risk assessment, asserted that the agency has not demonstrated a causal link between silica and lung disease at OSHA’s proposed permissible exposure limit of 50 ug/m3. The industry consultant’s opinion is grossly out-of-step with the abundance of literature concluding otherwise.” Please cite a specific study from the “abundance of literature” that, in your opinion, demonstrates the claimed causal relation using facts and data.

  8. Dr. Cox,
    An early synopsis of the literature was published by NIOSH in 2002. An update of that literature appears in OSHA’s peer reviewed risk assessment. Dr. Kyle Steenland’s testimony at the OSHA public hearing provided additional more recent research.

  9. #9 Tony Cox
    April 3, 2014

    Thank you. However, none of these is a study demonstrating causation. My question remains: Please name a single study that, in your opinion, does demonstrate causation using appropriate scientific and statistical methods of data analysis.

  10. #10 Lew Crist
    Chambersburg, PA
    April 8, 2014

    Unfortunately the people making the decisions don’t understand the big picture. There are industries such as the one I am in (foundry) that are going to be a casualty of this rule. The big sand company like Unimin can pass along price increases to pay for all the testing, etc. We are not able to do this since our Chinese competitors do not have to follow the same rules. The same people that are making these regs need to spend some time with buyers in this country educating them on what buying Chinese does to our environment. Work passing laws that keep workers employed in the country instead of regs that send work overseas only to pollute our grand children later.

  11. #11 Mary
    April 9, 2014

    @Tony Cox: If one thinks that a “single study” (of an exposure that it would be unethical to use in a randomized trial) can demonstrate causation, and (furthermore) that the lack of the existence of such a study is grounds for asserting no causation, then one’s training in causal inference is sorely inadequate.

  12. #12 Tony Cox
    April 10, 2014

    Mary, I might agree with you if epidemiological studies of causation required unethical experimentation. But they do not. Instead, valid causal inferences in epidemiology can be based on observational (non-experimental) data. Well-developed, rigorous methods for testing hypotheses about potential causation include conditional independence tests, panel data analysis, Granger causality tests, quasi-experimental designs and analysis, intervention analysis, change-point analysis, counterfactual and potential outcomes models, marginal structural models, and causal network models of change propagation. (For more on these, see e.g., http://www.ncbi.nlm.nih.gov/pubmed/23718912). These methods are used by epidemiologists to test whether causal hypotheses — such as that reductions in exposures preceded and help to predict and explain reductions in diseases — are consistent with data. They provide a more objective alternative to causal conclusions based on personal opinion or ideological beliefs.

    When Celeste writes that my opinion (that these methods have not been applied to produce solid, objective statistical evidence of causation at relevant exposure levels, but instead only unreliable subjective judgments have been used) “is grossly out-of-step with the abundance of literature concluding otherwise,” I think it appropriate to ask what papers, specifically, are included in this “abundance of literature.” I have been unable to find even one. I would be glad to be corrected if, in fact, there are any papers that do show evidence of causation, using any objective methods.