GHS Adoption in the Face of Daubert

Unfortunately, wholesale adoption of GHS hazard classification system principles will disrupt various chemical hazard assessment programs and heighten problems manufacturers and importers will face.

• By James M. Hofert, Daniel W. McGrath, Frederick J. Ufkes
• May 01, 2014

On March 26, 2012, OSHA adopted the United Nations Globally Harmonized System (GHS), an international approach to hazard communication. GHS was negotiated through a multi-year process by hazard communication experts from many different countries, international organizations, and stakeholder groups. While GHS is based on major existing hazard communication standards from around the world, including OSHA's hazard communications standard (1994), GHS employs separate and distinct criteria and methodology for hazard classification and categorization of chemical substances.

OSHA, in adopting GHS, recognized that diverse and sometimes conflicting international requirements created confusion among workers and companies seeking to effectively use information contained in various hazard communication documents. OSHA noted that labels and safety data sheets for foreign-made products included unfamiliar symbols and hazard statements. OSHA further determined that given the differences in hazard classification criteria, labels on U.S. product exports and foreign product imports could be inconsistently or incorrectly interpreted once they cross the border. OSHA's expectation in the adoption of GHS appears to be that international acceptance of GHS will promote utilization of consistent information on labels and safety data sheets for various chemical products sold worldwide.

Unfortunately, in its wholesale adoption of GHS, OSHA may have overlooked the fact that this harmonization system has detailed criteria and methodologies for determination of acute and chronic adverse health effects from exposure to given chemical substances that may be inconsistent with judicial standards of causation promulgated by federal and state courts following Daubert v. Merrell Dow Pharmaceuticals, 509 U.S. 579 (1993). Standardization of format, order of information, label elements, signal words, precautions, and hazard statements is laudable, but the substance and choice of a given standardized statement/precaution will still depend on the hazard risk assessment based on tools and methodology incorporated in GHS. To the extent that this methodology is inconsistent with Daubert,¹ chemical manufacturers/importers may face increased regulatory and litigation risks in attempting to designate hazard classification and categorization for a given chemical substance. GHS may require manufacturers/importers to reference health hazards that would not otherwise be recognized under Daubert and its progeny.  Manufacturers/importers' reference to health hazards pursuant to GHS may constitute admissions where causation is otherwise disputed.

In Daubert the Supreme Court of the United States determined that federal courts must evaluate the validity of scientific evidence that purports to establish a credible case of causation between exposure to a given substance and an adverse health effect before its admission into evidence. Under Daubert, the scientific proposition a party seeks to admit into evidence must not only be generally accepted, but also be validated through accepted scientific methodology. Both the scientific conclusion and methodology by which causation is established must meet an accepted framework applied by the court as part of its "gatekeeper" responsibility.

The Federal Judicial Center has produced a "Reference Manual On Scientific Evidence" for use by federal courts. Many state courts have adopted the principles enunciated in Daubert in order to establish proof of general causation. Courts following Daubert generally refuse to accept causation evidence based on extrapolation, for example, in the absence of reproducible studies demonstrating an adverse human health effect from exposure to a given chemical substance.

EU Courts Have Rejected Daubert
The important dates for purposes of application of the Hazard Communication Standard (HCS) as amended by GHS, as it pertains to manufacturers'/importers' responsibilities, include June 15, 2015 (the date manufacturers must comply with GHS); Dec. 1, 2015 (the date that distributors shall not ship a product unless it contains a GHS label); and June 1, 2016 (the date for updating alternative workplace labeling and hazard communication).

GHS has been adopted as a component of the European Union's regulatory framework for regulation of chemical substances. The European Regulatory Framework that adopted GHS was updated in December 2006 (REACH).² REACH is an integrated system for registration, evaluation, authorization, and restriction of chemicals presently managed by the European Chemical Agency (ECHA).³ These regulations require manufacturers and importers to draw up a detailed technical dossier, including information on each chemical manufactured, its potential uses, and its intrinsic hazards based on relevant studies. Hazard classification is determined in many cases by the European Commission for the Environment working group composed of recognized experts.⁴

Courts in the European Union will not overturn hazard classifications and/or categorizations made by the Commission and approved by member states unless a complaining party demonstrates a manifest error of assessment or misuse of power. The European Union judicature confers on the Commission broad discretion in determining the hazard classification and categorization of a chemical substance and will not substitute its assessment of scientific and technical facts for that of the Commission.⁵

Hazard classification under GHS is determined on the basis of the total weight of evidence as evaluated by application of expert judgment. Responsible parties are to review all information bearing on the hazard classification, including but not limited to in vitro testing, animal studies, human epidemiological and clinical studies, case reports, extrapolations, implementation of the Read Across Method based on the similarity of chemical molecules, as well as data based on structure activity relationship models. Several of these methods, used to demonstrate general causation, are frequently rejected by U.S. courts following Daubert.⁶ Human epidemiological evidence is not necessary to establish a classification or categorization under GHS if other relevant evidence supports classification and/or categorization.⁷

Courts in the European Union have, in turn, rejected Daubert, noting that admissibility of expert opinion evidence in Europe is governed by §79 of their Evidence Act, which enunciates a different and more liberal standard for establishing causation than §702 of the Federal Rules of Evidence.

The adoption of GHS may result in what several authors have described as "classification shock",⁸ the realization that a given chemical substance is more hazardous than previously described.⁹

Adoption of GHS would appear to reflect a relaxed standard of "proof of causation," at least as it relates to the definition of type, degree, and severity of a hazard associated with a given chemical substance as it appears on labels and within data sheets, particularly in cases of risk of injury from chronic exposure. Daubert and its progeny generally reject use of case reports alone, as well as extrapolation and/or structural activity relationship models among other evidence as demonstrating proof of causation in the absence of human data.¹⁰

Appendix A to the Hazard Communication Standard, as amended, consistent with GHS, appears to adopt these various methods for hazard classification and categorization contained in GHS as valid evidentiary principles to be considered in the completion of a hazard analysis. Federal OSHA modified its hazard communication standard to conform to GHS to, in part, facilitate international trade of chemical products and promote consistency in the classification and labeling of chemicals internationally. However, standardized warning symbols, pictograms, and phrases are, to a large degree, dependent on hazard classification and categorizations that, in turn, are dependent on the definition of acceptable methodology and proof as it pertains to demonstration of a causal link between exposure to a substance and a given health effect.¹¹

Under GHS, a chemical substance will be classified based upon type, degree, and severity of the hazard (hazard class and category). The GHS standards (commonly referred to as the "Purple Book") establish agreed hazard classifications and communication provisions with explanatory information on how to apply the system. The definition of health and physical hazards provided in Appendix A and B to HCS, as amended, and the definitional paragraphs of 29 C.F.R. 1910.200(c) appear to be consistent with criteria provided in the GHS "Purple Book."

Appendix A to HCS provides a general approach to classification of chemical substances, including "bridging principles." Hazard classification includes concepts such as acute toxicity and "carcinogenicity," among others. Hazard categorization not only reflects the degree or severity of adverse health effects, but also, in many cases, is defined by the strength of evidence supporting the purported health effect categorization.

GHS suggests a tiered approach for mixtures. In assessing a mixture, the manufacturer or importer must use available test data for the mixture as a whole, use "bridging principles" to extrapolate from other data, and/or estimate hazard type and severity based upon known information relating to the individual ingredients of the mixture.

Under previous HCS standards, a manufacturer/importer would generally simply determine whether a given substance caused a health effect or not. (For example, a given substance either was or was not a carcinogen.) Under the new standard, classification of a chemical substance as a carcinogen is categorized under one of two categories, Category 1 having two subcategories. Categories relating to carcinogenicity, for example, are as follows: Category 1 – known or presumed human carcinogen; Category 1(a) – known carcinogen based upon human evidence; Category 1(b) – presumed to be a carcinogen based on animal data, extrapolation, and other evidence; Category 2 – suspected carcinogen based on the strength of the evidence and additional considerations. Similar categorizations are applied to other adverse health effects associated with chemical exposure, including but not limited to reproductive toxicity, mutagenicity, and target organ systemic toxicity.

The potential problems in using the GHS hazard classification system, in light of long-standing precedent reflected in Daubert and its progeny, are clear based on review of opinions written by European Union courts upholding hazard classifications and categorizations under GHS.

In Etimine SA v. Secretary of State for Work and Pensions, the European Union court ruled against manufacturers contesting the classification of boric acid as a Category II reproductive and developmental toxin. The court dismissed arguments relating to the lack of human data in upholding the classification. The court accepted results from validated structural activity relationships and expert judgment consistent with GHS principles. The court also accepted extrapolation from animal data as support for the Category 1 classification. The court rejected criticism of methodology cited by chemical manufacturers used to perform the animal testing. The court held that the route of administration used in carrying out animal testing was not a matter of legal assessment but of scientific opinion.¹²

In Nichol Institute v. Secretary of State for Work and Pensions, the European Union court rejected a manufacturer's challenge that nickel sulfate and nickel chloride should be categorized as Category I carcinogens. This classification was based upon existing data not including epidemiological studies. In addition, experts, in coming to the conclusion that nickel nitrate should be defined as a Category I carcinogen, applied the Read Across Method to nickel nitrate, taking the view that the degree of water solubility of nickel nitrate and its chemical components were sufficiently similar to nickel sulfate and nickel chloride to classify it as a Category I carcinogen.¹³ The court found that the methodology and conclusion of government experts were consistent with GHR and REACH and that the Read Across Method was widely recognized by the European scientific community. The court noted that REACH legislation, in Article XIII, recognizes the importance of the use of alternative methods to evaluate human toxicity secondary to exposure to chemical substances other than by animal testing and epidemiological studies.

Problems for U.S. Manufacturers and Importers
Under Daubert, experts are required by courts to look at the reliability of a study's findings, its design, and the sufficiency of the data before suggesting that a HazCom-triggering event has occurred. In addition to study design, other factors that should be assessed under Daubert and its progeny include statistical significance and exposure and dose/response parameters. Extrapolation from animal data in the absence of human experience is generally frowned upon and denoted as a questionable practice under Daubert. Federal and state courts in this country have recognized that differences in animal metabolization of a substance often prevent extrapolation of data to humans.

In light of longstanding legal precedent reflected by Daubert and its progeny, the wholesale adoption of GHS will create potential problems for manufacturers/importers both in terms of future regulatory actions and common law toxic tort claims. The adoption of internationally consistent and accepted signal words, pictograms, and hazard statements is laudable; unfortunately, wholesale adoption of GHS hazard classification system principles will disrupt various chemical hazard assessment programs and heighten problems manufacturers and importers will face, inappropriately describing the hazards and precautions to be taken in the use of a given product.

Manufacturers may well face situations where application of GHS principles as applied to a given chemical substance require a warning about a health effect which U.S. courts would not otherwise require, in the absence of credible evidence as defined by cases following Daubert that the product can actually cause the health effect. This places the manufacturer/importer in a potential catch-22 situation. Should manufacturers/importers follow GHS methodology to the letter, they may be required to include Health Hazard risks on labels and data sheets not otherwise required under Daubert.

OSHA should consider clarification of its adoption of GHS methodology and parameters in relation to the classification and characterization of the health hazards of a given chemical substance to ensure consistency with established requirements relating to proof of causation under principles set forth by Daubert.

References
1. Daubert v. Merrill Dow Pharmaceuticals, 509 U.S. 579 (1993)
2. www.osha.gov/dsg/hazcom/hazcom-faq.html, http://ec.europa.eu/enterprise/sectors/chemicals/reach/index_en.htm, etimine v. secretary of state for work and pensions eu: case c-15/10 celex no. 610C0015(2011)
3. Ibid
4. Ibid
5. Ibid
6. In re Rezulin 2004 wl 288 4327 (S.D. NY 2004) Haggerty v. Upjohn Co. 950 FS 1160, 1164 S.D. FLA 1996 aff'd. 158 F.3d 588 (11th Cir. 1998) Wade Greauex v. Whitehall Labs, 874 FS 1441 Dvi 1994 aff'd. 46 F3d 1120 (3 Cir.1994). The authors recognize there's not necessarily unity of opinion among circuits.
7. Etimine Supra; Nickel Institute v. Secretary for Work & Pensions eu:case c 14/10 celex no. 610J0014 (2011)
8. Amaba Pty. Ltd. v. Dust Disease Tribunal & Booth 2010 Nswca 344 9. Newest Hazem Chemical Hazards GHS Classification Shock OHS 10/1/13
10. Glastetter v. Novartis Pharmaceutical Corp. 107 F. Supp. 2 1015 (E.D. Mo.2000) aff'd. per curiam 252 F.3d 986 A Circuit (2001), Goeway v. United States 886 F.Supp. 1268 (DSC 1995), Sorenson v. Shaklee 31 F.3d 638 (8th Cir. 1994.)
11. http://www.osha.gov/dsg/hazcom/ghs.html
12. See footnote 2.
13. See footnote 2.

This article originally appeared in the May 2014 issue of Occupational Health & Safety.