2

Historical and Legal Perspectives on the NAAQS Framework for Assessing Causality

Causal determinations in the U.S. Environmental Protection Agency (EPA) Integrated Science Assessments (ISAs) are used as part of the National Ambient Air Quality Standards (NAAQS) setting and review process. This process has an involved legal history, elements of which are important in considering causal determinations. The Preamble to the Integrated Science Assessments (EPA, 2015a) summarizes the general approach EPA employs in developing ISAs, including guidelines for evaluating weight of evidence, drawing scientific conclusions, and making causal judgments. These procedures were developed and refined during multiple NAAQS review cycles beginning in 2007, when the first draft ISAs were issued. The procedures have continued to evolve since the Preamble was first introduced. This chapter reviews the legal background for the NAAQS and then describes the historical development of the Preamble’s framework as described in the 2015 Preamble.

LEGAL BACKGROUND FOR THE NAAQS

The Clean Air Act (CAA) directs EPA to set NAAQS “to protect and enhance the quality of the Nation’s air resources so as to promote the public health and welfare and the productive capacity of its population.”¹ States are charged with determining how to reduce local emissions to meet these standards, assisted by federal emissions control programs such as those for motor vehicles or new stationary sources. Tribes also may elect to take this implementation responsibility. State or tribal implementation plans encompass requirements for monitoring ambient air quality, producing inventories, and modeling to determine what emissions reductions are needed, and implementing and enforcing measures to ensure attainment by specified deadlines. These processes are all overseen by EPA, which has authority to step in to regulate directly where tribes have not sought regulatory authority and in the case of recalcitrant states.

Statutory language, legislative history, and court opinions reviewing the standards establish that in setting the NAAQS, EPA must base its decisions on scientific evidence, err on the side of

___________________

¹ 42 U.S.C. § 7401(b)(1).

caution, and protect vulnerable people. The statute and case law also establish that NAAQS are to be set without consideration of compliance costs, in contrast to other provisions of the CAA which allow or require EPA to consider costs and feasibility in establishing emissions limits.

In section 108, the CAA directs the EPA administrator to publish and revise a list of common “criteria” air pollutants, “emissions of which, in his judgment cause or contribute to air pollution which may reasonably be anticipated to endanger public health or welfare.”² There are currently six air pollutants on the criteria pollutant list: carbon monoxide (CO), oxides of nitrogen (NOx), sulfur oxides (SOx), ozone (O₃) and related photochemical oxidants, particulate matter (PM), and lead (Pb). The statute also directs EPA to review and document the latest scientific knowledge as a basis for setting the NAAQS, specifying the scope of that review in section 108(a)(2). This provision directs the development of the scientific reviews that were initially known as “criteria documents” and are presently referred to as Integrated Science Assessments.

Section 109 of the CAA directs EPA to propose and then establish NAAQS for criteria pollutants and to revise them at 5-year intervals. The NAAQS are defined in section 109(b).

Added in the 1977 Amendments, section 109(d)(2)(A) establishes an independent scientific review committee that is charged with reviewing the criteria documents or ISAs and making recommendations to the administrator about revisions to the NAAQS. This seven-member committee is known as the Clean Air Scientific Advisory Committee (CASAC). The CAA specifies that the committee must include at least one member of the National Academy of Sciences, one physician, and one person representing state air pollution control agencies. The chartered CASAC has typically been supplemented by subcommittees whose members bring expertise specific to given air pollutants, health or welfare endpoints, or methodological approaches, although under the Trump administration these supporting committees were disbanded.

___________________

² 42 U.S.C. § 7408(a)(1)(A).

Over the years, most of EPA’s NAAQS decisions have been challenged in court, often by multiple parties and with a variety of claims of deficiencies. Since the beginning, court decisions on these legal challenges have emphasized the precautionary nature of the CAA. In an influential early decision, the District of Columbia Circuit considered a challenge to EPA’s regulation of lead additives in gasoline under the 1970 version of section 211, which directed the administrator to regulate fuel additives if their emissions “will endanger the public health or welfare.” The Court upheld EPA’s regulations over challenges to the scientific basis for the endangerment finding, stating:

In the 1977 Amendments, Congress changed the language of section 211 to emphasize the precautionary nature of the provision. In place of “will endanger the public health or welfare” the revised language is “may reasonably be anticipated to endanger the public health or welfare,” matching the language used for listing new criteria pollutants in section 108(a)(1)(A) (see 42 U.S.C. § 7408(a) (1)(A), above).

In the seminal case on the NAAQS (Lead Industries Association, Inc. v. EPA 647 F.2d 1130; D.C. Cir 1980), the Court reviewed the 1978 NAAQS decision for lead. In setting the standard, EPA focused on children aged 1 to 5 in urban areas, a vulnerable group who were exposed to lead through a variety of routes. Petitioners challenged the NAAQS decision in part by asserting that the subclinical effect that EPA referenced in developing the standard was not “adverse” and claiming that Congress only authorized EPA to protect against effects that were “clearly harmful.” The Court rejected the petition and affirmed the standard. Citing the legislative history of the 1970 and 1977 Amendments and affirming EPA’s decision, the Court found that the NAAQS were to be set to protect public health without consideration of economic or technological feasibility, and that the standards must protect those who are “particularly sensitive to the effects of pollution.” In summary, according to the Court, “Congress directed the Administrator to err on the side of caution.” Subsequent cases echo this language.

The courts have repeatedly affirmed that the NAAQS must protect sensitive or at-risk people and have remanded NAAQS decisions to EPA for failure to adequately consider these groups or for failure to explain how the standards are adequate to protect their members (e.g., Lead Industries Association, Inc. v. EPA 647 F.2d 1130 (D.C. Cir. 1978); American Lung Association v. EPA 134 F.3d 388 (D.C. Cir. 1998); and American Farm Bureau Federation v. EPA 559 F.3d 512 (D.C. Cir. 2009)). In American Farm Bureau Federation v. EPA, for example, public health groups had challenged the EPA Administrator’s 2006 decision to retain the annual primary PM_2.5 standard despite evidence that exposure at lower levels was associated with impaired lung function development in children. The Court remanded the standard to EPA for unreasonably dismissing studies of children’s lung function growth and failing to explain how, in the face of new scientific evidence, the existing standard was “requisite to protect the public health” including that of children, the elderly and other vulnerable subpopulations, while providing “an adequate margin of safety” (American Farm Bureau Federation v. EPA, 559 F.3d 512 (Court of Appeals, D.C. Cir. 2009, p. 519)).

HISTORICAL PERSPECTIVE ON THE 2015 PREAMBLE FRAMEWORK

Early versions of what became 2015 framework for causal determinations appeared in the first ISAs for nitrogen dioxide (NO₂) (EPA, 2008a) and sulfur dioxide (SO₂) (EPA, 2008b). Prior to

this, NAAQS reviews were initiated through comprehensive “Criteria Documents” prepared (after 1977) by staff of the EPA ORD. Criteria Documents were given that name due to the language used in CAA Section 108(a)(2) (42 U.S.C. § 7408(a)(2)). From 1979 to 2006, the criteria documents were followed by a separate “Staff Paper,” written by Office of Air Quality Policy and Standards personnel, which integrated science and policy analyses intended to inform choices on the indicator, averaging time, level and form of the standards (Bachmann, 2007).

In keeping with the precautionary nature of the CAA and limitations of scientific understanding at the time, the earliest criteria documents avoided drawing express conclusions about causality, especially for health endpoints. For example, the 1969 criteria document for sulfur oxides states

Later criteria documents such as those published in 1982 and 1996 for particulate matter employed a more holistic approach for inferring causality from multiple lines of scientific evidence (as will be discussed later in this report) but lacked a formal framework for causal determinations (EPA, 1982, 1996). The degree of formalism has evolved over time. For example, EPA discussed and employed variations of the Bradford Hill aspects of association (Hill, 1965) in the 2004 and 2006 Criteria Documents for PM and ozone, respectively (EPA, 2004, 2006a). However, in these NAAQS reviews, causality was considered more as a continuum or gradient reflecting varying degrees of confidence in a causal association, as illustrated by Table 2.1, reproduced from the 2007 Ozone Staff Paper (EPA, 2007a).

The CAA specifies that NAAQS reviews be completed every 5 years for an individual pollutant, however criteria documents were time and resource intensive to develop and were typically subject to multiple CASAC reviews. For example, the 2006 PM NAAQS was based on a final Criteria Document (EPA, 2004), preceded by four draft PM Criteria Documents (EPA, 1999, 2001, 2002, 2003), and was promulgated 9 years after the 1997 PM NAAQS (EPA, 1997), which came 10 years after the 1987 PM NAAQS (EPA, 1987). In 2006, EPA conducted an internal review (EPA, 2006c) of the NAAQS review process (EPA, 2006b) which led to revisions in the NAAQS review process summarized in memoranda from the deputy administrator (Peacock, 2006, 2007). This new process began with a technical workshop to identify key science and policy issues which supported development of an Integrated Review Plan (IRP). The IRP outlines EPA’s planned schedule and process for the NAAQS review and identifies a series of policy-relevant questions intended to focus the review on critical science and policy issues. This was followed by a draft ISA, intended to replace the Criteria Document with a more concise review of (primarily) the newest, most policy-relevant science, focused on guiding and supporting a separate, quantitative Risk and Exposure Assessment (REA). Following CASAC and public reviews of up to two drafts of the ISA and REA, EPA would publish a first draft of the Policy Assessment (PA) intended to bridge the gap between the scientific assessments presented in the ISA and REA and the policy judgments required by the administrator in deciding whether to retain or revise the standard. Initially this PA was issued in the form of an Advance Notice of Proposed Rulemaking (ANPR). That process was criticized by CASAC and was replaced with a straightforward PA document reflecting EPA staff advice and recommended options for revisions to the primary and secondary NAAQS. Following CASAC, public, and interagency review, a final rule would ideally be published within the statutory 5-year timeframe. This NAAQS review process is similar to that employed today, however the 5-year review cycle remains a challenge.

TABLE 2.1 Qualitative Characterization of Ozone Health Effect Outcomes

Characterization

Overall Confidence in Causal Relationship with Ambient Ozone

– Lung function decrements in healthy children – Lung function decrements in asthmatic children Lung function decrements in healthy adults – Respiratory symptoms in asthmatic children – Respiratory symptoms in healthy adults – Increased lung inflammation – Aggravation of asthma (i.e., increased medication usage, increased asthma attacks) – Respiratory-related hospital admissions – Respiratory-related emergency department visits – Respiratory-related doctor visits – Increased school absences – Respiratory-related mortality during the ozone season – Cardiorespiratory-related mortality during the ozone season – Total non-accidental mortality during the ozone season – Cardiovascular-related hospital admissions

SOURCE: EPA, 2007a.

The first draft ISAs issued under this new process for NOx—Health Criteria (EPA, 2007b) and SOx—Health Criteria (EPA, 2007c) incorporated many of the recommended changes from the newly proposed review process. However, they were criticized by CASAC, which concluded that they did not adequately describe the ISA review process, including EPA’s approach to literature identification and evidence evaluation. CASAC stated that a “clear framework was not established in the draft ISA for considering potential causal effects and non-causal associations of NOx with human health effects” (CASAC, 2007). CASAC offered similar criticism of the first draft SOx ISA, recommending that “a complete description of the approach to causal inference should be provided in the revised ISA” (CASAC, 2008).

EPA responded quickly to these criticisms, and within the year developed second draft ISAs for NOx (EPA, 2008a) and SOx (EPA, 2008b) Health Criteria, which both included extensive introductory sections and annexes describing the ISA procedures and introducing a new “EPA Framework for Causal Determinations.” The framework included procedures for literature review, study selection and evaluation of evidence. It presented a version of the Bradford Hill aspects of association (Hill, 1965) modified for application to air pollutants, and it introduced and documented the five-level hierarchy for classifying the “weight of evidence” for causation (see Table 1.1).

Similar introductory sections describing the EPA framework for causal determinations were repeated and enhanced in subsequent ISAs for PM, CO, O₃, Pb, and a unique combined SOx and NOx—Ecological Criteria ISA, issued between 2008 and 2013 (EPA, 2008c, 2009, 2010, 2013a,b).

Over the course of these NAAQS reviews, CASAC comments regarding the causal determination framework were generally favorable, but they occasionally criticized EPA for not adhering to the framework, or for not providing sufficiently clear descriptions of how the literature review, evidence synthesis, or other procedures outlined in the framework were applied. In some cases, CASAC disagreed with EPA’s choice of a specific causal category or commented that the justification for a causal category was not sufficiently persuasive or transparent. However, CASAC generally supported EPA’s continued use and refinement of the causal determination framework.³ In 2015, EPA finalized the stand-alone Preamble (EPA, 2015a) as the document that summarizes the general procedures EPA employs in developing ISAs, including guidelines for evaluating weight of evidence, drawing scientific conclusions, and making causal judgments.

In this report, the committee considers EPA’s procedures for developing ISAs and EPA’s framework for causal determinations to be those that are described in the 2015 Preamble. EPA has added refinements focused on assessing and documenting study relevance and study quality to the ISA process in ISAs issued after the Preamble was published (i.e., post 2015). This report comments on several of these newer refinements in subsequent chapters, but since they have not been formally proposed as intended future procedures in a document like the Preamble, it is not assumed that EPA intends to employ them in the future.

___________________

³ See, for example, the CASAC comments on the 2009 PM ISA (Samet, 2009) and on the 2011 ozone ISA (Samet, 2011).