Review of the Draft Interagency Report on the Impacts of Climate Change on Human Health in the United States (2015)

Answers to the Statement of Task Questions

The U.S. National Climate Assessment (NCA; Melillo et al., 2014) identified a number of ways in which climate change is affecting or is likely to affect people, infrastructure, natural resources, and ecosystems. Those impacts in turn have the potential for important current and future consequences for human health. Research on these impacts is active, with strong evidence to support some aspects and research still in progress for others. Therefore, there is a need to assess our understanding of how the impacts of climate change on the environment can create stressors that can affect human health in a number of dimensions both now and in the future.

Does the report accurately reflect the scientific literature? Are there any critical content areas missing from the report?

Throughout the Draft Assessment the authors have done a commendable job of identifying key components of and evaluating the scientific literature. The Committee identified some additional specific health impacts that could be included (detailed under the specific chapters below). The Committee also found that there were several areas where the consideration of the science could be enhanced.

The Committee has a primary concern with the authors’ selection of which aspects of potential climate change impacts on human health are included in the Draft Assessment. The process of selecting the specific health outcomes and case studies that were included in each chapter (for example, Lyme disease and West Nile virus in Chapter 4 and non-cholera Vibrios and toxic algae in Chapter 5) is not clear. In preparing a report such as this one, for which there are many health outcomes that could be considered, it is reasonable that only a subset of the possible health outcomes are highlighted. For both transparency and scientific rigor, and given that the scholarship is still evolving, the rationale behind the decisions about topics to be included should be described in greater detail. Without clear explanation of this rationale, it is difficult to determine whether topics are excluded because the literature is incomplete or because the impacts have been studied and have been found to be insignificant. If the same procedures were used in all of the chapters to review the available literature and determine which aspects to include, they should be described in Chapter 1. In addition, the Committee suggests a list of references used in the report be developed and provided as an appendix to this document, if possible.

The Committee further suggests that additional details should be provided on the system and criteria by which the authors chose what literature they reviewed from amongst the broader literature:

• Did they attempt to systematically identify all relevant literature?
• Were there explicit criteria for each set of chapter authors to make those choices?
• Can the report provide a clearer explanation of how these choices were made?

A second major concern the Committee identified was the inconsistent treatment of adaptive behavior throughout the document. Adaptive behavior by individuals and communities,

including attention to forecasts of extreme weather and air quality conditions, is an important potential mitigating factor in the assessment of human health impacts from climate change. That importance is described and discussed in detail in some chapters, e.g., Chapter 2 Extreme Heat, and the discussion of infrastructure in Chapters 5 and 7, but not discussed consistently across the chapters.

Although the Draft Assessment is clearly not intended to make policy recommendations for adaptation, it is designed to provide stakeholders with “updated information on the observed and projected impacts of climate change on human health and changes in risk to health” that “may help inform adaptation decisions in the public health arena” (79 FR 7419 [February 7, 2014]). The authors should review each chapter to ensure that they consistently discuss adaptive behavior in the context of that impact and describe, to the extent that there is available literature, potential adaptive behaviors and interventions that could moderate the health impacts. Authors could also consider building on Key Messages 3 and 4 from the 2014 NCA report. These Key Messages discuss how preparedness and prevention can help protect people from some climate change impacts (Key Message 3) as well as some potential co-benefits from responding to climate change in ways that can have positive outcomes for human health (Key Message 4). Additional consideration and discussion of these important findings would strengthen the Draft Assessment.

One further concern noted by the Committee was that, in general, the authors’ assessment of the literature on how climate change can alter environmental stressors is stronger than their assessment of the relationships of the stressors to human disease. The relative strength of their health assessments varies with the chapters; the Committee comments on individual chapters highlight those areas where the authors’ assessment of health impacts could be enhanced.

Are the approaches to quantitative modeling reasonable and adequately supported by existing literature?

The Committee reviewed all of the major quantitative analyses conducted or commissioned by the authors. Overall, the Committee found that these analyses were well done, that they were generally placed in the context of the broader quantitative literature where appropriate, and that they made valuable contributions to the literature and the Draft Assessment.

The Committee did have specific comments on several of the analyses which are summarized below and described in more detail in the chapter specific comments. The heat and air quality modeling (Chapters 2 and 3) were particularly useful, especially the analyses which factored in planned ozone precursor reductions to better understand likely climate-related impacts. The water and disease modeling was also well done, but the Committee suggests that the strength of the evidence would be enhanced if all of the underlying, publicly-created data were made available to other analysts to further explore, perhaps in a central archive that could be accessed by members of the scientific community (something which has not been made possible to date).

As a general point, methods for all featured modeling results, including a discussion of uncertainties related to modeling approach and information on how to access underlying data, should be provided either within the appropriate chapter or in the Technical Support Document (Appendix 1 of the Draft Assessment). At this stage, the Technical Support Document has only

general discussions of these issues.

Are the findings documented in a consistent, transparent, and credible way?

The Committee was impressed by the authors’ efforts to maintain a consistent approach to evaluating the evidence across a wide range of authors, agencies, and research organizations. This is no small task, and the organizers and leaders of the effort are to be praised for their efforts to maintain consistency of thinking and approach across the entire document. Having said that, the Committee found that there are several critical areas where the Draft Assessment can be enhanced:

• The assessment of likelihood and confidence: Although there is an effort to apply a consistent set of decision tools to assess the quality and strength of the evidence, the description (two sentences in Chapter 1) of the criteria being used to judge the likelihood and the confidence in effects that are described in this report is very limited. In addition, even though it is referenced in Chapter 1 as providing additional detail on the approaches used, the Technical Support Document provides only a general discussion of the sources of uncertainty without any additional details on how specifically that uncertainty should be applied to reaching likelihood and confidence judgments. Nor are there references to similar approaches that have been used in other settings to provide insight into the approaches used here. Report authors could consider including numerical information on the range of uncertainty, together with likelihood statements, in an effort to reduce any errors in interpretations (see Budescu et al., 2009).

  It is also unclear how these criteria for likelihood and confidence were applied in each chapter. The role played by each author in reaching these conclusions and the consistency of their applications across the multiple chapters is an important consideration here. In addition, there seems to be a disconnect in some of the chapters between the Key Findings and the text of the chapter in which “medium” findings are aggregates of findings that may be “high,” “medium,” and “low”. There may be opportunity to disaggregate confidence determinations in key findings sentences to clarify the likelihood and confidence of sub-findings.

• The description of likelihood and confidence: In addition to the issues about the criteria and their application, throughout the document there is inconsistency in how the likelihood and confidence statements are communicated in the chapters. The basic formula for quantifying health impacts (top of page 35) has two critical elements for which the quality of the science and the strength of the evidence must be evaluated: expected change in exposure and the exposure-dose-response relationship. In some chapters, these individual parts are rated and presented with individual confidence/likelihood statements, whereas others provide only a single evaluation. The Committee suggests providing statements for the confidence or likelihood of each aspect of the evaluation as a way to provide greater clarity and to enable a better understanding of the overall evaluation.

  Further, there are a wide range of descriptors applied throughout the document (e.g., “will,” “may,” “could,” “tends to,” “are likely to,”) to describe summary evaluations. Given that all statements of impact are accompanied by a statement of

• confidence/likelihood, the Committee would suggest that the standard form throughout should be, for example, “climate change will increase” (or “is increasing” in the case of current documented effects) followed by an appropriate statement of likelihood and/or confidence.

• The focus of this report is the health impacts of climate change. To improve clarity, the authors should strive to describe key findings by starting with the evidence of health impact (e.g., “Climate change will increase the frequency/prevalence/other of disease X [Confidence]. Changes in environmental stressor Y will result from climate change impact Z [Confidence]. Environmental stressor Y impacts the frequency/prevalence/other of disease X [Confidence]. Key issues that have an impact on the degree to which this stressor can impact human health include … Vulnerable populations include A, B, C …” This is the structure in some chapters (e.g., Chapter 2) but not consistently throughout.

• The order and linkages between the chapters: There is no guidance for the reader on how the various topics were assigned to the various chapters. Given that many aspects of health and climate change can overlap, a clear early description of what is in each chapter would be advantageous to the reader. While the Committee understands the challenges of inevitable overlap and has made its chapter-specific comments below in the order that they appeared in the Draft Assessment, the Committee encourages the authors to consider a revised ordering of chapters to enhance linkages, where possible, between related issues: Temperature-Related Death and Illness; Extreme Weather; Air Quality Impacts; Vectorborne Diseases; Water-Related Illnesses; Food Safety, Nutrition, and Distribution; Mental Health and Well-Being; and Climate-Health Risk Factors and Populations of Concern (potentially renamed simply “Populations of Concern” as noted below).

  After reorganization, the authors should consider reviewing all of the chapters with the goal to improve the linkages between the chapters (e.g., making clearer the linkages between water, shellfish contamination, and food safety) and confirm that topics are dealt with as consistently as possible.

• The Treatment of Vulnerability: the authors made a useful and correct decision both to identify vulnerable populations in each chapter and to dedicate an entire chapter at the end to summarizing those populations and vulnerabilities. There are, however, inconsistencies in how populations of concern are identified and described in each chapter. To improve the document, the Committee suggests that the definition of vulnerability be moved to Chapter 1 and that the authors carefully review whether the discussions of vulnerability in the individual chapters are consistent with that definition and the more detailed discussion in Chapter 9.

  The Committee thought highly of the organization of Chapter 9 by type of vulnerable population, but would consider renaming it to simply “Populations of Concern” to better communicate the importance of the Chapter to broader audiences. Also, within Chapter 9, populations of color, immigrant populations, and non-English speaking populations are lumped together. The text describes in more detail specific aspects of their vulnerability, but there is considerable overlap among the groups in other aspects of vulnerability like socioeconomic status (SES) and level of English proficiency (LEP); it would be helpful to add detail further distinguishing those cases where race/ethnicity in and of itself may add vulnerability rather than being a surrogate for SES, LEP, or some other vulnerability.

Are the report’s key messages and graphics clear and appropriate? Specifically, do they reflect supporting evidence, include an assessment of likelihood, and communicate effectively?

While we find little disagreement with the actual findings of the Assessment, we feel that both the key messages and the graphics could be enhanced.

• As noted above, each of the findings should be reviewed and reordered to describe the health related outcome(s) first wherever possible. Also, the Executive Summary opening pages could be enhanced by adding highlighted summary bullets of the major health impacts for which there is the highest likelihood and/or highest confidence, of both current and future effects. The rest of the Summary could then portray in concise fashion each of the specific findings (with perhaps shorter introductions).
• The Committee found that graphics, maps and graphs, which portrayed the evidence of actual or expected change in health effects (e.g., the ozone maps), were the most informative. Descriptive graphics intended to provide a roadmap from climate to health (e.g., Farm to Table in Chapter 6) seemed attractive for a broader public but do not appear to be the most effective way of communicating to the target audiences described in the Federal Register Notice (79 FR 7419 [February 7, 2014]) (i.e., public health officials, urban planners, decision makers, and other stakeholders). These descriptive graphics should be included in separate graphic material made available with the report to be used in communications with the public and others.
• The Overarching Graphic (Figure ES-1 and similar figures) needs work to ensure that the message is adequately conveyed to the desired audiences. It may be useful to have a simpler version (without the bulleted points) as the initial graphic to portray the key concepts being put forward. In addition, what are intended to be arrows linking the side compartments (“Social Determinants” and “Non-Climate Stressors”) are too subtle to be obvious and should be reconsidered. Also, the title “Non-Climate Stressors” is somewhat misleading as these are environmental stressors that can exacerbate the impacts of climate drivers on environmental degradation. Finally, some of the chapters have modified this figure to use different headings and slightly different concepts; if the goal of using a single graphic across multiple chapters is intended to make the impacts easier to understand, the uniformity of the message should be carefully reviewed.

Are the research needs identified in the report appropriate?

Although identification of research gaps or a research strategy is not a principal goal of the Draft Assessment, it is evident throughout the document that there are important needs for continuing monitoring and research. The sections on “Emerging Issues” and “Research Needs” identify a number of such areas needing additional monitoring and research. However, the sections are quite inconsistent throughout with variation among the chapters in the relative significance and importance of the needs. The Committee would suggest that the authors do a more consistent job of identifying the most important continuing needs in each chapter, and the Committee provides specific suggestions in the reviews of Chapters 3, 4, 5, and 6. More consistent descriptions of research needs would make it possible for USGCRP and member

agencies to follow up with an integrated research strategy.

The Committee also notes that, given the continuing uncertainties identified in each section of the Assessment, and in many cases the paucity of data for reducing those uncertainties, the Assessment could make an overarching recommendation for enhanced monitoring of exposure pathways and health effects as climate change proceeds. New tools and technologies for such monitoring are being developed at a rapid pace and could be brought to bear on informing future decisions on adaptation and mitigation.

Does the report meet its stated goals? Is the report responsive to the nation’s needs for information on the health impacts of climate change and their potential implications?

The Interagency Task Force has compiled a very good first summary and analysis of a range of potential human health effects that are being linked to or could be caused in the future by climate change. In doing that, they have sought to meet the objectives stated in the Federal Register (79 FR 7418 [February 7, 2014]):

The authors have evaluated a wide range of literature and commissioned and/or taken advantage of significant new quantitative efforts to estimate the likely future impact of climate change on both environmental stressors and human health. Despite their considerable efforts, however, the Draft Assessment does not fully meet the goals as outlined in the Federal Register, and the Committee offers a number of overarching suggestions (noted above in response to its Statement of Task) on how the authors can enhance their identification and assessment of the science and better communicate the conclusions to all of their target audiences. The Committee also offers specific comments on the individual chapters in the following sections and more detailed comments and suggested edits in Appendix A. We trust that in responding to these comments and suggestions that the authors can significantly enhance their final report and meet, if not exceed, their goals and the nation’s needs.