AMERICAN THORACIC SOCIETY DOCUMENTS

Recommendations for the Appropriate Structure, Communication, and Investigation of Tobacco Harm Reduction Claims An Official American Thoracic Society Policy Statement Frank T. Leone, Kai-Hakon˚ Carlsen, David Chooljian, Laura E. Crotty Alexander, Frank C. Detterbeck, Michelle N. Eakin, Sarah Evers-Casey, Harold J. Farber, Patricia Folan, Hasmeena Kathuria, Karen Latzka, Shane McDermott, Sharon McGrath-Morrow, Farzad Moazed, Alfred Munzer, Enid Neptune, Smita Pakhale, David P. L. Sachs, Jonathan Samet, Beth Sufian, and Dona Upson; on behalf of the American Thoracic Society Tobacco Action Committee

THIS OFFICIAL POLICY STATEMENT OF THE AMERICAN THORACIC SOCIETY WAS APPROVED OCTOBER 2018

Rationale: The tobacco harm reduction literature is replete with vague approval was achieved. Patient representatives ensured the language, far-reaching claims, and unwarranted certainty. The American document’s relevance to the lay public. Thoracic Society has increasingly recognized the need for a framework for fi reliably making such claims. Evidence-based standards improving the Results: Fifteen recommendations were identi ed, organized into scientific value and transparency of harm reduction claims are expected to four framework elements dealing with: estimating harm reduction improve their trustworthiness, clarity, and consistency. among individuals, making claims on the basis of population impact, appropriately careful use of language, and ethical considerations in Methods: Experts from relevant American Thoracic Society harm reduction. committees identified key topic areas for discussion. Literature search strategy included English language articles across Discussion: This statement clarifies important principles guiding Medline, Google Scholar, and the Cochrane Collaborative valid direct and inferential harm reduction claims. Ideals for effective databases, with expanded search terms including tobacco, communication with the lay public and attention to unique ethical addiction, smoking, cigarettes, nicotine, and harm reduction. concerns are also delineated. The authors call for formal systems Workgroup members synthesized their evidentiary summaries of grading harm reduction evidence and regulatory assurances of into a list of candidate topics suitable for inclusion in the final longitudinal surveillance systems to document the impact of harm report. Breakout groups developed detailed content maps of each reduction policies. topic area, including points to be considered for suggested recommendations. Successive draft recommendations were Keywords: addiction; prevention; smoking; e-cigarettes; tobacco modified using an iterative consensus process until unanimous dependence

Contents Anticipated Impact on Anticipated Population-based Overview Individuals Impact I. Claims of Harm Reduction II. Claims of Harm Reduction III. Harm Reduction Claims Must Must Be Accompanied by Must Be Accompanied by Be Carefully Constructed to Explicit Estimates of the Explicit Estimates of the Explicitly Avoid Overstatement or Misrepresentation

Supported in part by funds from the American Thoracic Society. The authors received no financial support for their participation in this project. None of the authors have relationships to tobacco industry sources. An Executive Summary of this document is available at http://www.atsjournals.org/doi/suppl/10.1164/rccm.201808-1443ST.

ORCID IDs: 0000-0002-3026-8521 (K.-H.C.); 0000-0002-5091-2660 (L.E.C.A.); 0000-0002-8587-8343 (M.N.E.); 0000-0002-9062-409X (H.K.); 0000-0002-1576-5394 (S.M.-M.); 0000-0002-4051-962X (S.P.). You may print one copy of this document at no charge. However, if you require more than one copy, you must place a reprint order. Domestic reprint orders: [email protected]; international reprint orders: [email protected]. Correspondence and requests for reprints should be addressed to Frank T. Leone, M.D., M.S., Comprehensive Smoking Treatment Program, Penn Lung Center, Suite 251 Wright-Saunders Building, 51 North 39th Street, Philadelphia, PA 19104. E-mail: [email protected]. Am J Respir Crit Care Med Vol 198, Iss 8, pp e90–e105, Oct 15, 2018 Copyright © 2018 by the American Thoracic Society DOI: 10.1164/rccm.201808-1443ST Internet address: www.atsjournals.org e90 American Journal of Respiratory and Critical Care Medicine Volume 198 Number 8 | October 15 2018

AMERICAN THORACIC SOCIETY DOCUMENTS

IV. The Investigation and/or Recognition of Need for Rapidly Harm Reduction Claims Based Regulation of Potential Harm Available Harm Estimates on Population Impact Estimates Reduction Strategies Must Pay Language as an Obstacle to Careful Use of Language in Harm Careful Attention to the Ethical Progress Reduction Claims Considerations Unique to This Methods Ethical Considerations When Concern Recommendations Claiming Harm Reduction Introduction Estimating the Potential Harm Discussion Analogies to Other Disciplines Reduction Impact on Individuals Patient Perspective

I prefer an attitude of humility achieving this goal is not yet possible, models to actual human experience corresponding to the weakness of our consensus on evidence-based standards should accompany the claim. intellectual understanding of nature and for improving the reliability and d Given our understanding of the protean of our own being. — transparency of harm reduction claims biopsychosocial consequences of Albert Einstein in a 1949 letter wouldbeexpectedtoimprovetheclaim’s nicotine dependence, the potential for trustworthiness for patients, provide causing or sustaining addiction should Overview clarity for clinicians, offer consistency of be accurately represented in risk approach for researchers, and set a future trade-off calculations. During the production of the Official direction for regulators. American Thoracic Society (ATS) Research To this end, this report seeks to II. Claims of Harm Reduction Must Statement on research needs in tobacco summarize evidentiary insights within Be Accompanied by Explicit Estimates dependence, the scientific literature on fi four identi ed criteria for valid harm of the Anticipated Population-based tobacco harm reduction was found to be fi reduction claims. The report makes fteen Impact replete with vague language, far-reaching key recommendations, including those d Population-based harm reduction claims claims, and unwarranted certainty as outlined here. to outcome (1). In light of the well- should be accompanied by a statement documented, intentionally deceptive past of known or suspected risk trade-offs practices of the tobacco industry, it is I. Claims of Harm Reduction Must Be incurred. Risk trade-off estimations important that the scientific and clinical Accompanied by Explicit Estimates of should include the impact on the entire community improve the reliability, the Anticipated Impact on Individuals community, not just the smoke-exposed subset. Calculations should account for reproducibility, and transparency of d Inference based on emerging or tobacco harm reduction claims by incomplete evidence is an acceptable potential tobacco use promotion and establishing clear and precise guidelines technique for estimating the anticipated incorporate relevant nonhealth outcomes for use in communications with the lay impact of harm reduction strategies on in their formulation. d and professional public. Such a shared individuals, particularly in instances It should not be assumed that harm framework is integral to the future where the impact on human illness is reduction claims are generalizable to all population subgroups. Claims should development of evidence-based policies, expected to be delayed. Inferential fi comparative assessment of varied methods should be explicitly described speci cally state the age cohort to which approaches to harm reduction, and when making harm reduction claims. they apply and the anticipated impact on adolescents and specify applicability to development of generally accepted research d Inferential claims relying on surrogate methods useful to future investigators. markers of disease, or on markers of pregnancy and lactation. The recommendations set forth in this exposure to risk-associated constituents, statement are not intended to suggest a should explicitly state the nature of, and III. Harm Reduction Claims Must position on the relative merits of harm degree of certainty in, the relationship Be Carefully Constructed to reduction per se, nor are these observations between markers and the specific harms Explicitly Avoid Overstatement or intended to settle the current debate over reduced. Misrepresentation the most appropriate approach to harm d When using number of cigarettes d Harm reduction claims should be reduction. This statement does not review consumed as a surrogate measure of presented in lay language, state the or endorse any individual approach to exposure, explicit care should be taken to perspective(s) from which they are derived, achieving harm reduction goals. Rather, account for anticipated compensatory and include an explicit range of possible this statement is intended to provide a adjustments in smoking topography outcomes expected from implementation. framework for establishing “standards of (i.e., physical characteristics of smoking d Harm reduction claimants should practice” related to tobacco harm reduction behavior, such as puff count, puff volume, exercise caution in language choices, claims. Clearly, complete remediation of average flow, puff duration, and interpuff explicitly minimizing the potential for tobacco smoke exposure remains the interval). exaggeration of effect size, misattribution primary objective and should never be d When based on animal or cell culture of protective effect, or other potential compromised as a matter of expediency. models, comprehensive narrative misunderstandings on release to the However, for circumstances in which statements of relevance of the chosen public.

American Thoracic Society Documents e91

AMERICAN THORACIC SOCIETY DOCUMENTS d Product-oriented population harm This framework represents a means of Nonetheless, popular misconceptions reduction claims should account for all assuring the public that threats to the validity persisted, changing the pattern of tobacco known (including unintended) uses or of harm reduction claims have been product use and providing an illusory modifications of the product in modeling minimized and of maximizing the public’s pathway for those who wished to continue impact, particularly when estimating confidence in such claims. Furthermore, we smoking while “avoiding” harm (9). impact on tobacco-naive users. call for several implementation steps to be During the 1980s, the notion that used, including the development of a formal nicotine caused a compulsive behavioral system of harm reduction evidence grading, a disorder in a manner similar to other IV. The Investigation and/or longitudinal surveillance system to document drugs of abuse was gaining acceptance, Regulation of Potential Harm the impact of a harm reduction interventions, culminating in the 1988 Surgeon General’s Reduction Strategies Must Pay and a careful reexamination of the ethical statement that cigarette smoking is the Careful Attention to the Ethical constraints on harm reduction research when behavioral manifestation of addiction (10). Considerations Unique to This effective harm elimination interventions exist. Consequently, the proposition emerged Concern that health professionals should mitigate d Given the established evidence of safety the harms of tobacco smoke exposure and efficacy of harm reduction methods Introduction among those patients so severely dependent for people unprepared to make a quit as to be unable to completely cease attempt, marginal harm reduction It has been well established that tobacco smoking. Nicotine replacement therapies calculations must use best available products, particularly combustible were developed, offering the possibility of options as the comparator, rather than cigarettes, causally increase risk for using medicinal nicotine therapies for harm the no-intervention condition. In all numerous diseases and other adverse reduction objectives when “cessation” was cases, harm reduction claims should be consequences in a dose-dependent manner otherwise impossible (11). accompanied by clear statements of their and that abstaining or reducing exposure Subsequent observational studies of methodologic decisions in these respects. consequently reduces this risk (2). Some Swedish cohorts using moist snuff (snus) d The hazards inherent to prospective benefits accrue immediately, and others suggested that increasing use of snus was population-based evaluations of tobacco continue to accumulate over a lifetime associated with favorable trends in incidence harm reduction strategies may qualify (3, 4). Accordingly, the concept of a safer of lung cancer and other tobacco-related as “greater than minimal risk.” Care cigarette has, for many years, seemed illnesses. The conversation regarding useful should be taken to assess research risks in straightforward: build a reduced-risk strategies for harm reduction had thus comparison to current standards of care for device through the practical application of extended to include the relative public reducing harm among people unprepared dose–response concepts (5). Unfortunately, health value of alternatives to smoked toquit,ratherthanincomparisontothe such past efforts proved unsuccessful and, tobacco, including chew and pipe tobacco no-intervention condition. ultimately, misleading (6). After the 1964 (12, 13). Once again, implied and direct d Harm reduction researchers must protect Surgeon General’s report causally linking harm reduction claims blossomed within the autonomy of subjects by discussing smoking to lung cancer, and in response to popular media. Alongside this growing the unique nature of risk inherent to declining rates of smoking in the United interest in tobacco products as harm harm reduction protocols. Promoting an States, tobacco companies began applying reduction devices, the Institute of Medicine understanding sufficient for subjects adjectives such as “light,”“mild,” or “low,” issued a landmark, comprehensive set of to legitimately assume risk protects with the implication of reduced risk, taking recommendations for formally assessing subjects from unforeseen or tragic advantage of popular assumptions about and regulating the harm reduction potential consequences of participation through cigarette constituent dose–response. Those of products introduced into the market fair methods that promote just outcomes. assumptions did not hold up to subsequent (Figure 1) (14). d Harm reduction policies should state scientific inquiry. the perspective(s) from which they The 1981 report of the Surgeon General Analogies to Other Disciplines are derived, guard against inequitable addressed the “changing cigarette” and The reduction of harm is a central precept of allocation of resources, provide assurance concluded that federal policy should medical care, routine in the management of of longitudinal surveillance, and inform continue to call for smoking cessation, chronic conditions like asthma, diabetes, the public of important developments in given the discordance between machine- or hypertension. The concept of “harm understanding. measured yields and the potential for harm reduction” has also evolved in the public d Harm reduction claims should be (7). The report pointed out that reduced- health arena, with a range of public health accompanied by a narrative statement of yield cigarettes might reduce harm, if users policies targeted at individuals, families, or the potential for unintended/involuntary did not change their smoking patterns. communities and designed to mitigate the impact on non–index users. However, compensatory behaviors did negative physical or social consequences Although elimination of risk should occur in response, and lower-yield of a variety of human endeavors (15). remain the goal of public health authorities, cigarettes did not lead to decreased risk Educational harm reduction strategies focus this statement highlights the special (2, 8). In fact, over recent decades, the risk on decreasing dangerous exposures by requirements for ethically pursuing the of lung cancer associated with smoking increasing awareness and modifying social reduction of harm within the complex social/ has increased despite a contemporaneous norms. For example, focused attention on biological context of tobacco dependence. trend of declining measured tar yield. the impact of driving drunk has helped e92 American Journal of Respiratory and Critical Care Medicine Volume 198 Number 8 | October 15 2018

AMERICAN THORACIC SOCIETY DOCUMENTS

normalize safe ride techniques such as BOX 1 designated drivers. Safer sex education Regulatory Principles aimed at teen students has reduced transmission of infectious disease, Regulatory Principle 1. Manufacturers of tobacco products, whether conven- tional or modified, should be required to obtain quantitative analytical data on the decreased teen pregnancy rates, and ingredients of each of their products and to disclose such information to the regu- promoted safer sexual decision making latory agency. among those engaged in sexual activity Regulatory Principle 2. All tobacco products should be assessed for yields of (16, 17). Alternatively, harm reduction nicotine and other tobacco toxicants according to a method that reflects actual strategies have also focused on minimizing circumstances of human consumption; when necessary to support claims, human the sequelae of exposures, without directly exposure to various tobacco smoke constituents should be assessed using appro- impacting behaviors. Examples include priate biomarkers. Accurate information regarding yield range and human expo- sure should be communicated to consumers in terms that are understandable and free or low-cost HIV testing, integrating not misleading. primary care services within substance Regulatory Principle 3. Manufacturers of all PREPs should be required to abuse treatment programs, and lung cancer conduct appropriate toxicological testing in preclinical laboratory and animal mod- screening programs. els as well as appropriate clinical testing in humans to support the health-related Perhaps most familiar are the strategies claims associated with each product and to disclose the results of such testing to that attempt to “manage” the variables the regulatory agency. that confer harm in circumstances where Regulatory Principle 4. Manufacturers should be permitted to market tobacco- the behavior itself may not be directly related products with exposure-reduction or risk-reduction claims only after prior modifiable. Public policies such as needle agency approval based on scientific evidence (a) that the product substantially reduces exposure to one or more tobacco toxicants and (b) if a risk reduction claim exchange programs (18, 19) and safe is made, that the product can reasonably be expected to reduce the risk of one or injection locations (20), wet shelters to more specific diseases or other adverse health effects, as compared with what- manage safe alcohol consumption among ever benchmark product the agency requires to be stated in the labeling. The the homeless (21), and free condom “substantial reduction” in exposure should be sufficiently large that measurable distribution in community health centers reduction in morbidity and/or mortality (in subsequent clinical or epidemiological (22) adopt a nonbinary understanding of studies) would be anticipated, as judged by independent scientific experts. risk and acknowledge that continued Regulatory Principle 5. The labeling, advertising, and promotion of all tobacco- related products with exposure-reduction or risk-reduction claims must be carefully adverse behavior need not confer continued regulated under a “not false or misleading” standard with the burden of proof on harm. Theoretically, promoting the use of the manufacturer, not the government. The agency should have the authority and nicotine delivery systems or alternative resources to conduct its own surveys of consumer perceptions relating to these tobacco products as a means of managing claims. addiction while reducing the exposure Regulatory Principle 6. The regulatory agency should be empowered to to cigarette smoke, divorced from any require manufacturers of all products marketed with claims of reduced risk of expectation of reduction in tobacco use tobacco-related disease to conduct post-marketing surveillance and epidemiological studies as necessary to determine the short-term behavioral and long-term health behaviors, could be considered an consequences of using their products and to permit continuing review of the accu- analogous example of this type of risk racy of their claims. management approach to harm reduction if Regulatory Principle 7. In the absence of any claim of reduced exposure or the erroneous assumptions that undermined reduced risk, manufacturers of tobacco products should be permitted to market past attempts are not repeated. The urgent new products or modify existing products without prior approval of the regulatory need for a fresh approach to the tobacco agency after informing the agency of the composition of the product and certifying epidemic remains intact. that the product could not reasonably be expected to increase the risk of cancer, For the purpose of this statement, when heart disease, pulmonary disease, adverse reproductive effects or other adverse health effects, compared to similar conventional tobacco products, as judged on considering claims of harm reduction as the basis of the most current toxicological and epidemiological information. they relate to tobacco smoke exposure and Regulatory Principle 8. All added ingredients in tobacco products, including the potential to impact the public health, we those already on the market, should be reported to the agency and subject to a refer exclusively to the 2001 definition comprehensive toxicological review. developed by the Institute of Medicine: “A Regulatory Principle 9. The regulatory agency should be empowered to set product is harm-reducing if it lowers total performance standards (e.g., maximum levels of contaminants; definitions of terms tobacco-related mortality and morbidity such as “low tar”) for all tobacco products, whether conventional or modified, or for even though use of that product may classes of products. involve continued exposure to tobacco- Regulatory Principle 10. The regulatory agency should have enforcement ” powers commensurate with its mission, including power to issue subpoenas. related toxicants (14). Regulatory Principle 11. Exposure reduction and risk reduction claims for drugs that are supported by appropriate scientific and clinical evidence should be Recognition of Need for Rapidly allowed by the FDA. Available Harm Estimates The recent introduction of novel tobacco Figure 1. Institute of Medicine recommendations for regulatory principles guiding the use of products, such as the electronic (e)-cigarette, “Reduced Harm” designation (reprinted from Reference 14). FDA = U.S. Food and Drug has reignited debate over the proper Administration; PREP = Potential Reduced-Exposure Products. placement of harm reduction within the

American Thoracic Society Documents e93

AMERICAN THORACIC SOCIETY DOCUMENTS overall tobacco public health strategy. product sold in the United States (32). collaboration with the tobacco industry was Concern over unsubstantiated harm Beginning in the 1970s, the tobacco involved in claim generation (41, 42). reduction claims and the potential threat to industry implicitly and falsely made harm Rather than adopt a policy position on public health that they represent has been reduction claims in its marketing of any particular approach to achieving harm the subject of debate for a decade (23). cigarettes with reduced yield of tar and reduction goals, the ATS Tobacco Action Previous frameworks for establishing an nicotine, forming the basis for litigation Committee sought to establish evidence- evidence-based approach to validating against the tobacco industry brought based “standards of practice” in making harm reduction claims have included core by the U.S. Department of Justice under tobacco harm reduction claims, expected components such as establishing standard the Racketeer Influenced and Corrupt to improve both claim trustworthiness methods for determining relative risk Organizations Act (33). and consistency of approach necessary for across products, clarification of the Clearly, neither the public health nor regulatory direction. relationships between toxicants and health the scientific basis for harm reduction outcomes, and monitoring the potential policies and product regulation would be population-based impact of these claims on expected to benefit from continued use of Methods tobacco initiation or delays in cessation. poorly defined terms such as “light” or The need for evidentiary certainty, “low tar.” In the absence of standardized To ensure balanced advice and input from a however, is necessarily counterbalanced by language, the commercial value of implicit variety of perspectives, an expert panel was the need for rapid implementation, given branding should be expected to remain convened in February 2016 that included the magnitude of the threat posed by enormous (34–36), particularly in those members drawn from several leadership continued tobacco smoke exposure. Some jurisdictions where regulatory and public committees of the American Thoracic authors have even called for post hoc scrutiny of tobacco products may be much Society (ATS). In addition to members of monitoring as an ethically acceptable less vigorous (37). Even in instances when the ATS Tobacco Action Committee, the method of evidence generation, due in large these traditional imprecise terms are expert panel included representation from part to tobacco’s unique position as a restricted from use, the implications of the Documents Development Committee, worldwide risk agent and the urgency to their meaning can carry over into branded Health Policy Committee, Ethics and prevent its otherwise inevitable casualties messaging and package design (38). Conflict of Interest Committee, and the (24, 25). Vague harm reduction language allows for Pediatrics Assembly. To ensure relevance As more potential harm reduction consumer misdirection, particularly when of project design to the broadest set of products are developed, how should we then commercial implications are large (39, 40). stakeholders, external support was obtained balance the need for rapid availability with Harm reduction strategies are primarily from two patient representatives identified the need for a reasonable assurance the based on principles of toxicology and through the ATS Patient Advocacy harm reduction claims are accurate? How human biology but are also distinctively Roundtable and from the Centers for should we make focused harm reduction a function of risk perception and Disease Control Office on Smoking and claims explicitly precise without sacrificing communication science. If the goal of harm Health. clarity for the user (26, 27)? The scientific reduction is to enhance the capacity of Work proceeded in three phases. In the community has a responsibility to counter adults to make an informed decision conceptualization phase, the project chair impulsive, nonscientific conclusions that concerning the use of tobacco, an moderated open-ended discussion among are often based on the enormity of the appropriate balance must be struck work group members, aimed at identifying problem, the political positions of the between avoiding irrelevance through the major themes that would go on to form agents, or the commercial potential of the excessive precision on one hand and the framework for recommendations. Key products. As a social institution, science has allowing for vagueness and innuendo topic areas along with possible points of an obligation to look beyond itself and through imprecision on the other. discussion were identified. Individuals help with global problems (28). It is our Attention to language is essential to within the workgroup were identified who responsibility to ensure information is developing a uniform database of could review the literature in each topic area, consistent with established scientific norms, reproducible results from which to prepare slide presentations that would maintaining the integrity of the data, the draw reliable conclusions. Unproven familiarize workgroup members with trustworthiness of conclusions, and the or premature reduced-risk claims, or available evidence, and set the stage for more transparency of aims (29, 30). By its very reduced-exposure claims that are in-depth discussion. A full workgroup nature, research on tobacco harm reduction mistakenly interpreted as reduced-risk meeting was convened using internet-based has been too often limited by inadequate claims, are a significant threat to the notion conferencing tools, during which each size, scope, or duration to yield useful of informed consent and autonomous of the topic areas was discussed in turn. guidance (31). control over risk management. Transcripts of the discussion were used to This experience has led to general generate broad outlines of content areas to Language as an Obstacle to Progress wariness of harm reduction claims within be explored further. Responding to the accumulating data on the tobacco control community. Continued Members searched several databases, the risks of cigarette smoking, the tobacco use of vague or misleading terms when including Medline, Google Scholar, and the industry began earnestly introducing filters making future harm reduction claims Cochrane Collaborative, to identify primary to their product lines in 1950. Filtered would be expected to continue to undermine sources and reviews. Searches were limited cigarettes quickly became the predominant trustworthiness, particularly when to articles published between 1950 and 2017 e94 American Journal of Respiratory and Critical Care Medicine Volume 198 Number 8 | October 15 2018

AMERICAN THORACIC SOCIETY DOCUMENTS and available in English. Expanded search of individual or population health would the strategy on human illness is expected to terms were used to identify the evidence, be most acceptable to the lay consumer? be delayed. In all cases, inferential methods including MeSH headings: Tobacco What regulatory standards (e.g., degree used should be explicitly described when (D014026), Smoking (D012907), Tobacco of certainty, long-term monitoring, etc.) making harm reduction claims. Products (D062789), Electronic Nicotine would give patients confidence in a harm In contrast to a factual claim (e.g., Delivery Systems (D066300), Nicotine reduction claim? How should complex and formaldehyde concentration measured in (D009538), and Harm Reduction or competing harm reduction claims be smoke from product A is lower than that (D040261), in addition to keywords unique communicated to patients in the setting of a measured from product B), inferential to each thematic area. Workgroup members clinical visit? claims are evaluative in nature, using a synthesized their evidentiary summaries The resulting final draft policy factual premise to support a conclusion into a list of candidate topics suitable for statement was subject to review by the (e.g., product A is safer than product B). inclusion in the final report. ATS Documents Editor, the ATS Board of Inferential claims may be explicit, typically Breakout group conference calls Directors, and anonymous peer reviewers, containing indicator words such as “thus” were scheduled 1 week after the initial per ATS policy. or “therefore,” or may be implicit, where meeting, with each call focused on a the critical link between the premise and single topic area and moderated by the the conclusion is alleged or assumed. An project chair. The breakout groups were Recommendations appropriately structured inferential claim charged with developing detailed, annotated should explicitly state the predetermined content maps of their topic areas, including Four standards for ensuring the choices made among the alternative points to be considered for suggested validity of any harm reduction claim inferential options that might have been recommendations. Workgroup members were identified, along with several used to estimate human risk from data that were encouraged to self-select to breakout important considerations relevant to each are not fully adequate or are not drawn sessions that most closely aligned with their (Table 1). Criteria I through III apply to directly from human experience. For expertise. All workgroup members and the specific construction of the claim itself, example, a harm reduction claim might external contributors participated in at least and criterion IV applies to the development assume normal renal function and average one of the topic area calls. and implementation of the claim. To be toxicant dose exposures; specification of In the document development phase, considered valid, harm reduction claims the impact of these assumptions on risk the workgroup held a full-day, in-person must meet all four criteria, specifically: modeling, and the variability associated meeting during which evidence-based I. Claims of harm reduction must be with the alternative conditions, should be content maps developed during the accompanied by explicit estimates of explicit (43). fi conceptualization phase were re ned into the anticipated impact on individuals. Inferential harm reduction claims a structured outline that would roughly II. Claims of harm reduction must be relying on surrogate biological markers of fi correspond to the nal format of the accompanied by explicit estimates of the disease, or on markers of exposure to risk- statement. At this stage, the workgroup was anticipated population-based impact. associated constituents, should explicitly fi asked to nalize the set of recommendations III. Harm reduction claims must be state which measures were used to estimate that would be included in the statement. carefully constructed to explicitly avoid harm, the relationship of the marker to the fi Writing assignments were generated for overstatement or misrepresentation. speci c harms reduced, and the degree to each of the workgroup members, who IV. The investigation and/or regulation of which certainty in this relationship has been were then charged with submitting a 250- potential harm reduction strategies established at the time the harm reduction to 500-word summary of their assigned must pay careful attention to the ethical claim is made. recommendation(s). These executive considerations unique to this concern. A search of the available literature summaries were edited for length and style reveals a wide variety of methods that have and then combined to form the document For example, a claim of the form Xis been used to develop the data foundation for first draft. healthier than Y would not be considered inferential tobacco harm reduction claims. In the prepublication phase, successive valid, because it fails to explicitly state the Inferential claims have been made using drafts were circulated to workgroup anticipated health effect on individuals and both direct (e.g., toxicant concentrations, members for review and modification, populations, and the use of vague terms nicotine concentrations, pharmacokinetic using an iterative consensus process until like “healthier” does not go far enough in metrics) and indirect (e.g., gene expression, unanimous approval was achieved. During explicitly avoiding overstatement. Specific airflow obstruction, cytokine production) this phase, patient representatives were recommendations guiding application of surrogate measures of effect. This flexibility asked to help the workgroup ensure the each of the framework criteria include the is in part driven by the biology of the document’s relevance to the lay public and following. harm being evaluated and in part by to compose a statement of relevance from the limitations of available technology. the patient perspective. Suggested questions Estimating the Potential Harm Flexibility is good, in that it allows for a to be considered included: What are the Reduction Impact on Individuals robust investigation of the target biology, types of harm reduction terms most Inference is an acceptable technique but can undermine confidence in the claim if frequently encountered by patients in lay for estimating the anticipated impact the scientific rationale for methods chosen literature? What terms seem misleading, or of harm reduction strategies on individuals, and assumptions made in risk assessment are to be avoided? Which kinds of measures particularly in instances where the impact of are not immediately clear. Given that these

American Thoracic Society Documents e95

AMERICAN THORACIC SOCIETY DOCUMENTS

Table 1. Framework for Valid Harm Reduction Claims effects of nicotine, and conditioned stimuli may also be relevant. In all cases Framework Criteria Important Considerations where harm reduction claims are based on observable changes in cigarette consumption, measures of compensation should be Claims of harm reduction must be d When is inference acceptable? When is accompanied by explicit estimates direct observation necessary? integrated into harm calculations. of the anticipated impact on d Most appropriate way of dealing with Comprehensive narrative statements individuals inherent uncertainty? regarding the relevance of laboratory models d Direct vs. indirect measures of impact? of harm to actual human experience should Influence of compensation? d How to deal with competing effects? accompany harm reduction claims when (i.e., reduced harm A = increased harm B) based on animal or cell culture models. d Hierarchy of evidence for “reduced harm” Qualitative descriptions of the designation? demonstrated strength of linkage between human physiology and the chosen surrogate Claims of harm reduction must be d What is the relationship of estimated accompanied by explicit estimates individual effects to anticipated population model should be explicit when basing of the anticipated population-based effects? (Ecologic fallacy?) inferential harm reduction claims on the impact d Assessing short- vs. long-term harms? results of such studies. In instances where Surveillance? the correlation is unknown or in doubt, d Direct vs. indirect measures of population effect? Requirements for complex this should be explicitly stated. Also, modeling? the relevance of exposure conditions, d How to deal with social/cultural/environmental including route of administration, dose harms? or concentration, timing, duration, etc., d — Managing risk trade-offs within group vs. to normative human exposure conditions between group d Non–index user harms? should be explicitly stated (44). Fetal/childhood/pregnancy/lactation When making a harm reduction claim, d Unintended uses/exposures the range of possible outcomes expected from d Hierarchy of epidemiologic evidence of implementation should be explicitly stated. “reduced harm”? A narrative description of the known or Harm reduction claims must be carefully d Why is precision important? Lessons from suspected mediators and moderators of the constructed to explicitly avoid prior deceptive practices harm reduction effect and a description of overstatement or misrepresentation d Variety of definitions currently in use how the effect of implementation is expected d fi Preferred terms/de nitions? to be distributed within the population The investigation and/or regulation of d Balancing need for action against need for should accompany all explicit claims of potential harm reduction strategies reassurance harm reduction. In addition, the temporality must pay careful attention to the d Implications for research methods? of the effect should be described explicitly ethical considerations unique to this d Monitoring unanticipated effects (both and the limits of certainty defined. When concern positive and negative) d Is compulsion (dependence) itself a should the user expect this claim to take “harm”? effect? How long will the effect last? Is d Mechanism for informing public of there need for longitudinal monitoring important developments? (Figure 2) (44)? New nicotine dependence is not just the antecedent to illness but a manifestation default decisions can have substantial effects with the product. Even in harm reduction of altered brain biology. Regardless of the on the results of risk assessments, it should strategies aimed at reducing exposure to impact harm reduction strategies have never remain a question as to whether the cigarette smoke by means of reducing the outside of the central nervous system, the best available scientific information and number of cigarettes consumed, it is possible potential for causing or sustaining addiction defaults have been used to examine the for users to alter topography enough to should be accurately represented in risk problem (44). minimize the anticipated risk reduction. For trade-off calculations. When using number of cigarettes example, consuming 10 cigarettes isn’tby Nicotine causes dependence through consumed as a surrogate measure of necessity less dangerous than consuming its action on nicotinic cholinergic receptors exposure, explicit care should be taken to 20 when the qualitative characteristics of in the mesolimbic centers of the brain account for anticipated compensatory the smoking behavior compensate for (47, 48). Nicotine addiction manifests as adjustments in smoking topography during the quantitative change in number of a compulsive disorder, with withdrawal smoking “reduction.” cigarettes consumed. There is evidence causing significant depressed mood, The extent to which smoking that the potential for compensation insomnia, frustration, anxiety, and agitation topography (i.e., puff frequency, puff varies significantly by product (45, 46). to a degree comparable to that caused by volume, hold duration, etc.) changes when Findings support the notion that smokers opiates, amphetamines, and cocaine (49). lower-yield products are used is important compensate for nicotine, but factors Relapse often follows prolonged periods for accurately assessing the risk associated including cigarette draw resistance, sensory of abstinence (50). Adolescent brains are e96 American Journal of Respiratory and Critical Care Medicine Volume 198 Number 8 | October 15 2018

AMERICAN THORACIC SOCIETY DOCUMENTS

trade-off estimations should include the BOX 2-3 Agency Guidance on Risk Characterization: Attention to Uncertainty impact on the entire community where A 1992 guidance memorandum reinforces principles enunciated in the 1983 Red Book and in appropriate, not just the smoke-exposed EPA’s 1986 risk-assessment guidelines and was a forerunner of later guidance documents. subset. Highly reliable data are available for many aspects of an assessment. However, scientific uncertainty is a Public health is by its nature a fact of life for the risk assessment process as a whole. . . . Scientists call for fully characterizing risk not to fi question the validity of the assessment, but to fully inform others about critical information in the assessment. communal good, with burdens and bene ts . . . Even though risk characterization details limitations in an assessment, a balanced discussion of reliable that often fall unevenly on different conclusions and related uncertainties enhances, rather than detracts, from the overall credibility of each as- sessment [Reprinted in NRC 1994, Appendix B, pp. 352-353]. subgroups. This potential for trade-offs The Risk Characterization Handbook (EPA 2000b) instructs risk assessors to, among other things, raises a particular set of challenges when “carry forward the key information from hazard identification, dose-response, and exposure assess- making harm reduction claims relevant to ment, using a combination of qualitative information, quantitative information, and information about tobacco. Although in financial systems the uncertainties” (p. 24) and “describe the uncertainties inherent in the risk assessment and the default positions used to address these uncertainties or gaps in the assessment” (p. 21). greatest rewards are often associated with After highlighting the emphasis on “transparency” in EPA’s 1995 risk-characterization policy (EPA the greatest risks, the risks of poorly 1995), the Staff Paper (EPA 2004a) notes that “one of the major comments on EPA risk assessment designed tobacco policy are both personal practices is that they do not characterize uncertainty and variability transparently enough” (p. 33). The and profound, making the calculus for statement of task for EPA (2004a) confirms that “this is an issue EPA is attempting to address” (p. 33). (See Box 2-4 for related peer-review commentary on one assessment.) considering such trade-offs much less clear. The two most frequently cited adverse concerns remain the unintentional Figure 2. National Research Council’s statement on the need for attention to scientific uncertainty in maintenance of tobacco use by persons who risk characterization (reprinted from Reference 44). EPA = Environmental Protection Agency. would otherwise quit and the initiation of tobacco use among persons who would otherwise never use tobacco. Judgments especially vulnerable to the negative regarding harms, intention to quit, cigarette regarding between-group risk tradeoffs—as consequences of nicotine exposure, given sales data, and even prevalence of smoking, when the health of one group benefits while the high degree of plasticity during this may be poor indicators of population-based that of another is harmed—should include period. Early exposure potentially alters harm. The relationship between such input from the affected groups within the the normal course of brain growth and surrogate measures and actual health decision-making algorithm. Within-group development, affecting learning, reasoning, behaviors is chaotic and weak, particularly risk trade-offs are also foreseeable; how do mental health, attention, impulse control, when indirect measures such as sales data we correctly value avoidance of a known and personality (2, 51–53). Nicotine has are used. Given that smoking is known to be risk if the trade-off involves increasing a powerful gateway effect on the brain, unhealthy, it is easy to assume that reduced another? Are disability-adjusted life-years making susceptibility to the addictive fi prevalence of smoking is de nitionally saved an appropriate metric for considering effects of other psychoactive drugs more less unhealthy. However, relationships the trade-off? Finally, in a setting of likely and recovery from other addictions between behavioral exposures and outcome limited resources, investment in one harm more difficult and conferring a long-term that are true in individuals may not be reduction strategy may delay or otherwise propensity to relapse (54, 55). Given proportionally true within populations, undermine the development of other, the complexity of nicotine’s effect on often requiring complex, multilevel potentially more effective, approaches. Is human psychosocial development, and the fl modeling to re ect the true anticipated there an economic metric appropriate for potential lifelong consequences of early population effects (56, 57). As such, use the trade-off calculation? Given that the exposure, the inherent primary harm of the of surrogate markers of population harm inherent risk trade-offs are complicated, compulsive disorder produced by nicotine reduction should be avoided unless there and that agreed-on metrics for judging should in no case be undervalued in is accompanying epidemiologic evidence acceptability are lacking, the evidentiary individual harm reduction estimates or of direct association in a representative standard in support of such claims needs trivialized in the discussions surrounding population. As an example, consider that to be high until precedent guidance is population-based approaches to reducing the association between elevated cholesterol available for the U.S. Food and Drug the impact of the epidemic. levels and ischemic heart disease was Administration and other regulatory not considered proof that reductions in agencies as they try to correctly set this Harm Reduction Claims Based on cholesterol would save lives until that balance (60). This high evidentiary Population Impact Estimates hypothesis was supported by evidence standard should be no higher than those Surrogate measures of population impact generated in a representative population applied to other conditions but also should should form the basis of inferential harm cohort (58, 59). Similar intellectual rigor not be lower. reduction claims only when the measures should apply to the impact of presumed Population-based harm reduction have a well-established link to actual harm tobacco harm reduction strategies on claims, accompanied by a statement of risk reduction outcomes. In all cases, the nature population health. Known mediators and trade-offs, should also incorporate known of the assumed link, and the degree of moderators should be disclosed. and/or suspected non–health-related variables certainty in this relationship, should be Population-based harm reduction in their disclosure. explicitly stated when making such claims. claims should be accompanied by a Few smokers perceive social approval Intermediate markers of population statement of known and/or suspected risk of their smoking, and most report that risk behaviors, such as knowledge/attitudes trade-offs incurred by the population. Risk tobacco companies cannot be trusted to tell

American Thoracic Society Documents e97

AMERICAN THORACIC SOCIETY DOCUMENTS the truth. Denormalization of tobacco use experimentation, the issue is not merely have an immediate impact than on work has been independently associated with whether a harm reduction claim accurately that includes inferential harm reduction intentions to quit smoking and probability represents the potential value to incumbent claims. Making research findings, including of abstinence at follow-up (61). The smokers but whether the “harmful but limitations, understandable to the public can relationship between social stigma and safer” message will inadvertently lead to help raise awareness and speed adoption smoking, particularly for adolescent, substantial harms in a younger cohort while but can raise unreasonable expectations minority, or otherwise disenfranchised benefitting an older cohort, because of of improved health if poorly executed. groups, raises questions about what impact differential effects of communication (65). The challenge inherent to meeting this the regulatory concept of “more acceptable It should not be assumed that harm important goal is to balance the need for than smoking” will have on tobacco use reduction claims are generalizable to simplification with avoiding distortion (77). prevalence over time. Popular assumptions pregnant women or to the developing fetus. Use everyday language and familiar examples. are that more people will be attracted to the Harm reduction claims should specifically Address the question “what does this mean more acceptable option. However, we have state their applicability to pregnancy and to me?” Anchor risk reduction estimates not adequately considered how tobacco- lactation. to other familiar, tangible examples of related stigma overlaps with other social Undeniably, the safest method of everyday life. identity stigmas, with evidence that some tobacco harm reduction during pregnancy is Harm reduction claims should exercise groups are in fact attracted to the less- to stop smoking. Smoking during pregnancy caution in language choices, explicitly acceptable options, potentially exacerbating may cause considerable harm to the fetus, minimizing the potential for exaggeration of health inequities (1). Use of mathematical which may have lifelong consequences. effect size, misattribution of protective effect, models to predict future public health Smoke exposure during pregnancy has been or other potential misunderstandings once scenarios should explicitly state which related to increased risk of preterm birth, released to the public. assumptions were used in calculations, small for gestational age births, sudden Inferential harm reduction claims the confidence intervals of values used, infant death (66), and structural airway should be based on data rather than the strength of available evidence used changes in newborn babies (67). Reduced theoretical or “common sense” rationales. as the basis for the model, and the results lung function at birth is related to reduced Objective data are critical to avoiding a of sensitivity analyses surrounding lung function 10 years later (68) and at repeat of baseless harm reduction claims all assumptions that influenced the 16 years of age with associated allergic made about numerous products in the past conclusions (62). Methods count—it is comorbidities (69). Even the grandchildren (78, 79). Resulting false perceptions of clear that inadequately concealed or of smoking grandmothers experience reduced harm (80) have made it difficult to nonblinded trials exaggerate intervention increased asthma prevalence during alter public behavior once suggestions of effect estimates. For this reason, it is childhood (70, 71). Compared with paradoxically increased harms came to light incumbent on the authors of harm continued tobacco use, epidemiologic (2). Scientific claims are particularly prone reduction claims to disclose potential evaluation of potential harm reduction to exaggeration (81, 82), often accompanied effects of methodological bias on the strategies among pregnant smoking women by press releases that can readily reach the claimed benefit, including reporting have suggested similar increases in preterm public, and may result in false perceptions sensitivity analyses restricted to data births (72), small for gestational age births or misunderstanding (83, 84). The importance generated in trials at low risk of bias (63). (73), and early neonatal mortality and of avoiding overreaching conclusions In all cases, harm reduction claims should stillbirths (74). Given that the underlying cannot be overstated. be accompanied by conflict of interest mechanisms for these findings have yet to Product-oriented population harm disclosures. be worked out, and at least a portion of the reduction claims should account for all It should not be assumed that harm reproductive harms appear to be caused by known, albeit perhaps unintended, uses or reduction claims are generalizable to all age nicotine (75, 76), conclusions regarding modifications of the product in the modeling groups. Harm reduction claims should impact on reproductive health should be or impact calculations, particularly when specifically state the age cohort to which they considered independently from similar estimating impact on tobacco-naive users. most ideally apply. strategies aimed at the general population. The observed uses of products Potential harms of tobacco smoke Consideration of population-based harm intended to reduce the harm of tobacco exposure can vary substantially by population reduction strategies should include smoke exposure often diverge from the subgroup. Child and adolescent brains consideration of impact on reproductive expected. Accessorization, disassembly, or have increased susceptibility to the health. combination with other products may not development of nicotine addiction be intended by the manufacturer, but use (64). A product that may reduce harms Careful Use of Language in Harm patterns that have significant prevalence for adults who are tobacco users, but Reduction Claims may have significant public health impact paradoxically increase risk of future tobacco Harm reduction claims should be presented and influence on harm reduction estimates use among children, may be seen as a in lay language. (85, 86). Even misapplication of intended net overall increase in population harm, As scientific content becomes uses, such as flavorant customization, vape given the relative societal implications increasingly available online, journal editors cloud competitions, direct ingestion, or of unintentionally promoting tobacco frequently require summaries that are excessive voltage adjustment of electronic use. Because adolescent development is written specifically for the lay public. nicotine delivery devices, may directly or characterized by risk-taking behaviors and Nowhere is this requirement more likely to inadvertently increase harms to tobacco-naive e98 American Journal of Respiratory and Critical Care Medicine Volume 198 Number 8 | October 15 2018

AMERICAN THORACIC SOCIETY DOCUMENTS users (87). Particularly because flavored ethically be estimated against the However, designs that allow only for products appear to be targeted to inactive comparator because of the strong abatement of harmful exposures when adolescents and young adults, harm possibility of introducing comparator bias harm-eliminating interventions exist may reduction claims should account for uses (90, 91). All harm reduction claims should warrant reevaluation in light of the that may have nothing to do with the make their methodology clear in these instructive lessons the lead paint harm reduction or cessation of tobacco use. respects. reduction study offers (95). Given the evidence establishing the Harm reduction researchers must efficacy and safety of pharmacologic harm Ethical Considerations When Claiming protect the autonomy of research subjects reduction methods for people unready or Harm Reduction by discussing the unique nature of the unwilling to make a quit attempt, marginal Hazards of prospective population-based risks of such research with those subjects harm reduction calculations should use the evaluations of tobacco harm reduction and by promoting their individual best available harm reduction approach strategies may qualify as “greater than understanding. as the comparator rather than the no- minimal risk.” Care should be taken to Autonomy is respected in human intervention condition. In all cases, harm assess research risks in comparison to subject research when participants have the reduction claims should be accompanied current standards of care for reducing harm capacity to understand the risks they are by clear statements of the methodologic among people unready or unwilling to assuming, are given the information decisions made in this respect. quit, rather than in comparison to the necessary to achieve that understanding, The foundation for the ethical no-intervention condition. have their understanding verified by a investigation, implementation, and There are inherent uncertainties to all researcher qualified to make such an regulation of harm reduction strategies new potential harm reduction strategies. assessment, and consent voluntarily to rests in the strong fiduciary commitment to Evaluation methods are aimed at reasonably assume the risks of the research. The nature public health and human rights. Ethically reducing these uncertainties. Although and severity of the risks of the research sound consideration of potential harm patients, physicians, and even researchers need not be known in their entirety to be reduction strategies requires an honest may believe a particular harm reduction acceptable to the subject. However, the fact evaluation of the evidence supporting approach has value, our fiduciary that certain risks are unknown must be efficacy, assurance that the public responsibility is executed through judicious included in the informed consent discussion understands the relative risks and benefits, investigation using methods that maximize if the participant’s assumption of risk is to and a plan to closely monitor the long-term potential benefit while minimizing potential be considered valid. health and behavioral effects, including risks. Because safe and effective harm Unlike current standards of care for unintended consequences. In these reduction methods exist, new proposed cessation and harm reduction, the long-term pursuits, appropriate choice of comparator harm reduction methods may represent risks of harm reduction methods are not yet is of utmost importance to fulfilling the greater than minimal risk if they are fully known—a fact that must be disclosed fiduciary role. Even the simplest analysis of evaluated against the established standard, during the consent process in a manner benefit requires comparison of the expected rather than in addition to that standard that meets the needs of the particular baseline effect to the observed effect under (92). In population-based studies, justifying research subject. Human subjects may harm reduction conditions (DE). For such a deviation from the standard of be vulnerable to coercion if a condition circumstances in which no known remedy practice on the basis of cost or other exists that has compromised their judgment is available, it is acceptable to estimate systematic barriers may be unethical, regarding how the research applies to their DE using the status quo ante (i.e., no particularly if the population includes individual circumstances. Given the nature intervention) as comparator. However, in vulnerable members of the community of ongoing marketing efforts, it remains a circumstances where a proven, accepted (e.g., adolescents, the mentally ill, particular responsibility of harm reduction remedy for the status quo ante does exist, polysubstance abusers) or unfairly favors investigators to avoid overemphasizing it is no longer ethically acceptable to the privileged (e.g., those with greater-than- unproven benefits or minimizing unknown estimate DE using the baseline condition average pharmaceutical insurance coverage risks. as comparator unless informed consent and access to health care). Both the Harm reduction researchers should is assured (88). Rather, the marginal Maryland Court of Appeals and U.S. protect research subjects from unforeseen or effectiveness compared with standard Office of Human Research Protections tragic consequences of their research interventions should be used (mDE). An determinations in the Johns Hopkins study through fair methods that promote just analysis that declares the superiority of of lead paint harm reduction rejected the outcomes. a new intervention using a comparator position that risk research is less ethically Fairness is ensured when a uniform that is no longer in practice, or is no problematic when performed on people process is consistently applied. Ideally, that longer considered to be standard of who are already disadvantaged by ongoing process promotes justice, or the equitable care, is not appropriate for assessing exposure. In light of both the affective distribution of benefits and harms within a mDE (89). In light of the well-developed dysregulation characteristic of nicotine group of people who have assumed the same evidence in favor of a standard for using addiction and the obvious life-or-death set of risks. Fairness is a process, justice pharmacologic methods of treating the consequences of uncontrolled tobacco is the result. Justice in research demands compulsive disorder of dependence even dependence, it remains an open question as that harms are minimized as much as among people unready or unwilling to to whether current smokers inherently possible across a group of research subjects, make a quit attempt,mDEcannolonger qualify as a vulnerable population (93, 94). regardless of the degree to which they

American Thoracic Society Documents e99

AMERICAN THORACIC SOCIETY DOCUMENTS voluntarily assumed the risk of harm. To estimates (96). It is possible that clinicians investment in tobacco control policies and fairly mitigate harm, consistent methods for will face competing concerns that are the development of effective therapeutic detecting, reporting, and responding to new difficult to balance when implementing strategies have resulted in a remarkable harms identified during investigation are harm reduction strategies. For example, overall reduction in tobacco use prevalence. imperative. These methods should be how does the potential for reducing future Yet, we continue to be frustrated by expressly discussed during the informed cancer risk compete with the potential to persistent and unacceptably high rates consent process. In the interest of fairness, it impact the effectiveness of concurrent of tobacco use, especially among children, is important for investigators and regulatory medications today, or with the economic the poor, and the disenfranchised (2). agencies to note the potential biases that impact of uncovered costs of long-term Long-established traditions, based on a may arise from new intellectual or commercial interventions? Finally, regulators may theoretical framework that artificially conflicts of interest. evaluate harm reduction claims from a divorces free will from the biological Adopted harm reduction strategies population-based perspective, within which functions of the brain, have been implicated should guard against inequitable allocation variation in individual results may be as limits to our creativity in solving this of resources. superseded by concerns for the community’s problem (101, 102). The obdurate nature of Harm reduction strategies should be well-being or the legal/political environment tobacco dependence has understandably aimed at mitigating not only the harms within which harm reduction policies are resulted in low implicit estimations of the accrued to the individuals engaging in enacted (97). probability of altering its course, and the target activity but also the harms Harm reduction claims should be these low expectations naturally lead accrued to others who may be affected. accompanied by a narrative statement of to disengagement and learned The unintended social costs of the potential for unintended impact on “hopelessness” among clinicians (103, implementation should remain of concern. non–index users. 104). From this perspective, desperation Poorly implemented approaches may Although it is true that nonusers may appear to warrant advocating for disproportionately accrue harms or can be exposed to tobacco smoke through unproven approaches to reducing harm in benefits to subsets of the population in second- and third-hand mechanisms as thehopeofsavinglivesthatwould an unfair manner. In that regard, harm well as through accidental ingestion, the otherwisebelost(105).Itispreciselywhen reduction strategies should be tailored only appropriate reference condition for the stakes are so high that fidelity to to the specific needs of vulnerable evaluating the potential impact of harm scientific principles is most crucial to the populations. Those most likely to be reduction claims on the nonuser portion of public health (106). targeted and affected by unsubstantiated the population is the “absent” condition. Harm reduction strategies should harm reduction claims, including Secondhand exposure refers to inhalation never represent a capitulation to the low-income populations who suffer by a nonuser in proximity to the user, easier, cheaper, or more expedient options. disproportionate tobacco-related and thirdhand exposures related to Elimination of risk is, and should remain, health burdens, should be part of the toxicant exposure from surface deposits, the goal of public health authorities. This discussion designing reduction policies either through dermal absorption or statement is not intended to endorse that serve them. Involving a wide range off-gassing/inhalation (98). Accidental “harm reduction” as a compromise of stakeholders in the development and ingestion occurs when a product is ingested position but rather to highlight the special implementation of evidence-based harm or absorbed unintentionally, such as when a requirements for ethically pursuing the reduction strategies will encourage toddler child ingests it or spills it on his/her reduction of harm within the complex culturally and ethically acceptable skin in the course of exploring the world. social/biological context of tobacco implementation. Products that have the potential to appeal dependence. In all cases, improving Harm reduction claims should to a young child and/or may be easy to the dissemination of interventions specifically state the perspective(s) from access increase risk of accidental ingestion. that are already known to be safe and which they are derived. Reduced risk products that have substantial effective remains the primary harm When evaluating the validity of a appeal to the nonuser may also serve as a reduction tactic. specific harm reduction claim and the data gateway to the higher-risk products, thus This statement clarifies several on which the claim was built, it is easy to increasing their use (99). Because harm important notions intrinsic to the forget that, in its broadest sense, the topic reduction claims are typically aimed at thoughtful application of harm reduction of “harms of tobacco smoke” and their people who may also be influenced by the principles. First, reductions in tobacco potential for modification is protean and risk sustained by others in their community toxicant exposure are too often conflated likely to mean very different things to (e.g., children, coworkers, etc.), all such with reductions in the harms of tobacco use. independent evaluators approaching the claims should provide an accurate accounting We understand this to be a false equivalence topic from different perspectives. Individual of the known impact on non–index user risk on many levels. Individually, we understand patients may be affected by harm reduction profiles (100). the relationship between risk and exposure claims in ways that are not easily reflected to be nonlinear. On a population level, we in scientific analysis. Externalities such as understand risk to be unequally distributed preexisting conditions, family history, or Discussion between subgroups. Imagine a theoretical the social acceptability of the proposed circumstance in which reductions in harm reduction strategy are likely to be Tobacco dependence represents a vexing tobacco use are concentrated within the considered differently in individual utility paradox. Several decades of massive subset of a population less susceptible to the e100 American Journal of Respiratory and Critical Care Medicine Volume 198 Number 8 | October 15 2018

AMERICAN THORACIC SOCIETY DOCUMENTS outcome of interest, rendered so by the responsibly assess the impact of harm Patient Perspective distribution of important genetic, social, or reduction claims if dependence, and all environmental moderators of disease. its consequent ill effects, are trivialized or It sometimes feels like people who smoke We also understand the impact that ignored. cigarettes are the modern-day lepers. behavioral compensation has on exposure, It is hard to imagine an area of Marginalized, stigmatized, and constantly regardless of measured toxicant levels. Facile regulatory science more prone to reminded that smoking kills, people fi assumptions proved to be ecologic fallacies con rmation and availability biases. affected by nicotine dependence can spend in the past and should not be repeated. Recommendations guiding the appropriate their adult years caught between the Second, the statement highlights the structure, communication, and metaphorical rock and hard place. Each day complexity of weighing risk trade-offs in investigation of tobacco harm reduction presents another opportunity to quit, yet fi such a chaotic system. Given that bene ts claims offer regulators and scientists a series each thought of quitting is accompanied of a particular approach to harm reduction of methodologic remedies to this problem by an impulse to leave it alone until are likely to accrue unequally within a and represent a means of assuring the public after the next cigarette. Days, months, population, it remains imperative that that these threats to validity have been yearsgobywaitingforthenexthealth those of us wishing to make harm reduction checked to the extent possible. To maximize ’ fi complication. Some of us face lung cancer, claims be judicious in estimating the the public s con dence in harm reduction some respiratory infections, and some magnitude of that difference—across age claims, several additional steps should progressive disability and breathlessness. groups, between sexes, and over time. be taken. For example, the scientific All of us face the dreadful possibility that Because evidence-based approaches to community should develop a formal system our actions will eventually hurt someone helping people remediate their tobacco of evidence grading, specific to the context we love. These are the hidden costs of use risks already exist, including in of harm reduction work, that incorporates tobacco use. circumstances characterized by reluctance measures of study design, consistency of It stands to reason that people who to discontinue the smoking behavior, the results, directness of the evidence, and smoke would naturally be attracted to marginal effectiveness of newly introduced precision of measures into a summary products that claim to reduce the harms interventions should only be assessed grade, useful for communicating confidence of smoking while allowing for continued against the effectiveness of established in the claim (107, 108). Regulators must tobacco use. It also stands to reason approaches, not against the no- provide assurance that longitudinal intervention condition. Choosing the surveillance systems are in place to that people in such a position would be comparator most relevant to each document the impact of a harm reduction vulnerable to the effects of misinformation, population subgroup is necessary to intervention, inform the public of exaggeration, and outright hucksterism manage the significant potential for developments that might alter their when products making harm reduction comparator bias within an analysis and the confidence in the claim, and serve as an claims are introduced to the market. fi potential for misleading estimations of risk early warning system for unintended This ATS policy statement re nes the during the conduct of research. In all consequences (109). conversation considerably, bringing us “ ” cases, authors should favor careful fidelity The known influence of cognitive closer to a reliable set of rules for fi to the scientific method, to avoid both false biases also affects our ethical responsibilities engendering con dence and sorting fact fi claims of harm reduction and rejection of in study design. The moral construct of from ction. valid harm reduction. It bears repeating informed consent is the basis for the We expect clinicians to use their best that for tobacco-naive youth, the only participant’s assumption of risk. This judgment while helping us minimize our ethically acceptable comparators are the requirement elevates the researcher’s risks. Thus, we expect data informing their tobacco-less condition. responsibility for clarity and transparency, judgment to be reliable and clear. Finally, this statement challenges particularly when an unreasonable or By articulating the ethical responsibilities of the premise that the development or premature expectation of benefit might be scientists and government regulators when maintenance of dependence can be anticipated among participants. Fairness making harm reduction claims, the ATS separated from a discussion of the harms mandates that the risk of harm reduction statement helps guide the future standards to be reduced. We know that dependence interventions be assessed as “greater by which we assess and communicate real- behaviors stem from addiction—a than minimal” if the risk profile of the world risk to the community. The additional pathologic disruption of normal brain intervention is unclear, because, in this effort inherent to these recommendations biology. We also know that this disruption context, “risk” refers to the impact of is a price worth paying if it helps us avoid is associated with unhealthy or maladaptive study participation and is not assessed in the mistakes of the past and reduce the sequelae that extend well beyond whether comparison to the risk of ongoing tobacco impact of tobacco on our children and smoking continues. Policy makers cannot use. grandchildren. n

American Thoracic Society Documents e101

AMERICAN THORACIC SOCIETY DOCUMENTS

This official policy statement was prepared by an ad hoc subcommittee of the ATS Tobacco Action Committee.

Members of the subcommittee are as Healthcare System, Loma Linda University, Author Disclosures: K.-H.C. served as a follows: Loma Linda, California; 4University of speaker for Boehringer Ingelheim. F.C.D. received research support from Medela; and 1 California San Diego School of Medicine, FRANK T. LEONE, M.D., M.S. (Chair) served on a data and safety monitoring board ˚ 2 UCSD, and VA San Diego Healthcare KAI-HAKON CARLSEN, M.D., PH.D. System, San Diego, California; 5Yale School for Olympus. M.N.E. served as a speaker 3 and received personal fees from Praxis DAVID CHOOLJIAN, M.D., J.D. of Medicine, Yale University, New Haven, 4 6 Pharmaceuticals. S.M. reports being the LAURA E. CROTTY ALEXANDER, M.D. Connecticut; Pulmonary and Critical Care Senior Director, Ethics and Conflict of Interest 5 12 FRANK C. DETTERBECK, M.D. Medicine and Pediatric Pulmonology, Johns Policies, for the American Thoracic Society; he 6 MICHELLE N. EAKIN, M.A., PH.D. Hopkins School of Medicine, Baltimore, derives salary support for this role and as such 1 7 SARAH EVERS-CASEY, M.P.H. Maryland; Pediatric Pulmonology, Baylor has a personal financial interest in the quality of 7 American Thoracic Society management of AROLD ARBER College of Medicine and Texas Children’s H J. F , M.D. 8 8 Hospital, Houston, Texas; Center for conflicts of interest. D.P.L.S. served on an PATRICIA FOLAN, R.N., D.N.P. advisory committee and developed and 9 Tobacco Control, Northwell Health, Great HASMEENA KATHURIA, M.D. Neck, New York; 9Boston University provided an internal one-day professional 10 training seminar for Pfizer. F.T.L., D.C., L.E.C.A., KAREN LATZKA Medical Center, Boston University, 11 S.E.-C., H.J.F., P.F., H.K., K.L., S.M.-M., F.M., SHANE MCDERMOTT Boston, Massachusetts; 10Public Advisory 12 A.M., E.N., S.P., J.S., B.S., and D.U. reported SHARON MCGRATH-MORROW, M.B.A., M.D. Roundtable and 11Ethics and Conflicts of 13 no relationships with relevant commercial FARZAD MOAZED, M.D. Interest, American Thoracic Society, New interests. 14 13 ALFRED MUNZER, M.D. York, New York; Division of Pulmonary and 6 ENID NEPTUNE, M.D. Critical Care Medicine, University of California 15 Acknowledgment: The authors thank Dr. Brian SMITA PAKHALE, M.D., M.SC. San Francisco, San Francisco, California; 14 King, Deputy Director for Research Translation, 16 Pulmonary Medicine, Washington Adventist DAVID P. L. SACHS, M.D. 15 Centers for Disease Control, Office on Smoking 17 Hospital, Takoma Park, Maryland; Ottawa JONATHAN SAMET, M.D., M.S. and Health, for his contributions. Dr. King 10 Hospital Research Institute, University of provided the panel with expert advice BETH SUFIAN, J.D. Ottawa, Ottawa, Ontario, Canada; 16Palo Alto 18 during the critical formative phases of the DONA UPSON, M.D. Center for Pulmonary Disease Prevention, project, helping the workgroup find clarity Palo Alto, California; 17Colorado School of 1 within an otherwise chaotic topic. They also Perelman School of Medicine, University of Public Health, University of Colorado Denver, thank of Mr. Courtney White and Ms. Kimberly 18 Pennsylvania, Philadelphia, Pennsylvania; Aurora, Colorado; and New Mexico Lawrence (ATS staff) for implementation 2 Institute for Clinical Medicine and Oslo Veterans Affairs Health Care System, assistance and methodological guidance, University Hospital, University of Oslo, Oslo, University of New Mexico, Albuquerque, New without which this group effort would have Norway; 3Veterans Affairs Loma Linda Mexico. been impossible.

References [accessed 2017 May 2]. Available from: https://profiles.nlm.nih.gov/ NN/B/B/S/N/. 1. Leone FT, Carlsen K-H, Folan P, Latzka K, Munzer A, Neptune E, et al.; 8. CDC’sOffice on Smoking and Health. Smoking and tobacco use; ATS Tobacco Action Committee; American Thoracic Society. An Surgeon General’s reports: complete report. Smoking and tobacco Official American Thoracic Society research statement: current use. Center for Health Promotion and Education. Office on Smoking understanding and future research needs in tobacco control and and Health; 2004 [accessed 2017 Aug 4]. Available from: http://www. treatment. Am J Respir Crit Care Med 2015;192:e22–e41. cdc.gov/tobacco/data_statistics/sgr/2004/complete_report/. 2. National Center for Chronic Disease Prevention and Health Promotion 9. Kozlowski LT, Goldberg ME, Yost BA, White EL, Sweeney CT, Pillitteri (US) Office on Smoking and Health. The health consequences of JL. Smokers’ misperceptions of light and ultra-light cigarettes may smoking—50 years of progress: a report of the Surgeon General. keep them smoking. Am J Prev Med 1998;15:9–16. Atlanta, GA: Centers for Disease Control and Prevention (US); 2014 10. Davis RM, Novotny TE, Lynn WR, editors. The health consequences of [accessed 2017 Dec 13]. Available from: http://www.ncbi.nlm.nih. smoking: nicotine addiction: a report of the Surgeon General. Center gov/books/NBK179276/. for Health Promotion and Education. Office on Smoking and Health; 3. Gerber Y, Myers V, Goldbourt U. Smoking reduction at midlife and 1988 [accessed 2017 Aug 7]. Available from: http://profiles.nlm.nih. lifetime mortality risk in men: a prospective cohort study. Am J gov/NN/B/B/Z/D/. Epidemiol 2012;175:1006–1012. 11. Kozlowski LT, Strasser AA, Giovino GA, Erickson PA, Terza JV. Applying 4. Jia H, Lubetkin EI. Dose–response effect of smoking status on quality- the risk/use equilibrium: use medicinal nicotine now for harm reduction. adjusted life years among US adults aged 65 years and older. J Public Tob Control 2001;10:201–203. Health (Oxf) 2017;39:e194–e201. 12. Henningfield JE, Fagerstrom KO. Swedish Match Company, Swedish 5. Farone WA. Harm reduction: 25 years later. Tob Control 2002;11: snus and public health: a harm reduction experiment in progress? 287–288. Tob Control 2001;10:253–257. 6. Thun MJ, Burns DM. Health impact of “reduced yield” cigarettes: a 13. Foulds J, Ramstrom L, Burke M, Fagerstrom ¨ K. Effect of smokeless critical assessment of the epidemiological evidence. Tob Control tobacco (snus) on smoking and public health in Sweden. Tob Control 2001;10:i4–i11. 2003;12:349–359. 7. U.S. Department of Health and Human Services, Public Health Service, 14. Stratton K, Shetty P, Wallace R, Bondurant S. Clearing the smoke. Office on Smoking and Health. The health consequences of smoking: National Academies Press (US); 2001 [accessed 2017 Apr 30]. the changing cigarette: a report of the Surgeon General. Department Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/ of Health and Human Services publication No. (PHS) 81-50156; 1981 PMC1747561/pdf/v010p00189.pdf. e102 American Journal of Respiratory and Critical Care Medicine Volume 198 Number 8 | October 15 2018

AMERICAN THORACIC SOCIETY DOCUMENTS

15. Center for Addiction and Mental Health. Harm reduction [accessed 37. Ciura B. Philip Morris International’s new e-cigarette could help it grow. 2018 Sept 25]. Available from: https://ontario.cmha.ca/harm- The Motley Fool; 2014 [accessed 2017 May 3]. Available from: reduction/. https://www.fool.com/investing/general/2014/12/08/philip-morris- 16. Kirby D. Emerging answers: research findings on programs to reduce internationals-new-e-cigarette-could.aspx. teen pregnancy. Washington, DC, National Campaign To Prevent 38. Dewhirst T. Into the black: Marlboro brand architecture, packaging and Teen Pregnancy; 2001 [accessed 2017 May 3]. Available from: marketing communication of relative harm. Tobacco Control 2018; https://eric.ed.gov/?id=ED456171. 27:240–242. 17. Centers for Disease Control and Prevention (CDC). HIV, other STD, and 39. Sparks M. Building healthy public policy: don’t believe the misdirection. pregnancy prevention education in public secondary schools — 45 Health Promot Int 2011;26:259–262. states, 2008–2010. MMWR Morb Mortal Wkly Rep 2012;61:222–228. 40. Mercincavage M, Saddleson ML, Gup E, Halstead A, Mays D, Strasser 18. Aspinall EJ, Nambiar D, Goldberg DJ, Hickman M, Weir A, Van Velzen E, AA. Reduced nicotine content cigarette advertising: how false beliefs et al. Are needle and syringe programmes associated with a and subjective ratings affect smoking behavior. Drug Alcohol Depend reduction in HIV transmission among people who inject drugs: a 2017;173:99–106. systematic review and meta-analysis. Int J Epidemiol 2014;43: 41. Chapman S. The role of doctors in promoting smoking cessation. BMJ 235–248. 1993;307:518–519. 19. Abdul-Quader AS, Feelemyer J, Modi S, Stein ES, Briceno A, Semaan 42. Gruning ¨ T, Gilmore AB, McKee M. Tobacco industry influence on science S, et al. Effectiveness of structural-level needle/syringe programs to and scientists in Germany. Am J Public Health 2006;96:20–32. reduce HCV and HIV infection among people who inject drugs: a 43. National Research Council (US) Committee on the Institutional Means systematic review. AIDS Behav 2013;17:2878–2892. for Assessment of Risks to Public Health. Risk assessment in the 20. Potier C, Laprevote ´ V, Dubois-Arber F, Cottencin O, Rolland B. federal government: managing the process. Washington, D.C.: Supervised injection services: what has been demonstrated? A National Academies Press (US); 1983 [accessed 2017 May 5]. systematicliteraturereview.Drug Alcohol Depend 2014;145: Available from: http://www.ncbi.nlm.nih.gov/books/NBK216620/. 48–68. 44. National Research Council (US) Committee on Improving Risk Analysis 21. Jones L, Hughes K, Atkinson AM, Bellis MA. Reducing harm in drinking Approaches Used by the US EPA. Science and decisions: advancing environments: a systematic review of effective approaches. Health risk assessment. Washington, D.C.: National Academies Press (US); Place 2011;17:508–518. 2009 [accessed 2017 May 5]. Available from: http://www.ncbi.nlm. 22. Free C, Roberts IG, Abramsky T, Fitzgerald M, Wensley F. A systematic nih.gov/books/NBK214630/. review of randomised controlled trials of interventions promoting 45. Scherer G, Lee PN. Smoking behaviour and compensation: a review of effective condom use. J Epidemiol Community Health 2011;65: the literature with meta-analysis. Regul Toxicol Pharmacol 2014;70: 100–110. 615–628. 23. Zeller M, Hatsukami D; Strategic Dialogue on Tobacco Harm Reduction 46. Strasser AA, Souprountchouk V, Kaufmann A, Blazekovic S, Leone F, Group. The Strategic Dialogue on Tobacco Harm Reduction: a vision Benowitz NL, et al. Nicotine replacement, topography, and smoking and blueprint for action in the US. Tob Control 2009;18:324–332. phenotypes of e-cigarettes. Tob Regul Sci 2016;2:352–362. 24. Kozlowski LT. The truncation of moral reasoning on harm reduction by 47. Centers for Disease Control and Prevention (US), National Center for individuals and organizations. Addiction 2015;110:1070–1072. Chronic Disease Prevention and Health Promotion (US), Office on 25. Sarewitz D. Allow use of electronic cigarettes to assess risk. Nature Smoking and Health (US). How tobacco smoke causes disease: the 2014;512:349. biology and behavioral basis for smoking-attributable disease: a 26. Kozlowski LT. Harm reduction, public health, and human rights: report of the Surgeon General. Atlanta, GA: Centers for Disease smokers have a right to be informed of significant harm reduction Control and Prevention (US); 2010 [accessed 2014 Apr 14]. Available options. Nicotine Tob Res 2002;4:S55–S60. from: http://www.ncbi.nlm.nih.gov/books/NBK53017/. 27. Kozlowski LT, Edwards BQ. “Not safe” is not enough: smokers have 48. Benowitz NL. Nicotine addiction. N Engl J Med 2010;362:2295– a right to know more than there is no safe tobacco product. Tob 2303. Control 2005;14:ii3–ii7. 49. Fiore MC, Jaen ´ CR, Baker TB, Bailey WC, Benowitz NL, Curry SJ, et al. 28. Douglas HE. The moral responsibilities of scientists: tensions between Treating tobacco use and dependence: 2008 update. Clinical autonomy and responsibility. American Philosophical Quarterly; Practice Guideline. Rockville, MD: US Department of Health and 2012 [accessed 2017 May 3]. Available from: http://www.academia.edu/ Human Services. Public Health Service; 2008. 987446/The_moral_responsibilities_of_scientists_tensions_between_ 50. Goriounova NA, Mansvelder HD. Short- and long-term consequences autonomy_and_responsibility_. of nicotine exposure during adolescence for prefrontal cortex 29. ICSU Committee on Freedom and Responsibility in the conduct of neuronal network function. Cold Spring Harb Perspect Med 2012;2: Science (CFRS). Advisory note: “bias in science publishing” [accessed a012120. 2018 Sept 25]. Available from: https://council.science/cms/2017/04/ 51. Counotte DS, Spijker S, Van de Burgwal LH, Hogenboom F, ICSU_CFRS_Advisory_Note_Bias_in_Science_Publishing.pdf. Schoffelmeer ANM, De Vries TJ, et al. Long-lasting cognitive 30. Pain E. The social responsibilities of scientists. Science, AAAS; 2013 deficits resulting from adolescent nicotine exposure in rats. [accessed 2017 May 3]. Available from: http://www.sciencemag. Neuropsychopharmacology 2009;34:299–306. org/careers/2013/02/social-responsibilities-scientists. 52. Leslie FM, Loughlin SE, Wang R, Perez L, Lotfipour S, Belluzzia JD. 31. Lindson-Hawley N, Hartmann-Boyce J, Fanshawe TR, Begh R, Farley Adolescent development of forebrain stimulant responsiveness: A, Lancaster T. Interventions to reduce harm from continued tobacco insights from animal studies. Ann N Y Acad Sci 2004;1021: use. Cochrane Database Syst Rev 2016;10:CD005231. 148–159. 32. Schuman LM. Patterns of smoking behavior. NIDA Res Monogr 1977; 53. Murthy VH. E-cigarette use among youth and young adults: a major 17:36–66. public health concern. JAMA Pediatr 2017;171:209–210. 33. RICO convictions of major tobacco companies affirmed. Washington, 54. Kandel ER, Kandel DB. Shattuck lecture: a molecular basis for nicotine D.C., The Daily Washington Law Reporter; 2011 [accessed 2017 as a gateway drug. N Engl J Med 2014;371:932–943. May 5]. Available from: http://www.dwlr.com/blog/2011-05-12/rico- 55. Leone FT, Evers-Casey S. Developing a rational approach to tobacco convictions-major-tobacco-companies-affirmed. use treatment in pulmonary practice: a review of the biological basis 34. Gendreau PL, Vitaro F. The unbearable lightness of “light” cigarettes: of nicotine addiction. Clin Pulm Med 2012;19:53–61. a comparison of smoke yields in six varieties of Canadian “light” 56. Robinson WS. Ecological correlations and the behavior of individuals. cigarettes. Can J Public Health 2005;96:167–172. Int J Epidemiol 2009;38:337–341. 35. Anonymous. 2016 Annual meeting of shareholders. Altria; 2016 57. Wakefield J. Multi-level modelling, the ecologic fallacy, and hybrid [accessed 2017 May 3]. Available from: http://media.corporate-ir. study designs. Int J Epidemiol 2009;38:330–336. net/media_files/IROL/80/80855/2016ASM/Presentation.pdf. 58. Kannel WB. Some lessons in cardiovascular epidemiology from 36. Knapp DE. Brand equity. Risk Management 1999;46:71–74. Framingham. Am J Cardiol 1976;37:269–282.

American Thoracic Society Documents e103

AMERICAN THORACIC SOCIETY DOCUMENTS

59. Lloyd-Jones DM, O’Donnell CJ, D’Agostino RB, Massaro J, Silbershatz 80. O’Connor RJ, Bansal-Travers M, Cummings KM, Hammond D, H, Wilson PWF. Applicability of cholesterol-lowering primary Thrasher JF, Tworek C. Filter presence and tipping paper color prevention trials to a general population: the Framingham Heart influence consumer perceptions of cigarettes. BMC Public Health Study. Arch Intern Med 2001;161:949–954. 2015;15:1279. 60. Faden R, Shebaya S. Public health ethics. Stanford Encyclopedia of 81. Woloshin S, Schwartz LM, Casella SL, Kennedy AT, Larson RJ. Press Philosophy. 2016 [accessed 2017 May 7]. Available from: https://plato. releases by academic medical centers: not so academic? Ann Intern stanford.edu/archives/win2016/entries/publichealth-ethics/. Med 2009;150:613–618. 61. Hammond D, Fong GT, Zanna MP, Thrasher JF, Borland R. Tobacco 82. Sumner P, Vivian-Griffiths S, Boivin J, Williams A, Venetis CA, Davies A, denormalization and industry beliefs among smokers from four et al. The association between exaggeration in health related science countries. Am J Prev Med 2006;31:225–232. news and academic press releases: retrospective observational 62. Hamby DM. A review of techniques for parameter sensitivity study. BMJ 2014;349:g7015. analysis of environmental models. Environ Monit Assess 1994;32: 83. Etter J-F, Kozlowski LT, Perneger TV. What smokers believe about light 135–154. and ultralight cigarettes. Prev Med 2003;36:92–98. 63. Wood L, Egger M, Gluud LL, Schulz KF, Juni ¨ P, Altman DG, et al. 84. O’Connor RJ, McNeill A, Borland R, Hammond D, King B, Boudreau C, Empirical evidence of bias in treatment effect estimates in controlled et al. Smokers’ beliefs about the relative safety of other tobacco trials with different interventions and outcomes: meta-epidemiological products: findings from the ITC collaboration. Nicotine Tob Res study. BMJ 2008;336:601–605. 2007;9:1033–1042. 64. Chen J, Millar WJ. Age of smoking initiation: implications for quitting. 85. Krishnan-Sarin S, Morean M, Kong G, Bold KW, Camenga DR, Cavallo “ ” Health Rep 1998;9:39–46. DA, et al. E-cigarettes and dripping among high-school youth. 65. Savitz DA, Meyer RE, Tanzer JM, Mirvish SS, Lewin F. Public health Pediatrics 2017;139:e20163224. “ implications of smokeless tobacco use as a harm reduction strategy. 86. Talih S, Balhas Z, Salman R, Karaoghlanian N, Shihadeh A. Direct ” Am J Public Health 2006;96:1934–1939. dripping : a high-temperature, high-formaldehyde emission electronic – 66. Anderson HR, Cook DG. Passive smoking and sudden infant death cigarette use method. Nicotine Tob Res 2016;18:453 459. syndrome: review of the epidemiological evidence. Thorax 1997;52: 87. Bartschat S, Mercer-Chalmers-Bender K, Beike J, Rothschild MA, 1003–1009. J¨ubner M. Not only smoking is deadly: fatal ingestion of e-juice-a case report. Int J Legal Med 2015;129:481–486. 67. Elliot J, Vullermin P, Robinson P. Maternal cigarette smoking is 88. Emanuel EJ, Wendler D, Grady C. What makes clinical research ethical? associated with increased inner airway wall thickness in children who JAMA 2000;283:2701–2711. die from sudden infant death syndrome. Am J Respir Crit Care Med 89. National Institutes of Health. HTA 101: V. Economic analysis methods 1998;158:802–806. [accessed 2018 Sept 25]. Available from: https://www.nlm.nih. 68. Haland˚ G, Carlsen KCL, Sandvik L, Devulapalli CS, Munthe-Kaas MC, gov/nichsr/hta101/ta10107.html. Pettersen M, et al.; ORAACLE. Reduced lung function at birth and 90. Asfar T, Ebbert JO, Klesges RC, Relyea GE. Do smoking reduction the risk of asthma at 10 years of age. N Engl J Med 2006;355: interventions promote cessation in smokers not ready to quit? Addict 1682–1689. Behav 2011;36:764–768. 69. Lødrup Carlsen KC, Mowinckel P, Hovland V, Haland˚ G, Riiser A, 91. Mann H, Djulbegovic B. Comparator bias: why comparisons must Carlsen K-H. Lung function trajectories from birth through puberty address genuine uncertainties. JLL Bulletin [accessed 2017 May 10]. reflect asthma phenotypes with allergic comorbidity. J Allergy Clin Available from: http://www.jameslindlibrary.org/articles/comparator- Immunol 2014;134:917–923, e7. bias-why-comparisons-must-address-genuine-uncertainties/. 70. Li YF, Langholz B, Salam MT, Gilliland FD. Maternal and grandmaternal 92. Emanuel EJ, Grady CC, Crouch RA, Lie RK, Miller FG, Wendler DD, smoking patterns are associated with early childhood asthma. Chest editors. The Oxford Textbook of Clinical Research Ethics. Reprint – 2005;127:1232 1241. edition. Oxford: Oxford University Press; 2011. ˚ 71. Magnus MC, Haberg SE, Karlstad Ø, Nafstad P, London SJ, Nystad W. 93. Cheetham A, Allen NB, Yucel ¨ M, Lubman DI. The role of affective ’ Grandmother s smoking when pregnant with the mother and asthma dysregulation in drug addiction. Clin Psychol Rev 2010;30: in the grandchild: the Norwegian Mother and Child Cohort Study. 621–634. – Thorax 2015;70:237 243. 94. Macklin R. Bioethics, vulnerability, and protection. Bioethics. 2003 fl 72. Baba S, Wikstrom ¨ A-K, Stephansson O, Cnattingius S. In uence of [accessed 2017 Aug 20]. Available from: http://www.readcube.com/ smoking and snuff cessation on risk of preterm birth. Eur J Epidemiol articles/10.1111/1467-8519.00362. – 2012;27:297 304. 95. Lewin T. US investigating Johns Hopkins study of lead paint hazard. 73. Baba S, Wikstrom ¨ A-K, Stephansson O, Cnattingius S. Changes in The New York Times. 2001 Aug 24 [accessed 2017 May 10]. snuff and smoking habits in Swedish pregnant women and Available from: http://www.nytimes.com/2001/08/24/us/us-investigating- risk for small for gestational age births. BJOG 2013;120: johns-hopkins-study-of-lead-paint-hazard.html. – 456 462. 96. Cameron J. WHO. Valuing water, valuing livelihoods. Chapter 11. IWA 74. Baba S, Wikstrom ¨ A-K, Stephansson O, Cnattingius S. Influence Publishing; 2011 [accessed 2017 May 6]. Available from: https://www. of snuff and smoking habits in early pregnancy on risks for scribd.com/document/272337704/Valuing-Water-Valuing-Livehoods. stillbirth and early neonatal mortality. Nicotine Tob Res 2014;16: 97. U.S. Food and Drug Administration. Rules, regulations & guidance. 78–83. Family smoking prevention and tobacco control act: an overview 75. Wongtrakool C, Roser-Page S, Rivera HN, Roman J. Nicotine alters lung [accessed 2018 Sept 25]. Available from: https://www.fda.gov/ branching morphogenesis through the a7 nicotinic acetylcholine tobaccoproducts/labeling/rulesregulationsguidance/ucm246129. receptor. Am J Physiol Lung Cell Mol Physiol 2007;293:L611– htm. L618. 98. Farber HJ, Groner J, Walley S, Nelson K; Section on Tobacco Control. 76. Wongtrakool C, Wang N, Hyde DM, Roman J, Spindel ER. Prenatal Protecting children from tobacco, nicotine, and tobacco smoke. nicotine exposure alters lung function and airway geometry through Pediatrics 2015;136:e1439–e1467. a7 nicotinic receptors. Am J Respir Cell Mol Biol 2012;46:695– 99. Wills TA, Sargent JD, Gibbons FX, Pagano I, Schweitzer R. E-cigarette 702. use is differentially related to smoking onset among lower risk 77. Dube ´ CE, Lapane KL. Lay abstracts and summaries: writing advice for adolescents. Tob Control 2016;26:534–539. scientists. J Cancer Educ 2014;29:577–579. 100. Vallerand RJ, Toward A. Hierarchical model of intrinsic and extrinsic 78. Harris B. The intractable cigarette ‘filter problem’. Tob Control 2011;20: motivation. Adv Exp Soc Psychol 1997;29:271–360. i10–i16. 101. Brady JV, Lukas SE, editors. Testing drugs for physical dependence 79. Pauly JL, Mepani AB, Lesses JD, Cummings KM, Streck RJ. Cigarettes potential and abuse liability. Department of Health and Human with defective filters marketed for 40 years: what Philip Morris never Services; 1984 [accessed 2018 Sept 25]. Available from: https:// told smokers. Tob Control 2002;11:I51–I61. archives.drugabuse.gov/sites/default/files/monograph52.pdf. e104 American Journal of Respiratory and Critical Care Medicine Volume 198 Number 8 | October 15 2018

AMERICAN THORACIC SOCIETY DOCUMENTS

102. Dackis C, O’Brien C. Neurobiology of addiction: treatment and 106. Leone FT, Douglas IS. The emergence of e-cigarettes: a triumph public policy ramifications. Nat Neurosci 2005;8:1431– of wishful thinking over science. Ann Am Thorac Soc 2014;11:216–219. 1436. 107. Guyatt GH, Oxman AD, Kunz R, Vist GE, Falck-Ytter Y, Schunemann ¨ 103. Leone FT, Evers-Casey S, Graden S, Schnoll R. Behavioral economic HJ; GRADE Working Group. What is “quality of evidence” and why insights into physician tobacco treatment decision-making. Ann Am is it important to clinicians? BMJ 2008;336:995–998. Thorac Soc 2015;12:364–369. 108. Schunemann ¨ HJ, Oxman AD, Brozek J, Glasziou P, Jaeschke R, Vist 104. Leone FT, Evers-Casey S, Graden S, Schnoll R, Mallya G. Academic GE, et al.; GRADE Working Group. Grading quality of evidence and detailing interventions improve tobacco use treatment among strength of recommendations for diagnostic tests and strategies. physicians working in underserved communities. Ann Am Thorac BMJ 2008;336:1106–1110. Soc 2015;12:854–858. 109. World Health Organization. Public health surveillance [accessed 105. E-cigarettes. The BMJ [accessed 2018 Sept 25]. Available from: 2018 Sept 25]. Available from: http://www.who.int/immunization/ http://www.bmj.com/campaign/e-cigarettes. monitoring_surveillance/burden/vpd/en/.

American Thoracic Society Documents e105