# Required Warnings for Cigarette Packages and Advertisements

> Briefs, arguments, decisions, and more.

URL: https://www.frixlaw.com/law-library/documents/fr%3A2011-15337

## Record

- **Collection:** Federal Register
- **Document type:** Rule
- **Published:** June 22, 2011
- **Citation:** 76 FR 36628

## Text

DEPARTMENT OF HEALTH AND HUMAN SERVICES
Food and Drug Administration
21 CFR Part 1141
[Docket No. FDA-2010-N-0568]
RIN 0910-AG41
Required Warnings for Cigarette Packages and Advertisements

AGENCY:

Food and Drug Administration, HHS.

ACTION:

Final rule.

SUMMARY:

The Food and Drug Administration (FDA) is amending its regulations to add a new requirement for the display of health warnings on cigarette packages and in cigarette advertisements. This rule implements a provision of the Family Smoking Prevention and Tobacco Control Act (Tobacco Control Act) that requires FDA to issue regulations requiring color graphics, depicting the negative health consequences of smoking, to accompany the nine new textual warning statements required under the Tobacco Control Act. The Tobacco Control Act amends the Federal Cigarette Labeling and Advertising Act (FCLAA) to require each cigarette package and advertisement to bear one of nine new textual warning statements. This final rule specifies the color graphic images that must accompany each of the nine new textual warning statements.

DATES:

This rule is effective September 22, 2012. See section VIII of this document,
Implementation Date,
for additional information. The incorporation by reference of a certain publication listed in the rule is approved by the Director of the Federal Register as of September 22, 2012.

FOR FURTHER INFORMATION CONTACT:

Gerie Voss or Kristin Davis, Center for Tobacco Products, Food and Drug Administration, 9200 Corporate Blvd., Rockville, MD 20850-3229, 877-287-1373,
gerie.voss@fda.hhs.gov
or
kristin.davis@fda.hhs.gov.

SUPPLEMENTARY INFORMATION:

Table of Contents

I. Introduction

A. Purpose and Overview

B. Background

II. Need for the Rule and Responses to Comments

A. Cigarette Use in the United States and the Resulting Health Consequences

1. Smoking Prevalence and Initiation in the United States

2. Health Consequences of Smoking

B. Inadequacy of Existing Warnings

C. Consumers' Lack of Knowledge of the Health Risks

D. Larger, Graphic Warnings Communicate More Effectively

E. Need To Refresh Required Warnings

III. FDA's Selection of Color Graphic Images

A. Methodology for Selecting Images

B. FDA's Research Study

1. Study Design

2. Use of FDA's Study Results in Selection of Images

3. Comments on FDA's Research Study

C. Comments to the Docket

1. Comments Submitting Research on FDA's Proposed Required Warnings

2. Other Comments

D. Selected Images

1. “WARNING: Cigarettes are addictive”

2. “WARNING: Tobacco smoke can harm your children”

3. “WARNING: Cigarettes cause fatal lung disease”

4. “WARNING: Cigarettes cause cancer”

5. “WARNING: Cigarettes cause strokes and heart disease”

6. “WARNING: Smoking during pregnancy can harm your baby”

7. “WARNING: Smoking can kill you”

8. “WARNING: Tobacco smoke causes fatal lung disease in nonsmokers”

9. “WARNING: Quitting smoking now greatly reduces serious risks to your health”

10. Image for Advertisements With a Small Surface Area

E. Non-Selected Images

1. “WARNING: Cigarettes are addictive”

2. “WARNING: Tobacco smoke can harm your children”

3. “WARNING: Cigarettes cause fatal lung disease”

4. “WARNING: Cigarettes cause cancer”

5. “WARNING: Cigarettes cause strokes and heart disease”

6. “WARNING: Smoking during pregnancy can harm your baby”

7. “WARNING: Smoking can kill you”

8. “WARNING: Tobacco smoke causes fatal lung disease in nonsmokers”

9. “WARNING: Quitting smoking now greatly reduces serious risks to your health”

10. Image for Advertisements With a Small Surface Area

IV. Comments Regarding Textual Warning Statements

A. Changes to Textual Warning Statements

B. Attribution to the Surgeon General

C. Foreign Language Translations

V. Description of the Final Rule

A. Overview of the Final Rule

B. Description of Final Regulations and Responses to Comments

1. Section 1141.1—Scope

2. Section 1141.3—Definitions

3. Section 1141.10—Required Warnings

4. Section 1141.12—Incorporation by Reference of Required Warnings

5. Section 1141.14—Misbranding of Cigarettes

6. Section 1141.16—Disclosures Regarding Cessation

VI. Comments Regarding Implementation Issues

A. Technical Issues Regarding Compliance

B. Textual Statement Color Formats

C. Random Display and Rotation of Warnings

VII. Legal Authority and Responses to Comments

A. FDA's Legal Authority

B. First Amendment Commercial Speech Issues

C. Takings Under the Fifth Amendment

VIII. Implementation Date

IX. Federalism

X. Environmental Impact

XI. Analysis of Impacts

A. Introduction and Summary

B. Comments on the Preliminary Regulatory Impact Analysis

1. General

2. Need for the Rule

3. Benefits

4. Costs

5. Distributional Effects

6. Impact on Small Entities

C. Need for the Rule

D. Benefits

1. Reduced Cigarette Smoking Rates

2. Quantifying Benefits That Accrue to Dissuaded Smokers

3. Reduced Fire Costs

4. Summary of Benefits

E. Costs

1. Number of Affected Entities

2. Costs of Changing Cigarette Labels

3. Ongoing Costs of Equal and Random Display

4. Market Testing Costs Associated With Changing Cigarette Package Labels

5. Advertising Restrictions: Removal of Noncompliant Point-of-Sale Advertising

6. Government Administration and Enforcement Costs

7. Summary of Costs

F. Cost-Effectiveness Analysis

G. Distributional Effects

1. Tobacco Manufacturers, Distributors, and Growers

2. National and Regional Employment Patterns

3. Retail Sector

4. Advertising Industry

5. Excise Tax Revenues

6. Government-Funded Medical Services, Insurance Premiums, and Social Security

H. International Effects

I. Regulatory Alternatives

1. 24-Month Compliance Period

2. 6-Month Compliance Period

3. Alternative Graphic Images

4. Summary of Regulatory Alternatives

J. Impact on Small Entities

1. Description and Number of Affected Small Entities

2. Description of the Potential Impacts of the Final Rule on Small Entities

3. Alternatives to Minimize the Burden on Small Entities

XII. Paperwork Reduction Act of 1995

XIII. References

I. Introduction

A. Purpose and Overview

The Tobacco Control Act was enacted on June 22, 2009, amending the Federal Food, Drug, and Cosmetic Act (FD&C Act) and FCLAA, and providing FDA with the authority to regulate tobacco products (Pub. L. 111-31; 123 Stat.

1776). Section 201 of the Tobacco Control Act modifies section 4 of FCLAA (15 U.S.C. 1333) to require that the following nine new health warning statements appear on cigarette packages and in cigarette advertisements:

• WARNING: Cigarettes are addictive

• WARNING: Tobacco smoke can harm your children

• WARNING: Cigarettes cause fatal lung disease

• WARNING: Cigarettes cause cancer

• WARNING: Cigarettes cause strokes and heart disease

• WARNING: Smoking during pregnancy can harm your baby

• WARNING: Smoking can kill you

• WARNING: Tobacco smoke causes fatal lung disease in nonsmokers

• WARNING: Quitting smoking now greatly reduces serious risks to your health.

Section 201 of the Tobacco Control Act also states that “the Secretary [of Health and Human Services] shall issue regulations that require color graphics depicting the negative health consequences of smoking” to accompany the nine new health warning statements.

As discussed in the preamble to the proposed rule (75 FR 69524 at 69525, November 12, 2010), cigarette smoking kills an estimated 443,000 Americans each year, most of whom began smoking when they were under the age of 18 (Ref. 1). Tobacco use is the foremost preventable cause of premature death in the United States, and has been shown to cause cancer, heart disease, lung disease, and other serious adverse health effects (Ref. 2). The U.S. Government has a substantial interest in reducing the number of Americans, particularly children and adolescents, who use cigarettes and other tobacco products in order to prevent the life-threatening health consequences associated with tobacco use (section 2(31) of the Tobacco Control Act).

Although FCLAA has required the inclusion of textual health warnings on cigarette packages and in cigarette advertisements for many years, there is considerable evidence, which was presented in the preamble to the proposed rule (75 FR 69524 at 69529 through 69531) and is discussed in section II.B of this document, that the existing cigarette health warnings are given little attention or consideration by viewers. A 2007 report from the Institute of Medicine (IOM) described the warnings as “invisible” (Ref. 3), and found that they fail to communicate relevant information in an effective way. The warnings currently in use in the United States also fail to include any graphic component, despite the evidence in the scientific literature that larger, graphic health warnings promote greater understanding of the health risks of smoking and would help to reduce consumption (
see
75 FR 69524 at 69531 through 69533). In proposing this regulation and preparing this final rule, we found substantial evidence indicating that larger cigarette health warnings including a graphic component, like those being required in this rule, would offer significant health benefits over the existing warnings. Consistent with Executive Order 13563, this regulation is “based on the best available evidence” and has allowed “for public participation and an open exchange of ideas.”

B. Background

On November 12, 2010, as directed by section 201 of the Tobacco Control Act and in the interest of public health, we issued a proposed rule seeking to modify the warnings that appear on cigarette packages and in cigarette advertisements to include color graphic images depicting the negative health consequences of smoking; these images were proposed to accompany the nine new textual warning statements set forth in section 201 of the Tobacco Control Act (
see
75 FR 69524). The Agency received more than 1,700 comments to the docket for the November 12, 2010, notice of proposed rulemaking (NPRM) on required warnings for cigarette packages and advertisements. Comments were received from cigarette manufacturers, retailers and distributors, industry associations, health professionals, public health or other advocacy groups, academics, State and local public health agencies, medical organizations, individual consumers, and other submitters. These comments are summarized and responded to in the relevant section(s) of this document. Similar comments are grouped together by the topics discussed or the particular portions of the NPRM or codified language to which they refer.

To make it easier to identify comments and FDA's responses, the word “Comment,” in parenthesis, appears before the comment's description, and the word “Response,” in parenthesis, appears before FDA's response. Each comment is numbered to help distinguish among different comments. Similar comments are grouped together under the same comment number. The number assigned to each comment is purely for organizational purposes and does not signify the comment's value or importance or the order in which it was received.

II. Need for the Rule and Responses to Comments

A. Cigarette Use in the United States and the Resulting Health Consequences

1. Smoking Prevalence and Initiation in the United States

In explaining the need for the proposed rule, we provided information in the NPRM on smoking prevalence and initiation rates among adults and children in the United States. As stated in the NPRM (75 FR 69524 at 69526), approximately 46.6 million U.S. adults (or 20.6 percent of the adult population) are cigarette smokers (Ref. 4). Moreover, almost half (46.3 percent) of youth in grades 9 through 12 in the United States have tried cigarette smoking, and 19.5 percent of youth in grades 9 through 12 are current cigarette smokers (Ref. 5 at p. 10). Smoking rates among U.S. adults have shown virtually no change during the 5-year period from 2005 to 2009 (Ref. 4), and smoking rates among U.S. youth have not decreased from 2006 to 2009 (Ref. 6).

Furthermore, each year millions of U.S. adults and children become new smokers. Data from the 2008 National Survey on Drug Use and Health indicate that 2.4 million persons aged 12 or older in the United States smoked cigarettes for the first time in the past 12 months (Ref. 7 at p. 59). In addition, these data indicate that almost 1 million Americans aged 12 or older started smoking cigarettes daily within the past 12 months (Ref. 7 at p. 60).

In other words, approximately 6,600 people aged 12 or older in the United States become new cigarette smokers every day, and more than 2,500 individuals become new daily cigarette smokers every day (Ref. 7 at pp. 59-60). Moreover, nearly 4,000 of the people who become new cigarette smokers every day and nearly 1,000 of the individuals who become new daily cigarette smokers every day are children under the age of 18 (Ref. 7 at pp. 59-60). These statistics for youth smokers are particularly concerning, as studies suggest that the age people start smoking can greatly influence how much they smoke per day and how long they smoke, which in turn influences their risk of tobacco-related disease and death (Refs. 8, 9, and 10).

FDA received many comments that were strongly supportive of the proposed rule, some of which provided data and information consistent with that in the NPRM regarding cigarette use prevalence and initiation in the United States (75 FR 69524 at 69526 through 69527). Many of these comments also stated that smokers would be more

likely to quit smoking and that nonsmokers would be less likely to start smoking if cigarette advertisements and packages display, visually and graphically, the health effects of cigarettes. Most of these comments expressed a belief that the required warnings would help reduce the existing and future use of cigarettes. Some comments that were supportive of the proposed rule discussed the smoking prevalence and initiation rates in the United States in particular populations. These comments, and FDA's responses, are summarized in the following paragraphs.

(Comment 1) Multiple comments indicated that people with less education and lower incomes have higher smoking prevalence rates in general. One comment from a health care association indicated that women of low educational background have higher smoking prevalence rates and that many of these women still are not aware of cigarettes' impact on life expectancy, heart disease, and pregnancy.

(Response) We agree that adults with low education levels have higher than average smoking prevalence rates. For example, as discussed in the NPRM (75 FR 69524 at 69526), 49.1 percent of adults with a General Education Development certificate (GED) and 28.5 percent of adults with less than a high school diploma were current smokers in 2009, compared with 5.6 percent of adults with a graduate degree (Ref. 4). We also agree that graphic health warnings may be particularly important communication tools for these smokers, as there is evidence suggesting that countries with graphic health warnings demonstrate fewer disparities in health knowledge across educational levels (Ref. 11 at p. 18 and Ref. 3 at p. 295).

(Comment 2) Multiple comments noted that smoking rates vary by race and ethnicity, with American Indians/Alaska Natives having the highest rates. One comment also noted that the health and economic costs of smoking vary by race and ethnicity. For example, the comment stated that African-American smokers suffer disproportionately from smoking-related diseases, including lung cancer, heart disease, and strokes (
citing
Ref. 12), and called for measures to address these disparities.

One comment from a State public health agency indicated that racial minority populations and economically disadvantaged populations have smoking prevalence rates that are two to three times higher than the general population.

(Response) We agree that smoking rates vary by race and ethnicity and socioeconomic status. For example, prevalence data from 2009 for current U.S. adult cigarette smokers indicate that, among racial/ethnic groups, adults reporting multiple races had the highest smoking prevalence (29.5 percent), followed by American Indians/Alaska Natives (23.2 percent) (Ref. 4). We also agree that economically disadvantaged populations have higher smoking prevalence rates. For example, data from 2009 indicate that the prevalence of current smoking was higher among U.S. adults living below the Federal poverty level (31.1 percent) than among those at or above this level (19.4 percent) (
Id.
). We have selected required warnings that will help effectively convey the negative health consequences of smoking to a wide range of population groups, including different racial and ethnic groups and different socioeconomic groups, and that can help both to discourage nonsmokers from initiating cigarette use and to encourage current smokers to consider quitting. For additional information regarding our selection of required warnings to reach a broad range of population groups, see section III of this document regarding our selection of the final images.

(Comment 3) Multiple comments stated that tobacco use disparities exist among lesbian, gay, bisexual, and transgender individuals. One comment from a community organization stated that lesbian, gay, bisexual, and transgender individuals smoke at rates almost 50 percent to 200 percent higher than the rest of the population and strongly supported the proposed rule.

(Response) We agree that evidence suggests that gay, lesbian, bisexual, and transgender populations have higher smoking rates than their heterosexual counterparts (Ref. 13). The required warnings will help convey information about various health risks of smoking to individuals from a wide range of demographic groups and will help encourage smoking cessation and discourage smoking initiation.

(Comment 4) One comment from a nonprofit research organization indicated that members of the U.S. military have rates of smoking that are unacceptably high, particularly among younger members. The comment detailed the negative outcomes of smoking to military personnel, including lower physical performance, an increased risk of injury during physical tasks, a greater number of days sick and unable to report for duty, poorer job performance, and a higher likelihood of premature discharge from active duty, and stated that smoking and its negative effects among active duty personnel costs the military an estimated $1 billion annually in health care and lost productivity (Ref. 14). The comment also referred to evidence suggesting the tobacco industry has targeted military members and fought efforts to reduce tobacco product consumption by military personnel, and indicated that the proposed rule is an important step in protecting military members from the health harms of cigarette use and will likely decrease cigarette use among military personnel.

(Response) We agree that members of the U.S. military have higher smoking prevalence rates than the general population; approximately 20.6 percent of the U.S. adult population smoke cigarettes, while data from 2008 indicate that 31 percent of active duty military personnel smoke cigarettes (Ref. 15). We agree that the required warnings will help convey information about various health risks of smoking to a wide range of individuals, including members of the U.S. military and veterans who began smoking while in military service, and that the required warnings will encourage smoking cessation and discourage smoking initiation in these individuals.

2. Health Consequences of Smoking

Smoking is responsible for at least 443,000 premature deaths per year in the United States, and each year cigarettes are responsible for approximately 5.1 million years of potential life lost (Ref. 1). Annual direct health care expenses due to smoking total approximately $96 billion, and annual productivity losses due to premature deaths alone from cigarette smoking total approximately $96.8 billion (
Id.
).

The Agency received many comments that were supportive of the proposed rule, some of which reiterated the health risks of smoking described in the NPRM (75 FR 69524 at 69527 through 69529) and stressed the need for measures, such as graphic health warnings, to curb smoking in the United States in order to improve health and to reduce the massive health care costs attributable to tobacco-related illnesses. Some of these comments cited data demonstrating that smoking is the leading cause or most powerful risk factor for particular diseases, such as chronic obstructive pulmonary disease (COPD), bladder cancer, and atherosclerosis.

However, FDA also received multiple comments disputing the health risks of smoking. These comments and FDA's responses are summarized in the following paragraphs.

(Comment 5) One comment from an individual expressed a belief that addiction to nicotine is 99 percent

psychological and only 1 percent pharmacological, and that nicotine is no more addictive than caffeine.

(Response) We disagree with the assertion that nicotine addiction does not have a substantial physiologic component. While we acknowledge that behavioral processes play a role in initiation and maintenance of nicotine addiction, nicotine is a powerful pharmacologic agent that acts in a variety of ways at different sites in the body. As stated in the NPRM, nicotine causes physical dependence characterized by withdrawal symptoms that usually accompany nicotine abstinence (75 FR 69524 at 69528). Regarding the relative addictiveness of nicotine and caffeine, caffeine is distinct from nicotine in its abuse liability, which includes a consideration of multiple factors, including the dependence potential of a substance and the degree to which it produces adverse effects (
see
Ref. 16 at p. 304). Caffeine produces only minimal disruptive physiological effects and, unlike nicotine from tobacco products, caffeine is generally not used in ways that are considered to be of significant adverse health effect (
see Id.
at pp. 285 and 304).

(Comment 6) One comment stated that nicotine withdrawal is the only medical condition that is irrefutably caused by cigarettes.

(Response) We disagree with this comment. While nicotine addiction is one negative health effect of cigarette smoking, it is not the only medical condition irrefutably caused by cigarettes. As detailed in the 2004 report of the Surgeon General, “The Health Consequences of Smoking,” which summarizes thousands of peer-reviewed scientific studies and was itself peer-reviewed, cigarettes have been shown to cause an ever-expanding number of diseases and conditions, including lung cancer, laryngeal cancer, oral cavity and pharyngeal cancers, esophageal cancer, bladder cancer, pancreatic cancer, kidney cancer, stomach cancer, cervical cancer, acute myeloid leukemia, all the major clinical cardiovascular diseases, COPD, and a range of acute respiratory illnesses (Ref. 2).

Maternal smoking during pregnancy causes a reduction in lung function in infants, and women who smoke during pregnancy are more likely to experience premature rupture of the membranes, placenta previa, and placental abruption (
Id.
at pp. 508 and 576). Smoking also increases rates of preterm delivery and shortened gestation, and women who smoke are twice as likely as nonsmokers to have low birth weight infants; smoking also increases the risk of sudden infant death syndrome (SIDS) (
Id.
at pp. 569, 576, 587 and 601).

Children who smoke experience impaired lung growth and an early onset of lung function decline (
Id.
at pp. 508-509, 2004 SG). Smoking during adulthood also leads to a premature onset of accelerated age-related decline in lung function (
Id.
at p. 509). Smoking also results in poor asthma control and causes a range of respiratory symptoms in children, adolescents, and adults, including coughing, phlegm, wheezing, and shortness of breath (
Id.
).

Furthermore, cigarette smokers have poorer overall health status compared to nonsmokers, and an increased risk of adverse surgical outcomes related to wound healing and respiratory complications compared to nonsmokers. Smokers are also at an increased risk for hip fractures, and smoking increases the risk for periodontitis, cataract, and the occurrence of peptic ulcer disease in persons who are
Heliobacter pylori
positive (
Id.
at pp. 717-719, 736, 777, 780, and 813).

In addition, exposure to secondhand smoke has been shown to cause a variety of negative health effects in nonsmokers, including lung cancer, cardiovascular disease, and respiratory symptoms (
see
Ref. 17).

(Comment 7) Some comments were submitted by individuals disputing the negative health consequences of smoking that are described in the graphic warnings. These comments generally indicated that the individuals submitting the comments were smokers, and that they and/or their family members (who were exposed to secondhand smoke) had not experienced negative health effects from smoking.

(Response) We disagree with these comments. Cigarette smoking has been shown to cause a wide range of negative health consequences, as detailed in the previous response. Furthermore, it can be years before some of the negative health consequences of smoking clinically manifest. Thus, the personal health status of the individuals submitting these comments could change in the future. A scientific determination that a product causes a particular negative health consequence is based on data from large groups of individuals, and the fact that an individual product user has not experienced (or has not yet experienced) a particular negative health consequence does not mean the product does not cause that harm.

Moreover, to the extent these comments indicate that many smokers do not fully understand the serious health risks of cigarettes or do not believe that these risks apply to them, they illustrate the need for health warnings that effectively communicate the negative health consequences of smoking to consumers. For additional information regarding consumers' lack of knowledge of smoking risks, see section II.C of this document.

(Comment 8) One comment stated that cigarettes are a minor public health concern compared to obesity and alcohol, and that cigarette use results in less health care costs than medical treatment for the obese.

(Response) As discussed in the NPRM, cigarette smoking is the leading cause of preventable death and disease in the United States (Ref. 4). Furthermore, cigarettes are responsible for health care expenditures and productivity losses resulting in a combined economic burden of approximately $193 billion per year (Ref. 1). The total costs of smoking to society are much higher, as the estimate for productivity losses does not include costs associated with smoking-related disability, employee absenteeism, or costs associated with secondhand-smoke attributable disease morbidity and mortality (
Id.
).

We disagree that cigarettes are a minor public health concern, even as compared to other public health issues, and also disagree with the implication that the public health issue of smoking should not be addressed because other public health issues exist. The required warnings will have a significant, positive impact on public health (75 FR 69524 at 69526), and as a result will help mitigate the single largest cause of preventable death and disease in the United States.

B. Inadequacy of Existing Warnings

In the preamble to the proposed rule, FDA explained how cigarette packages and advertisements can be effective channels for communication of important health information, particularly given that pack-a-day smokers are potentially exposed to warnings more than 7,000 times per year (75 FR 69524 at 69529). However, the existing warnings have suffered from three crucial problems: (1) They have not changed in more than 25 years, (2) they often go unnoticed, and (3) they fail to convey relevant information in an effective manner. FDA also explained that larger, graphic warnings communicate the health risks of smoking more effectively. The preamble to the proposed rule presented extensive evidence from other countries' experiences with graphic warnings as well as information from the 2007 IOM Report (75 FR 69524 at 69531). On the

basis of the available scientific evidence, the IOM concluded that larger, graphic warnings would promote greater public knowledge of the health risks of using tobacco and would help reduce consumption (Ref. 3).

We received numerous comments regarding the adequacy of the existing warnings that appear on cigarette packages and advertisements. The large majority of these comments supported our analysis of the existing warnings, but a few comments disagreed with this analysis. These comments, and our responses, are summarized in the following paragraphs.

(Comment 9) A substantial number of comments, including those from health institutions, nonprofit organizations, academics, and consumers, agreed with FDA's conclusion that the existing warnings that appear on cigarette packages and advertisements are ineffective at conveying the health risks of smoking (75 FR 69524 at 69529 through 69531).

However, one comment stated that the current warnings were “fine.” Two comments expressed the belief that the existing warnings have worked successfully in the current information environment.

(Response) We disagree with the comments stating that the existing warnings that appear on cigarette packages and advertisements are effective. As several other comments noted, the Surgeon General has long recognized that the cigarette warnings are deficient. For example, in its 1994 report the Surgeon General noted that the warnings had become ineffective due to their size, shape, and familiarity (Ref. 18). That same year, the IOM concluded that the warnings were “inadequate * * * and woefully deficient when evaluated in terms of proper public health criteria” (Ref. 19 at p. 237). Yet those same warnings are still in use more than 16 years after the Surgeon General's report and 26 years after their inception. Accordingly, we conclude that the existing warnings for cigarettes do not adequately communicate the health risks of smoking.

C. Consumers' Lack of Knowledge of the Health Risks

In the preamble to the proposed rule, FDA described how the existing warnings that currently appear on cigarette packages and advertisements have largely gone unnoticed by both smokers and nonsmokers (75 FR 69524 at 69530). FDA also provided clear evidence that the warnings have failed to convey appropriately crucial information such as the nature and extent of the health risks associated with smoking cigarettes (75 FR 69524 at 69530 through 69531).

FDA received many comments regarding the level of consumers' knowledge regarding the health risks of smoking. Several comments stated that consumers are adequately informed about the risks of smoking or even overestimate the risks of smoking, while many other comments explained that consumers lack knowledge about a wide variety of smoking risks. A summary of these comments, and our responses, is included in the following paragraphs.

(Comment 10) Several comments, including comments from tobacco product manufacturers and individual consumers, objected to the new required warnings, in part because they claimed that consumers already know the health risks associated with smoking. The submitters expressed the belief that the new warnings are unnecessary, because the new warnings provide information that the public has been aware of for many years.

(Response) We disagree. Many comments provided significant evidence to support the notion that consumers, including those in communities with low literacy rates and military personnel, actually lack knowledge or underestimate the risks associated with smoking. As discussed in this document, this lack of knowledge may involve either an incomplete understanding of the statistical risks or a failure to understand the personal (as opposed to the statistical) risks (
see also
section XI.B.2 of this document). There is also a possibility that the risks are not considered at the time of purchase, even if they are understood—a special problem for those who are deciding whether to start to smoke. The requirements adopted here should help to counteract all of these problems.

While most smokers understand that smoking poses certain statistical risks to their health, many fail to appreciate the severity and magnitude of those risks (Refs. 20 and 21), and there is evidence that even when smokers appreciate the statistical risk, they underestimate the personal risk that they face (Ref. 22). A 2007 survey found that two in three smokers underestimate the chance of a smoker developing lung cancer compared to a nonsmoker (Ref. 23). The survey also found that up to a third of smokers erroneously believe that certain activities, such as exercise and taking vitamins, could “undo” most of the effects of smoking (
Id.
).

Other research also highlights how smokers underestimate the health effects of smoking. For example, in a 2008 survey, more than one-quarter of current smokers did not agree that smoking increases a person's chances of getting cancer “a lot” (Ref. 24). Furthermore, one study, involving smokers' perception of their personal risk, found that only 40 percent of current smokers believed they had a higher-than-average risk of cancer and only 29 percent believed they had a higher-than-average risk of heart disease (Ref. 25). Even among heavy smokers (those who smoke at least 40 cigarettes per day), less than half believed they were at increased risk for these diseases (
Id.
). In another demonstration of underestimation of personal risk, a study found that adolescent smokers underestimated their personal risk, even if they had an accurate sense of the statistical risk (Ref. 22).

A 2005 study of smokers in the United States and three other countries found that there were significant gaps in smokers' knowledge about the risks of smoking and that smokers living in countries where health warnings referred to specific disease consequences of smoking were much more likely to be aware of those consequences (Ref. 26). The study concluded that smokers are not fully informed about the risks of smoking, and that warnings that are graphic, larger, and more comprehensive in content are more effective in communicating the health risks of smoking (
Id.
).

Thus, even if consumers are aware of certain negative health consequences of smoking, such as lung cancer and emphysema, and even if they are aware of certain statistical risks, many smokers underestimate their personal risks, and many Americans are under-informed about other health risks associated with smoking. For example, while nearly all daily smokers in one study correctly identified that smoking caused lung cancer (99 percent) and emphysema (97 percent), a lower percentage of respondents correctly identified smoking as causing low birth weight babies (88 percent), worsened asthma (85 percent), miscarriages (76 percent), other cancers (69 percent), head and neck cancers (68 percent), cervical cancer (48 percent), stomach ulcers (46 percent), reproductive difficulties (44 percent), osteoporosis (41 percent), and SIDS (40 percent) (Ref. 27). In fact, research indicates that most people know only one or two of the many diseases causes by smoking. One survey found that while a majority of people knew that smoking caused life-threatening illnesses, more than half of the respondents were unable to name a smoking-related illness other than lung cancer (Ref. 28). Similarly, researchers

found that when asked about health risks of smoking, 39 percent of respondents either answered incorrectly or said they did not know (Ref. 29).

Americans also lack adequate understanding of the addictive nature of cigarettes. Although one comment provided local surveys showing that adults already know that cigarettes are addictive, there is also evidence that many adolescents do not appreciate the addictive nature of cigarettes. The 2007 IOM Report explained that “adolescents misperceive the magnitude of smoking harms and the addictive properties of tobacco and fail to appreciate the long-term dangers of smoking, especially when they apply the dangers to their own behavior” (Ref. 3 at p. 93). In addition, one survey found that fewer than 5 percent of daily smokers in high school think that they still will be smoking at all in 5 years, yet more than 60 percent of high school smokers are regular daily smokers 7 to 9 years later (Ref. 30). Another survey found that only 7.4 percent of adult smokers and 4.8 percent of young smokers expected to smoke longer than 5 years when they started, but 87 percent of these adults and 76 percent of these youth reported that they had been smoking for more than 5 years (Ref. 31).

There is also evidence that certain demographic groups are even less aware of the negative health consequences of smoking, which is particularly concerning in light of the evidence that these groups also have some of the highest smoking prevalence rates (
see
section II.A.1 of this document). For example, research shows that knowledge of smoking risks is lower among people with lower incomes and fewer years of education (Refs. 32 33 and 24). Smokers in the military also underestimate the actual risk of serious disease and substantially underestimate their own risks (a point that fits well with the evidence of underestimation of personal risks) (Refs. 34 35 and 36).

In addition to underestimating the risks smoking poses to their own health, Americans underestimate the health effects of secondhand smoke on others. In the 2010 Report, “How Tobacco Smoke Causes Disease: The Biology and Behavioral Basis for Smoking-Attributable Disease,” the Surgeon General concluded that “many of the effects from active smoking can be observed in persons involuntarily exposed to cigarette smoke” (Ref. 37). In addition, individual studies have shown that secondhand smoke triggers childhood asthma and is associated with both heart disease and cancer (Ref. 17). Yet, most parents believe that smoke exposure has little or no negative impact on children's asthma (Ref. 38), and a 2009 study found that nearly one-fifth of Americans do not believe that secondhand smoke is dangerous to nonsmokers (Ref. 39).

There is a final point. Even if many people do have an accurate understanding of the statistical risk, and even if, in the abstract, many smokers also have an accurate understanding of their personal risk, that understanding may be too abstract to be thought of at the time of purchase, especially (but not only) for those who are starting to smoke. Efforts to make the relevant risks salient are justified and indeed required under the Tobacco Control Act.

(Comment 11) A few comments claimed that adults actually overestimate the risks of smoking-related disease, and stated that this further underscores the lack of a need for graphic health warnings. In particular, one comment referred to a Montana survey in which adults believed that smoking caused colon cancer.

(Response) We disagree with these comments. While the Montana survey referred to in one of the comments indicates that some consumers are not aware of the precise relationship between smoking and certain diseases (for example, the 2004 Surgeon General's report notes that the evidence is suggestive but not sufficient to infer a causal relationship between smoking and colorectal cancer (Ref. 2 at p. 26)), we are aware of significant research indicating that many consumers are not sufficiently aware of the risks associated with smoking, as discussed in the previous response. We find that the weight of evidence clearly demonstrates that many consumers lack adequate knowledge about the health risks of smoking—especially the personal risks. In addition, the comments claiming that adults overestimate smoking's risks fail to take into account consumers' lack of knowledge of other health risks due to smoking, like the dangers of secondhand smoke, reproductive difficulties, and miscarriages, as described in the previous response.

D. Larger, Graphic Warnings Communicate More Effectively

Since Canada first introduced graphic health warnings for cigarettes in 2001, an extensive evidence base has been developed to examine the effects of graphic health warnings in Canada and in the more than 30 other countries that have adopted similar requirements for graphic health warnings on cigarettes. As FDA extensively discussed in the NPRM (75 FR 69524 at 69531 through 69533), the research literature indicates that large graphic health warnings, such as those being required in this rule, are more likely than text-only warnings to (1) get consumers' attention, (2) influence consumers' awareness of cigarette-related health risks, and (3) affect smoking intentions and behaviors. FDA received many comments on the efficacy of large, graphic warnings, as well as comments regarding the potential for any rebound effect from the use of graphic warnings. Those comments, and FDA's responses, are summarized in the following paragraphs.

(Comment 12) A wide variety of comments, including those from health institutions, nonprofit organizations, and academics, agreed with FDA's findings in the NPRM that larger, graphic warnings are effective.

However, several comments stated that the changes in the format and placement of the warnings being proposed, including the use of graphic images, will not result in reductions in cigarette use given the experiences in other countries. For example, one comment noted that Health Canada's own data found, among other things, that there was no statistically significant decline in smoking incidence consumption for adolescents or adults after the introduction of graphic warnings. This comment expressed the belief that Canada's warnings have been ineffective and that FDA's graphic health warnings will be similarly ineffective.

(Response) For the reasons stated in the NPRM, we conclude that larger, graphic warnings are effective in conveying the health risks of smoking, influencing consumer awareness of these risks, and affecting smoking intentions. We disagree with comments stating that the change in format and placement of the warnings will not be effective. The set of required warnings we have selected will satisfy our primary goal, which is to effectively convey the negative health consequences of smoking on cigarette packages and in advertisements, and this effective communication can help both to discourage nonsmokers, including minor children, from initiating cigarette use and to encourage current smokers to consider cessation to greatly reduce the serious risks that smoking poses to their health.

The research literature clearly indicates that larger, graphic warnings are effective at communicating the health risks associated with smoking, encouraging users to quit smoking, and discouraging nonsmokers from beginning to smoke. We already included significant research to

substantiate this conclusion in the preamble to the proposed rule, and the comments did not specifically dispute this analysis (
see
75 FR 69524 at 69531 through 69532). In addition, as we noted in the NPRM, the available evidence demonstrates that graphic health warnings are (1) more likely to be noticed than text-only warnings, (2) more effective for educating smokers about the health risks of smoking and for increasing the time smokers spend thinking about the health risks, and (3) associated with increased motivation to quit smoking (
Id.
). As several comments noted, evidence from countries with graphic health warnings also indicates that such warnings are an important information source for younger smokers, and that pictures are effective in conveying messages to children (Ref. 40 at pp. 3, 20, and 24-26). These important effects of graphic warnings are sustained longer than any impact from text-only warnings (Ref. 41).

Further, the data from Health Canada does not indicate that the warnings have been ineffective at conveying the health risks of smoking and impacting smoking intentions. We cited several studies in the preamble (including data from Health Canada) that illustrated the effectiveness of the Canadian graphic health warnings, which have been found effective at providing youth and adult smokers with health information, making consumers think about the health effects of smoking, and increasing smokers' motivations to quit smoking, among other things (
see
75 FR 69524 at 69532). For example, national surveys conducted on behalf of Health Canada indicate that approximately 95 percent of youth smokers and 75 percent of adult smokers report that the Canadian pictorial warnings have been effective in providing them with important health information (Ref. 3 at p. 294).

(Comment 13) One comment suggested that the new required warnings will have a greater impact on nonsmokers who inadvertently view cigarette packages than on smokers and, therefore, will not be effective in achieving FDA's goals.

(Response) We are not aware of any evidence to substantiate this comment. Further, our required warnings are intended to have an impact on
both
smokers and nonsmokers. As stated in the preamble to the proposed rule, “the new required warnings are designed to clearly and effectively convey the negative health consequences of smoking on cigarette packages and in cigarette advertisements, which would help both to discourage nonsmokers, including minor children, from initiating cigarette use and to encourage current smokers to consider cessation to greatly reduce the serious risks that smoking poses to their health” (75 FR 69524 at 69526). Therefore, the warnings are intended to have an impact on nonsmokers as well as smokers, and the required warnings will effectively communicate the negative health consequences of smoking to both of these important audiences.

(Comment 14) Several comments, including comments from cigarette manufacturers and individual consumers, expressed concerns that the new required warnings on cigarette packages and advertisements would cause people not to look at packages or cause them to hold their cigarettes in decorative cases. The comments also indicated that some of the proposed images would induce youth to purchase cigarettes rather than deter them from smoking, because the new images would be striking to youth. These comments stated that this “rebound effect” would undermine the intent of the warnings and decrease their effectiveness.

(Response) We disagree. Comments expressing concerns about a potential rebound effect did not provide persuasive scientific evidence to demonstrate such an effect is likely to occur (or that it would have sufficient magnitude to be a significant concern). The comments referenced older studies that did not specifically address graphic warnings on cigarette packages and advertisements, and also referred to a qualitative study conducted on the European Union's graphic warnings, in which some focus group participants commented that some warnings were humorous or that they were not persuasive in educating consumers about dental diseases associated with smoking (Ref. 42). When weighing this qualitative information against the quantitative research available, including evidence from countries with graphic health warning requirements, as well as the findings of the expert panel of the IOM in its 2007 report (
see
Ref. 3), the information referenced in the comments is not persuasive. (While focus groups can provide useful information, it is well known that they are not as reliable as real-world evidence for drawing conclusions about causal relationships and generalizing results to the population as a whole (Ref. 43).)

Furthermore, we note that in the European Union qualitative study referenced in the comments, the researchers concluded that pictures have the potential to add a powerful element to health warning messages and that the old text-only messages were not working (Ref. 42 at p. 43). They also noted that some of the warning messages the comments referred to, including the referenced dental disease image, provoked a highly emotional response in all the countries surveyed despite the comments from certain focus group participants (
Id.
at p. 35). The research literature suggests that images that evoke emotional responses can increase the likelihood smokers will reduce their smoking, make an attempt to quit, or quit altogether (Ref. 44).

While one comment said that the failure of fear-inducing messages based on health effects is “well-known in areas outside of smoking prevention,” the comment did not provide sufficient evidence of such failure in the area of smoking prevention. In fact, as some comments discussed, there is scientific evidence relating to cigarette graphic health warnings illustrating the success of fear-inducing messages (
see, e.g.,
Ref. 44). For example, one comment referred to research that found that smokers exposed to Canada's graphic health warnings generally did not try to avoid the fear-inducing messages, and that any avoidance engaged in by smokers does not appear to undermine quitting intentions or attempts (
citing
Ref. 45). Similarly, researchers analyzing data related to graphic warnings found that:

[T]here is no evidence that pictorial warnings lead to boomerang effects. An analysis of data from the ITC Four Country Survey found that the Australian pictorial warnings, introduced in 2005, led to greater avoidant behaviours (
e.g.
covering up the pack, keeping it out of sight, or avoiding particular labels), compared to Canada, the United Kingdom, and the USA. Importantly, those smokers who engaged in avoidant behaviours were no less likely to intend to quit or to attempt to quit replicating the findings of a study of the Canadian warnings. Thus, although pictorial warnings can lead to avoidance and defensive reactions, such reactions are actually indicators of positive impact.

(Ref. 46,
citing
Refs. 20 and 44). To the extent that smokers engage in any defensive avoidance with respect to the new required warnings, we are adding a reference to a cessation resource to give smokers an immediate way to act upon this impulse and access cessation assistance. The research literature suggests that such a reference is effective in diminishing potential avoidance effects in response to messages that arouse fear (
see
Ref. 40 at pp. 39-41). See section V.B.6 of this document for additional information regarding our rationale and authority for including a reference to a cessation resource in the required warnings.

(Comment 15) Several comments expressed concern about the potential

effectiveness of the new required warnings, particularly those that are fear-based, with certain portions of the population. These comments expressed the following concerns: (1) Many youths and young adults are rebellious and will be attracted to what they perceive as the “forbidden fruit;” (2) fear-based warnings fail with groups that have low self-esteem; (3) fear-based warnings fail with adolescents, because they tend not to be influenced by health-based deterrents; and (4) the new required warnings are “high fear messages” that may actually inhibit reductions in smoking, because they decrease a person's perceived ability to quit smoking. These comments expressed the belief that the new required warnings would be ineffective.

(Response) While acknowledging the concerns, we disagree. It is true that messages that induce fear, pointing to a risk, may not be effective when people are unaware of how to reduce the risk, but in this case, the best way to reduce the risk is clear. We have chosen a balanced set of images, including those that may arouse fear and those that may generate other emotional responses in certain individuals in order to reach a diverse population of smokers and nonsmokers, as well as youth, young adults, and adults. Furthermore, as is explained in more detail in section III.B of this document, we conducted a research study to quantitatively evaluate the relative efficacy of the proposed required warnings in communicating the health harms of smoking to adults (aged 25 or older), young adults (aged 18 to 24), and youth (aged 13 to 17). The nine selected required warnings showed positive effects on important study measures in all study populations, including youth, relative to the text-only control. In particular, as is discussed in more detail in section III of this document, the selected required warnings showed strong impacts on the salience measures in our research study, including emotional and cognitive measures.

The research literature suggests that these measures are likely to be related to behavior change. For example, the literature suggests that risk information is most readily communicated by messages that arouse emotional reactions (
see
Ref. 45), and that smokers who report greater negative emotional reactions in response to cigarette warnings are significantly more likely to have read and thought about the warnings and more likely to reduce the amount they smoke and to quit or make an attempt to quit (Ref. 44). The research literature also suggests that warnings that generate an immediate emotional response from viewers can confer negative feelings about smoking and undermine the appeal and attractiveness of smoking (Ref. 45 and Ref. 40 at pp. 37-38). In addition, research has shown that younger adolescents are more likely to notice and think about health warnings that include graphic images (Ref. 47).

The required warnings will effectively communicate the negative health consequences of smoking, and we do not agree that they will have unintended negative effects among younger population groups.

(Comment 16) One comment expressed concern that the new graphic images on cigarette packages and advertisements would actually make cigarette smokers sicker, as the images would increase smokers' anxiety and damage their self-esteem.

(Response) We disagree. We are not aware of any scientific evidence to support this claim. In fact, as discussed in the preamble to the proposed rule, the available evidence suggests that graphic health warnings can benefit the public health by increasing smokers' intentions to quit and reducing the likelihood of initiation by nonsmokers (75 FR 69524 at 69532).

(Comment 17) A few comments stated that fear-based warnings fail to work when the message being conveyed is already clearly understood and does not provide new information. These comments expressed the view that, because consumers already understand the risks associated with smoking, the new required warnings would not be effective in achieving FDA's goals.

(Response) We disagree. As explained in section II.C of this document, there is substantial evidence demonstrating that the premise of these comments is not correct and that many consumers do not adequately understand the personal risks associated with smoking.

E. Need To Refresh Required Warnings

As amended by the Tobacco Control Act, FCLAA includes provisions that can help prevent or delay the wear out of the new required warnings. For example, section 4(c)(1) of FCLAA (15 U.S.C. 1333(c)(1)) indicates that the required warnings on cigarette packages must be randomly displayed in each 12-month period, in as equal a number of times as is possible on each brand of the product, and be randomly distributed throughout the United States, in accordance with a warning plan approved by FDA. Section 4(c)(2) of FCLAA requires the warnings to be rotated quarterly in cigarette advertisements, also in accordance with a warning plan approved by FDA.

Nevertheless, as stated in the NPRM, we intend to monitor the effects of the new required warnings once they are put into use. We will conduct research and keep abreast of scientific developments regarding the efficacy of various required warnings and the types and elements of various warnings that improve efficacy. As stated in the NPRM, we will use the results of our monitoring and such research to help determine whether any of the textual warning statements or accompanying graphic images should be revised in a future rulemaking (75 FR 69524 at 69534). This commitment to continued empirical testing is consistent with Executive Order 13563, section 1, which states that our regulatory system “must measure, and seek to improve, the actual results of regulatory requirements.”

FDA received numerous comments regarding the need periodically to refresh the warnings to minimize wear out, which we have summarized and responded to in the following paragraphs.

(Comment 18) Many comments, including comments from health institutions, nonprofit organizations, and academics, suggested that FDA should refresh the graphic warnings on a regular basis because consumers can become habituated to and ignore warnings. The comments referred to scientific research on the effectiveness of graphic warnings for cigarette packages and advertisements, which strongly recommends that warnings be periodically refreshed to maintain their effectiveness and impact on consumers (Refs. 18, 42, 44, and 26). The comments suggested a wide range of timeframes as to when FDA should refresh the graphic warnings. One comment suggested that FDA track the effectiveness of the required warnings on a quarterly basis and that the results of any testing be made publicly available. One comment suggested that FDA establish a conclusion that new graphic warnings for cigarette packages and advertisements will be required at no more than a 2-year interval. A few comments also suggested that FDA establish a target schedule for reconsideration and revision of the warnings, which would include ongoing consumer research and re-examination of the effectiveness of the required warnings.

(Response) We agree that refreshing the required warnings on a periodic basis can help maintain their effectiveness. Researchers have found that graphic images and text messages are likely to have greater impact at the time they are introduced and that

meaningful impact of the warnings may decline with repeated exposure (Ref. 41). Rotating a variety of cigarette warnings and updating the warnings periodically is likely to minimize the negative effects of overexposure (Ref. 3).

However, we are not aware of any research that warrants the selection of a particular timeframe for future iterations of required warnings. As stated by several comments, there is no definitive rate at which the warnings will wear out, as it depends on many factors including the variety of message executions, exposure level, and the appeal of the message.

We recognize the value of conducting ongoing evaluation of the effects of the required warnings after they enter the marketplace. We also intend to monitor and evaluate the effects of the required warnings, and to monitor the warnings for potential wear out. In addition, we will keep abreast of scientific developments regarding the efficacy of various required warnings and the types and elements of various warnings that improve efficacy. As noted, this monitoring is consistent with Executive Order 13563, which recognizes the importance of measuring “actual results” and of analyzing significant rules after they are in effect to determine whether they should be “modified, streamlined, expanded, or repealed so as to make the agency's regulatory program more effective or less burdensome in achieving the regulatory objectives.”

When we determine that changes to the required warnings are appropriate (including changes to the textual warning statements and/or the color graphic images) because they would promote greater public understanding of the risks associated with smoking, we can exercise our authority to initiate a new rulemaking to modify the required warnings under section 202(b) of the Tobacco Control Act (adding subsection (d) to section 4 of FCLAA).
1

1
Section 202(b) of the Tobacco Control Act amends section 4 of FCLAA (15 U.S.C. 1333) to add a new subsection (d), “Change in Required Statements.” However, section 201 of the Tobacco Control Act also amends section 4 of FCLAA to add a new subsection (d), “Graphic Label Statements.”

III. FDA's Selection of Color Graphic Images

A. Methodology for Selecting Images

When we issued the NPRM, we proposed color graphic images to accompany the nine textual warning statements required by Congress in section 201 of the Tobacco Control Act. In all, we proposed 36 potential required warnings, consisting of the color graphic images FDA developed and the nine textual warning statements from the Tobacco Control Act. These 36 proposed required warnings were made available as electronic files in portable document format (.pdf) and displayed in the document entitled “Proposed Required Warning Images,” which was included in the docket for the proposed rule. The proposed required warnings were also made available on FDA's Web site. Consistent with section 4 of FCLAA, 2 versions of each of the 36 proposed required warnings were developed; one with the textual warning statement in black font on a white background, and one with the textual warning statement in white font on a black background.

As explained in the preamble to the proposed rule (75 FR 69524 at 69534 through 69535), in considering and developing appropriate color graphic images to accompany the nine textual warning statements set forth in section 201 of the Tobacco Control Act, FDA assessed the graphic warnings that other countries have required, and worked with various experts in the fields of health communications, marketing research, graphic design, and advertising to develop 36 proposed required warnings. Each of the proposed color graphic images depicted the negative health consequences of smoking, and the themes and subjects depicted in each image illustrated the message conveyed by the accompanying textual warning statement.

The NPRM explained that we planned to select 9 final required warnings from among the 36 proposed required warnings. We sought comments on what color graphic images to require in this final rule, including comments on the 36 proposed color graphic images included with the NPRM.

In addition, as is described in more detail in section III.B of this document, we conducted research on the 36 proposed required warnings to evaluate the relative effectiveness of the proposed color graphic images and their accompanying textual warning statements at conveying information about various health risks of smoking, and additionally, at encouraging smoking cessation and discouraging smoking initiation.

In order to determine which color graphic images to require in the final rule, we considered a number of factors. First, we considered the relative effectiveness of the proposed required warnings based on the strength of effect the different color graphic images had on the various endpoints and across the populations included in our study (see section III.B of this document for a more detailed description of the research study).

In addition, we considered the substantive public comments received in the docket related to the 36 proposed required warnings (see section III.C of this document for more information on the comments received; the comments relating to each image are summarized and responded to in sections III.D and III.E of this document). We also considered the comments received in the docket that suggested that we use other images in the required warnings, including images that have been used in other countries' graphic health warnings. However, as discussed in more detail in the following comment summaries and in section III.B of this document, we selected images for the nine required warnings from among the images we developed and proposed. Our research study, among other information, indicated these required warnings will effectively communicate the negative health consequences of smoking to a wide range of population groups. As explained in the comment responses throughout this section III, the comments submitted to the docket did not persuade us that other images, including images used in other countries' graphic health warnings, were more appropriate for use in the required warnings than the images we selected.

Furthermore, we considered the relevant scientific literature in the docket, and in particular the extent to which the literature supported or refuted aspects of the images and the extent to which the literature helped determine the appropriate weight to give to other information (including the appropriate weight to give to the various endpoints considered in our research study).

We also considered the variety and diversity reflected in the images in making selection decisions in order to ensure that the final set of required warnings effectively communicates risk information to a diverse range of audiences, including audiences that have been targeted by tobacco industry marketing efforts. We took into account the importance of selecting a set of required warnings that includes a diversity of styles (
e.g.,
photographic versus illustrative), themes, and human images (
e.g.,
race, gender, age). This is consistent with the evidence base for graphic health warnings from countries that have already implemented such warnings, which indicates that variety is important in enhancing the noticeability and salience of warnings and broadening their relevance for target groups (Ref. 40 at p. 46 and Ref. 48 at

p. 9), and which suggests that warnings that include pictures of people should broadly represent the ethnic/racial profile of the relevant country (Ref. 11).

We also considered whether to have one image accompany each of the textual warning statements set forth in section 201 of the Tobacco Control Act.

We received multiple comments regarding our proposal to select 9 final required warnings and our proposal to select them from among the 36 proposed color graphic images that were made available with the NPRM. We have summarized and responded to these comments in the following paragraphs (we also received a number of comments on the proposed color graphic images themselves; these comments are summarized in sections III.D and III.E of this document. In addition, we received a number of comments regarding our research study, which assessed the relative effectiveness of the 36 proposed color graphic images; these comments are summarized in section III.C of this document).

(Comment 19) Several comments suggested that FDA select more than one graphic image for each new textual warning statement. The comments reasoned that by limiting the warnings to one graphic image per textual statement, the health warnings would effectively communicate to fewer segments of the smoking and nonsmoking populations. Some comments also suggested that selecting more than one image per warning statement would counteract wear out of the required warnings. One comment suggested that FDA develop multiple series of images and require that each series be used one at a time to delay wear out.

(Response) We decline to select more than one image for each warning statement as suggested in these comments. We believe that the set of nine required warnings we selected will be sufficient at this time to achieve our goal of effectively communicating the negative health consequences of smoking and to prevent wear out of the required warnings for several years. Furthermore, the nine selected required warnings will appeal to a diverse range of audiences, and, as discussed in section III.D of this document, the images selected showed significant effects on important measures in our research study across the three study populations (adults, young adults, and youth).

We intend to monitor the effects of these required warnings once they are put into use. We will conduct research and keep abreast of scientific developments regarding the efficacy of various required warnings and the types and elements of various warnings that improve efficacy. Given the significant changes being made to the text, format, and placement of the existing warnings by this rule, it will be valuable to obtain relevant data on the effects of the complete set of required warnings as soon as possible. If we were to expand the number of required warnings, it could delay an assessment of efficacy of the warnings under conditions of real-world use. We intend to use the results of our monitoring and of research conducted on the required warnings once they are in public use to determine whether changes should be made to the required warnings in a future rulemaking, including changes to add new images or to modify the existing required warnings. Accordingly, at this time we decline to select more than nine images.

(Comment 20) Multiple comments suggested that FDA use graphic warning images that have been tested or used in other countries instead of or in addition to one or more of the images that FDA proposed. Some of these comments indicated that images that are in use in other countries would be more effective and educational than some or all of FDA's proposed images.

(Response) We decline to follow this suggestion. FDA's research study evaluated the 36 proposed required warnings. The results from this research study suggest that the nine selected required warnings will effectively communicate negative health consequences of smoking to a diverse range of audiences. Moreover, if we were to select images that were not evaluated in our study, it would be difficult to objectively assess the relative efficacy of such images compared to the 36 proposed images. Compared to the information provided by our research study, the supporting information in the comments did not convince us that the images suggested by those comments would more effectively communicate the negative health consequences of smoking than the images we have selected in this final rule.

(Comment 21) A number of comments suggested that FDA use other images than those published with the proposed rule. For example, some comments recommended that FDA use images that depict real people with real diseases and not models. A few recommended that FDA include images that show negative cosmetic effects of smoking, such as stained fingers and bad breath, in order to impact adolescents concerned about body image. One comment suggested that FDA portray a picture of an obituary, while another recommended the use of an image depicting the amount of money smokers spend to purchase cigarettes every year.

(Response) We decline to select the images suggested in these comments. Each of the required warnings selected by FDA was quantitatively tested to assess its relative effectiveness in communicating the negative health consequences of smoking. In selecting the set of nine required warnings, we considered the results of our research study and a number of other factors and have concluded that the nine selected required warnings effectively communicate the negative health consequences of smoking. In addition, we are adopting the nine textual warning statements mandated by Congress in section 4(a)(1) of FCLAA. The images selected were designed to correlate with those warning statements; the available evidence base highlights the value of the text and images in graphic health warnings relating to one another in a meaningful way (
see
Ref. 40 at p. 41). Including images inconsistent with the textual warning statements could confuse consumers and detract from the effectiveness of the warnings. Furthermore, some of our selected images do show the negative cosmetic effects that can occur as a result of the health consequences of smoking. Moreover, some of the images proposed for use in the comments, such as an image showing the amount of money smokers spend to purchase cigarettes, would not be consistent with the statutory requirement that the required warnings depict the negative health consequences of smoking.

B. FDA's Research Study

As explained in the NPRM (75 FR 69524 at 69535), we conducted research on the 36 proposed required warnings. Specifically, we conducted an Internet-based consumer research study with over 18,000 participants that quantitatively examined the relative efficacy of the 36 proposed color graphic images in communicating the harms of smoking to 3 target groups: Adult smokers (age 25 or older), young adult smokers (aged 18 to 24), and youth (aged 13 to 17) who currently smoke or who are susceptible to smoking.

The purpose of the study was to: (1) Measure consumer attitudes, beliefs, and intended behaviors related to cigarette smoking in response to the proposed color graphic images and their accompanying textual statements; (2) determine whether consumer responses to the proposed color graphic images and their accompanying textual statements differed across various groups based on age, smoking status, or

other demographic variables; and (3) evaluate the relative effectiveness of the proposed color graphic images and their accompanying textual warnings statements at conveying information about various health risks of smoking, and additionally, at encouraging smoking cessation and discouraging smoking initiation.

We placed a report (Ref. 49;
see also
Ref. 50
2

) that described the research study and presented the results of the analyses from the research study in the docket for the proposed rule and announced the report's availability by a notice in the
Federal Register
on December 7, 2010 (
see
75 FR 75936 at 75936 through 75937) so that the public had an opportunity to comment on the results.

2
While the numerical results reported in the study report (Ref. 49) were correct, and while all of the results discussed in this rule are accurately described, some of the descriptors contained in the study report were in error. An errata sheet for the study report (Ref. 50), which lists all the errors and the corrections, has been prepared and is being placed in the docket. These errors did not adversely impact commenters' ability to convey their assessment of the images and the study results in their comments. To the extent some comments included inaccurate statements about the study results in their significant comments as a result of the errors, we recognized the inaccuracy and were able to discern the material points in the comment and evaluate them appropriately, as is reflected in the comment summaries and responses.

This section briefly describes the design of FDA's research study and key endpoints examined in the research study; a full description of the study and the several hundred pages of data and data analyses are contained in the study report and accompanying appendices (Ref. 49) that was placed in the docket for the proposed rule. This section also describes how the results from this research study informed the selection of the final required warnings.

FDA received numerous comments in the docket related to the research study; this section also includes a summary of the substantive comments received about the research study and FDA's responses to these comments.

1. Study Design

FDA's research study evaluated the required warnings proposed for each of the nine warning statements against a text-only control (which contained the warning statement without any accompanying color graphic image). Study participants were randomly assigned to be exposed to either one of the 36 proposed required warnings (treatment groups) or one of the 9 textual warning statements (control groups). Treatment groups for each target population (adults, young adults, and youth) viewed a hypothetical pack of cigarettes that included one of the proposed required warnings, which appeared on the upper 50 percent of the pack, while the control group viewed a hypothetical pack of cigarettes with a warning statement (but no warning image), which appeared on the side of the pack. Furthermore, among adults, an additional treatment group viewed a hypothetical advertisement that included one of the proposed required warnings, which encompassed approximately 20 percent of the upper right area of the advertisement, while a control group viewed a hypothetical advertisement with a warning statement in the same location (but without a warning image) that was presented using the size and format currently required by FCLAA. The study tested the relative efficacy of each proposed required warning relative to the text-only control for that warning statement for the various outcomes measured.

Each respondent viewed either a single cigarette package or advertisement that displayed one of the proposed required warnings or a text-only warning. Respondents answered questions about their immediate reactions to the cigarette package or advertisement, related attitudes and beliefs about smoking, as well as intentions to quit or start smoking. At the end of the survey, subjects were asked to recall which warning statement and image they saw earlier in the survey to assess the accuracy of recall. In addition, 1 week after completing the survey, subjects were re-contacted and asked to recall the warning statement and image to which they were exposed. Overall, the following key outcomes were measured after exposure to one of the required warnings or the text-only control, and/or at 1 week follow-up:

• Salience—The study examined emotional and cognitive responses to the cigarette packages and advertisements that bore health warnings. Participants provided ratings of their responses to the packages and advertisements. The ratings were aggregated to create two scales: (A) An emotional reaction scale, which included ratings on how the warning made the respondent feel, such as “depressed,” “discouraged,” and “afraid”; and (B) a cognitive reaction scale, which included ratings on what the respondent thought about the warning, such as “believable,” “meaningful,” and “convincing”.
3

Regression analyses were used to assess the relative impact of treatment conditions on ratings as compared to the text-only control.

3
Some additional cognitive measures, including the reaction item “the pack was difficult to look at” (or, for the adult sample viewing the print ad, “the ad was difficult to look at”) were also evaluated but were not reported as part of the composite cognitive reaction scale. These items were not sufficiently correlated with the other cognitive measures to include in the composite measure.

• Recall—The study measured participants' recall of the nine warning statements after exposure to either one of the proposed required warnings or the text-only control (baseline). Participants were also re-contacted after 1 week and asked about their recall of the warning statement they had viewed (1 week follow-up). The results were analyzed to determine whether exposure to the proposed required warnings elicited higher recall of the warning statements than exposure to the text-only controls. In addition, in the treatment groups (
i.e.,
participants who viewed one of the proposed required warnings), recall of the image was assessed at baseline and at 1-week follow-up. Because the control group did not view an image, the impact of the proposed required warnings on image recall was measured against one of the proposed required warnings for each warning statement that had been selected to be the referent image and statistically assessing whether recall of the images associated with the other proposed required warnings was higher or lower than recall of the referent image.

• Influence on Beliefs—The study assessed whether the proposed required warnings had a significant impact on beliefs about the health risks of smoking to regular smokers relative to the text-only control, as well as whether they had a significant impact on beliefs about the health risks of secondhand smoke exposure to nonsmokers relative to the text-only control.

• Behavioral Intentions—The study assessed whether the proposed required warnings may have a significant impact on cessation, by assessing smokers' intentions to quit smoking (
i.e.,
asking participants how likely it is that they would try to quit smoking within the next 30 days). In youth, the study assessed whether the proposed required warnings may have a significant impact on potential initiation, using a measure of how likely youth felt they were to be smoking 1 year from now.

As the study report (Ref. 49) explains, the outcomes examined were selected based on established theories of message processing and health-related behavior change, which suggest that immediate emotional and cognitive reactions to messages, and recall of messages, are part of a process that eventually leads to

changes in beliefs and intentions and ultimately to behavior change.

2. Use of FDA's Research Study Results in Selection of Images

As described in section III.A of this document, in order to determine which color graphic images to require in the final rule, we considered a number of factors, including the results from our research study. We carefully examined the research results for the 36 proposed required warnings on all the different outcomes in determining which images to require in this final rule. However, the responses on the salience measures served as a primary basis for distinguishing among the 36 proposed required warnings for a number of reasons.

First, many of the proposed required warnings elicited significant impacts on the salience measures (emotional and cognitive measures), which the research literature suggests are likely to be related to behavior change (Ref. 51). For example, the literature suggests that risk information is most readily communicated by messages that arouse emotional reactions (
see
Ref. 45), and that smokers who report greater negative emotional reactions in response to cigarette warnings are significantly more likely to have read and thought about the warnings and more likely to reduce the amount they smoke and to quit or make an attempt to quit (Ref. 44). The research literature also suggests that warnings that generate an immediate emotional response from viewers can result in viewers attaching a negative affect to smoking (
i.e.,
feel bad about smoking), thus undermining the appeal and attractiveness of smoking (Ref. 45 and Ref. 40 at pp. 37-38).

In comparison to the salience measures, fewer of the proposed required warnings elicited significant impacts on the beliefs measures in our research study, and on the whole the proposed required warnings did not elicit strong responses on the intentions measures. Given the design of our research study, where participants had only a single exposure to one proposed required warning, it is not surprising that the proposed required warnings did not consistently show effects on these beliefs and intentions measures, which are more eventual outcomes in the behavior change process than the salience responses, which occur more immediately. However, this does limit the utility of these longer-term measures in discriminating across the proposed required warnings. Thus, given the design of the study, the results on the salience measures, which the research literature indicates are predictors of more eventual behavioral outcomes, were considered to be more meaningful than the results on the beliefs and intentions measures in discriminating between the images.

In addition, we gave greater weight to outcomes on the salience measures than to outcomes on the statement recall measures for several reasons. First, there is evidence to suggest that, while recall of associated warning message statements may be reduced in the short term by moderately or highly graphic pictorial warnings versus text-only controls or less graphic pictorial warnings, these warnings still increase intentions to quit through evoked emotional responses (Ref. 52). Second, as described previously, participants in the research study were exposed to a single viewing of the proposed required warnings, which does not allow for assessment of the effect that repetitive viewing of the required warnings may have on recall. Recall can be expected to increase in real world settings where consumers will be exposed to the warnings multiple times. Third, recall was generally high for all the proposed required warnings, even where there was not a significant difference compared to the text-only control or where recall was significantly lower for the proposed required warning than for the text-only control. For example, for the nine required warnings that we selected for use in this final rule, the research study shows that recall of both the textual warning statements and the color graphic images was high at both baseline and at 1-week follow-up, exceeding 50 percent on all measures, and, in many cases, exceeding 80 percent.

3. Comments on FDA's Research Study

FDA received a number of comments related to its research study in the docket for the proposed rule, which are summarized and responded to in the following paragraphs.

a.
Study design.
Several comments addressed the cross-sectional design of the study.

(Comment 22) Several comments, including comments from cancer researchers, nonprofit organizations, and academics noted that participants in the study were exposed to a proposed required warning only once in a controlled environment. These comments stated that the single exposure study design makes it impossible to assess long term or actual effects of the proposed required warnings. Two of these comments recommended that FDA conduct longitudinal research or post-market surveillance to assess actual long-term effects.

(Response) We agree that the study design does not permit us to reach firm conclusions about the long-term, real-world effects of the proposed required warnings on the measured outcomes. As noted previously, the purpose of the study was not to assess actual effects but to assess the relative effects of the proposed required warnings on various outcomes. Data on the relative effects of the various proposed required warnings provided a more objective and scientific basis to help select which required warnings should be included in the final regulation. A cross-sectional design with a single exposure under experimental conditions is appropriate for assessing relative effects. For absolute effects, the scientific literature presented in the preamble to the proposed rule provides a substantial basis for our conclusion that the required warnings will effectively communicate the health risks of smoking, thereby encouraging smoking cessation and discouraging smoking initiation.

However, we recognize the value of conducting an ongoing evaluation of the effects of the required warnings after they enter the marketplace, and we intend to monitor and evaluate their ability to effectively communicate the negative health consequences of smoking. This evaluation will provide information regarding whether the required warnings effectively reach the appropriate target audiences, wear out of the required warnings, and whether and what changes to the required warnings may be appropriate in any future rulemaking on this subject.

(Comment 23) A comment from tobacco product manufacturers stated that a longitudinal study demonstrating that the required warnings would have actual effects on smoking prevalence was necessary to support the final regulation.

(Response) We appreciate the value of longitudinal studies but disagree that such a study is necessary to support the final regulation. As discussed previously, our research study assessed the relative efficacy of the 36 proposed required warnings published with the NPRM, and the cross-sectional study design was appropriate for that purpose. The scientific literature presented in the preamble to the proposed rule provides a substantial basis for our conclusion that the required warnings will effectively communicate the health risks of smoking, thereby encouraging smoking cessation and discouraging smoking initiation.

(Comment 24) Several comments discussed behavioral models similar to that described in FDA's research study (
see
Ref. 49) and explained how those models provide a rationale for how health warnings can effectively communicate risk information about the harmful effects of tobacco use. For example, one comment from a researcher working on an international project to evaluate the impact of graphic health warnings for tobacco products stated that the primary objectives of health warnings are to educate and inform smokers and nonsmokers about the many negative health consequences of smoking and to provide information that can enhance their efficacy for quitting. The comment noted that effective health warnings increase knowledge and thoughts about the harms of cigarettes, the extent to which the smoker could personally experience a smoking-related disease, and as a result, increase motivation to quit smoking. Another academic who also is conducting research on graphic health warnings commented that a wide variety of research suggests that health warnings with pictures are significantly more likely to draw attention, result in greater information processing, and improve memory for warnings than text-only warnings. A comment from a researcher with expertise in risk perceptions and decisionmaking stated that changes in smoking behavior based on warning labels appear to require four steps: (1) Immediate, negative affective reactions to the potential consequences of smoking; (2) associations of these emotional reactions to smoking cues; (3) increases in perceptions of the risks of smoking, and finally (4) increases in quit contemplation and reductions in smoking behaviors.

(Response) We agree that the design of our research study is consistent with established social science models (in psychology, economics, and related fields) of risk communication and health behavior change. The purpose of graphic health warnings is to effectively communicate the negative health consequences of cigarette use to smokers and nonsmokers, which is critical given the seriousness of these consequences. Greater understanding of those health effects will motivate some smokers to stop smoking and prevent some nonsmokers from starting to smoke. The preamble to the proposed rule presented a detailed discussion of the scientific literature to substantiate our conclusion that graphic health warnings can be an effective means of communicating important health information about the risks of smoking (
see
75 FR 69524 at 69531 through 69533). These comments provide additional support for that conclusion.

b.
Study results.
Several comments discussed the results from FDA's research study.

(Comment 25) Several comments, including comments from academics, nonprofit organizations, and health professional organizations, stated that FDA's research study provides data consistent with the overall literature demonstrating the effectiveness of graphic health warnings. For example, one comment stated that in general the study results are consistent with prior findings that the addition of graphic images to health warnings is beneficial in comparison to text-only warnings. Another comment stated that, based upon the FDA study and the existing scientific literature, it is possible to conclude that the proposed graphic warnings are likely to be effective.

Other comments, including comments from tobacco product manufacturers, advertising industry associations, and a public policy organization, asserted that FDA's research study fails to provide evidence of efficacy. These comments stated that the study did not show evidence that the proposed required warnings would actually affect prevalence of smoking, and failed to demonstrate sufficient evidence that the proposed required warnings would significantly affect consumer knowledge of the risks of smoking or actual behavior change.

(Response) We agree that the study is generally consistent with the existing scientific evidence demonstrating that graphic health warnings can effectively communicate the negative consequences of cigarette smoking, and by doing so, can encourage smoking cessation and discourage smoking initiation. We disagree that the study results do not support the efficacy of the warnings. We presented substantial research in the preamble to the proposed rule supporting the efficacy of graphic health warnings (75 FR 69524 at 69531 through 69534), and the results of our research study are consistent with that research.

c.
Study outcome measures.
Numerous comments discussed the key outcomes measured in FDA's research study.

(Comment 26) FDA received a wide variety of comments concerning the use of emotional reactions to assess the relative effectiveness of the proposed graphic warnings. A number of comments, including those from academics, medical institutions, and public health groups, supported the inclusion of emotional reaction measures. These comments stated that graphic health warnings that elicit strong emotional reactions, especially negative feelings, are more effective in communicating the negative health consequences of smoking and in motivating healthier behaviors than warnings that do not elicit emotional reactions, and indicate that these effects are well established in the scientific literature.

For example, one comment stated that the scientific literature shows that graphic depictions of the negative health effects of smoking arouse reasonable fears and are associated with greater consideration of health risks, increases in motivations to quit, and ultimately with attempts at cessation. Another comment stated that theoretical models and studies in communications and social psychology suggest that graphic health warnings can be effective because they elicit greater emotional engagement with the information provided and it is that engagement that drives behavior change. Another comment from an academic researcher stated that considerable psychological research suggests that risk is more readily communicated by information that arouses emotional associations with the activity. Emotional reactions can be readily accessed from memory by mere presentation of the stimulus, and appear to be powerful predictors of smoking behavior. Yet another comment stated that growing evidence from controlled experiments and survey research indicates that, compared to text-only warnings, graphic health warnings evoke stronger emotional responses and increase motivations to quit or not start smoking. The comment indicated that these studies are consistent with cognition and neuroscience research demonstrating that relative to linguistic or text information, imagery-based information can be processed more rapidly, evoke stronger emotional responses, induce greater cognitive processing and attitude change and can be recalled more easily.

However, other comments stated that reliance on emotional measures for assessing graphic health warnings is inappropriate. A joint comment from tobacco product manufacturers stated that the study measured only the effect of eliciting strong emotional and cognitive reactions, which confirms that the warnings were intended not to inform consumers with purely factual and uncontroversial information, but rather to shock consumers into adopting the Government's preferred course of conduct. Another tobacco product manufacturer commented that, to the extent FDA selected images based on emotional or cognitive reactions and not

on ability to inform consumers about the health risks of smoking, the regulations would not pass constitutional muster. A comment from a public policy organization commented that emotional and cognitive responses are irrelevant measures of effectiveness if there is no behavior response.

(Response) On the basis of our review of the relevant scientific literature and the feedback received in the docket, we conclude that our inclusion of emotional reaction measures to evaluate the relative effects of the 36 proposed required warnings was appropriate and is consistent with well-established principles in the scientific literature. As discussed in the study report that was placed in the docket (Ref. 49) and in other comments summarized in previously in this document, eliciting strong emotional and cognitive reactions to graphic warnings enhances recall and information processing, which helps to ensure that the warning is better processed, understood, and remembered. Thus, these responses can enhance the effective communication of the health warning message. These responses in turn influence short-term outcomes, such as later recall of the message and changes in knowledge, attitudes, and beliefs related to the dangers of tobacco use and exposure to secondhand smoke. As attitudes and beliefs change, they eventually lead to changes in intentions to quit or to start smoking and then later can lead to lower likelihood of smoking initiation and greater likelihood of successful cessation.

We disagree that use of emotional reaction measurements demonstrates the Agency's intent to advocate a preferred position or course of conduct. Each of the nine graphic warnings required by the final regulations communicates negative health consequences of smoking that are well-established in the scientific literature. Consistent with the Tobacco Control Act, the purpose of these required warnings is to communicate effectively and graphically the very real, scientifically established adverse health consequences of smoking. The overall body of scientific evidence indicates that health warnings that evoke strong emotional responses enhance an individual's ability to process the warning information, leading to increased knowledge and thoughts about the harms of cigarettes and the extent to which the individual could personally experience a smoking-related disease. Increased knowledge and thoughts about the negative consequences of smoking, in turn, are reasonably likely to result in more informed and healthier behaviors, such as trying to quit smoking or deciding not to start.

(Comment 27) We also received two comments concerning the cognitive measure used in the study. A comment filed by tobacco product manufacturers observed that “looks cool” was one of the measured cognitive reactions. The comment stated that the study analysis omits responses on whether the warnings “looked cool,” and contended that if a substantial number of participants viewed a warning as “looks cool,” the warning would be unlikely to have the intended effect. The comment concluded that the ratings for the “looks cool” measure do not appear to have been neutral; the comment stated that regression results for the “looks cool” measure indicates that this measure elicited one of the strongest estimated effects of the study and the results go in the opposite direction of effectively communicating health risk information.

(Response) We disagree that data concerning the “looks cool” outcome was omitted or that the results for this outcome go in the opposite direction of the intended effect of communicating the negative health consequences of smoking. Although the “looks cool” outcome was not included in the reported composite cognitive measure, the study report (Ref. 49) includes the results for this measure in its appendices. The measure was reverse coded, so that a higher value corresponded with the intended directionality for other measures. Thus, a high value for “looks cool” corresponds to a response of “strongly disagree” from the respondent. The data presented in the appendices demonstrate that for each of the nine selected required warnings, significantly more participants disagreed that the warning “looked cool” than participants who viewed the text-only control warning. Eight of the nine required warnings elicited significantly higher ratings than the text-only control warning across all target audiences. Ratings for the ninth required warning, which includes the textual statement “WARNING: Quitting smoking now greatly reduces serious risks to your health,” show that significantly more adults disagreed that the selected required warning “looked cool.” Responses for young adults and youth were in the appropriate direction, but the responses were not significantly different from the text-only control warning.

(Comment 28) We also received a comment concerning the believability measure. This comment raised a concern that some of the 36 proposed required warnings may be perceived as unrealistic because they did not vividly portray immediate health risks, which could lead some smokers to discount the warning. The comment recognized that a believability measure was included in the study as part of the cognitive reaction scale, but stated that specific results for believability were not reported, and recommended that FDA examine the mean scores of the specific believability items in conjunction with other important measures included in the study.

(Response) We agree with the comment that believability is a helpful measure for assessing the relative effectiveness of warning images. All of the selected images scored significantly higher than the controls on the cognition measures, which included ratings on how meaningful the warning was, whether it was informative, and whether it was believable. While the results do not include mean scores for believability and other individual measures, the appendices include the parameter estimates from regression analyses on these individual measures. The results show that, in most cases, the images selected for the nine required warnings scored significantly better than the control with respect to believability.

(Comment 29) One comment stated that the statement recall measure is less important and less relevant to decisions about smoking than negative affective reactions because the warning statements are now believed by smokers and nonsmokers.

(Response) Statement recall was appropriately included as part of the assessment of the relative effectiveness of the 36 proposed required warnings. As discussed in section II.C of this document, while both smokers and nonsmokers have some understanding about some of the risks of smoking, there are significant gaps in their knowledge, including about the magnitude and severity of the risks of smoking. We also note that, as explained in section III.B.2 of this document, although we carefully examined the research results on all the study measures for the 36 proposed required warnings, including recall, the responses on the salience measures served as a more important basis than recall for distinguishing among the 36 proposed required warnings.

(Comment 30) A joint comment submitted by tobacco product manufacturers asserted that the study fails to demonstrate that the published graphic warnings will have any discernible effects on smoking risk beliefs.

(Response) We disagree with this comment. Four of the nine selected required warnings did show a significant impact on beliefs about the health risks of smoking relative to the text-only control among at least one study population. In addition, there is substantial evidence in the scientific literature showing that graphic health warnings effectively increase consumer understanding of the health risks of smoking. In the preamble to the proposed rule (75 FR 69524 at 69531 through 69533), we presented substantial research showing that graphic health warnings significantly increase consumer thoughts about and understanding of the health risks of smoking after they were introduced in other countries. In addition, as discussed previously in this document, considerable scientific evidence shows that health warnings that elicit strong emotional and cognitive reactions are better processed and more effectively communicate information about the negative health consequences of smoking. Each of the nine required warnings elicited strong effects on the emotional and cognitive reaction scales, which indicates that these warning will effectively communication information about the negative health consequences of smoking.

Based on the results of our research study and the existing scientific literature, we conclude that graphic health warnings, including the nine selected required warnings, are likely to increase consumer knowledge and understanding of the health risks of smoking.

(Comment 31) A comment submitted by tobacco product manufacturers criticized the study's use of intentions to measure behavioral change and stated that FDA should have presented data showing actual effects on behavior.

(Response) We disagree that intentions are an inappropriate variable for assessing potential behavioral changes. While measures of intended behavioral outcomes do not perfectly predict a future behavior outcome, it is a necessary precursor. The scientific literature indicates that one's intentions to quit smoking must be increased before one makes the actual quit attempt. Thus, we conclude that it was appropriate in our research study to assess quit intentions as a proxy for behavior change. In accordance with Executive Order 13563, after the rule is in effect we will be undertaking analysis to better understand the behavioral effects of the warnings.

(Comment 32) Several comments raised concerns that the lack of strong statistically significant results concerning intentions in FDA's research study is an indication that the required warnings will not be effective. For example, a comment submitted by tobacco product manufacturers stated that the results of FDA's research study show that graphic health warnings will not result in a statistically significant reduction in youth initiation or overall prevalence of smoking, and thus, confirms that the warnings will not be effective.

(Response) We disagree that our study results indicate that the required warnings will not be effective. It is important to recognize that FDA's research study was not designed or intended to produce evidence demonstrating actual effects on behavior. Rather, the study was designed to provide data concerning the relative effects of the graphic health warnings in order to provide a more objective and scientific basis for our selection of the set of nine required warnings in the final regulation. There is considerable evidence in the scientific literature demonstrating that graphic health warnings effectively increase awareness of the health risks of smoking, which is the principal purpose of the warnings, and that this awareness in turn can influence smoking intentions and behaviors. We included significant research to substantiate this conclusion in the preamble to the proposed rule (
see
75 FR 69524 at 69531 through 69533). For example, as discussed in the proposed rule, a 2007 report from an expert IOM panel that evaluated the existing scientific evidence on health warnings concludes that the available scientific evidence indicates that larger, graphic health warnings would promote greater public understanding of the health risks of using tobacco and would help to reduce consumption (Ref. 3).

FDA's research study cannot be viewed in isolation from the overall body of scientific evidence evaluating the efficacy of graphic health warnings. While the research study itself did not provide evidence of strong effects on intentions (which, as noted in section III.B.2 of this document, is not surprising given the single-exposure design of the study), the overall body of scientific literature does provide sufficient evidence that the required warnings, by increasing public understanding of and thoughts about the health risks of smoking, will be effective in encouraging smoking cessation and discouraging smoking initiation.

A number of comments provide additional support for our conclusion. For example, a comment from a researcher conducting an international longitudinal study on graphic health warnings states that studies show that graphic depictions of smoking's adverse effects on the body are associated with greater consideration of health risks, increases in motivations to quit smoking, and ultimately, attempts at cessation. A comment by a researcher with expertise in risk perceptions and decisionmaking concludes that emotional associations to smoking appear to be powerful predictors of smoking behavior and may well be causally implicated in efforts to either stop or start smoking.

(Comment 33) A comment from tobacco product manufacturers stated that the responses to the “smoking urges” questions included in the study would provide a better measure for assessing whether the proposed required warnings affected smoking behavior and, referring to the responses regarding these questions, the comment asserts that, on balance, seeing the proposed required warnings increased the desire to have a cigarette rather than decreased it.

(Response) We disagree that our research study shows that, on balance, seeing the proposed required warnings increased the desire to have a cigarette. The “smoking urges” measures were reverse coded, so that a higher value corresponded with the intended directionality for other measures in the study. Thus, a high value corresponds to a response of “strongly disagree” from the respondent. The data presented in the study report appendices (Ref. 49, study report) show that, for three of the nine selected required warnings, significantly more participants in at least one target group disagreed with the statement that they wanted a cigarette than participants exposed to the text-only control warning. For one of the selected required warnings, significantly more adult participants who viewed the warning on a cigarette pack disagreed that they wanted a cigarette, but significantly more adults who viewed the warning in a cigarette advertisement agreed. For one of the selected required warnings, significantly more participants in one target audience agreed that they wanted a cigarette than participants exposed to the text-only control warning. Results for the remaining selected required warnings and sample groups were not significantly different from the text-only control warning.

Thus, on balance, the study does not show that exposure to the final set of nine images increased the desire to smoke a cigarette among study participants. As discussed in the previous response, the overall body of

scientific literature provides ample evidence that the required warnings, by increasing public understanding of and thoughts about the health risks of smoking, are likely to encourage smoking cessation and discourage smoking initiation. Data from our research study regarding “smoking urges” provide no basis for calling into question that evidence.

d.
Study limitations and issues regarding methodology.
A number of comments discussed a wide variety of issues concerning limitations of FDA's research study and raised various issues concerning the study methodology.

(Comment 34) Several comments, including comments from health institutions, nonprofit organizations, and academics, raised concerns that the demographics of FDA's research study did not include adequate sample sizes for minority populations and persons of lower income or lower education status. These comments noted that the findings of the study therefore may not be relevant to populations with high smoking prevalence and to those consumers who might be most impacted by graphic health warnings. Some of the comments recommended further testing in these populations.

(Response) We recognize the importance of reaching populations with high smoking prevalence, including various racial/ethnic groups and persons of lower income or lower education status. The study report provides analyses of the relative effects of the images within various sub-groups, separating samples by gender, race, and education. The analyses, for the most part, confirm that the relative effects of the images are consistent across groups. As such, we have determined that the required warnings will help to effectively convey the negative health consequences of smoking to a wide range of audiences, including different racial and ethnic populations and different socioeconomic groups.

(Comment 35) A comment from tobacco product manufacturers criticized the study methodology because it did not include a nationally representative sample of participants and claimed that this failure biased the study results. The comment stated that the study report (Ref. 49, study report) fails to disclose basic sampling information and provides no indication that those conducting the study adjusted for the effect of choosing participants by soliciting volunteers. The comment concluded that this failure was significant because the participants in the study may not reflect the population of interest and may bias the statistical estimates.

(Response) We disagree that the study results are invalid due to the demographic composition of the sample. The research study was not intended to be a survey of the national population, but rather a study using random assignment to study conditions. The study included individuals from certain target groups, particularly current smokers and youth who may be susceptible to initiation of smoking. Statistical methods were used to assess the relative impact of each of the proposed required warnings on various outcomes, rather than to assess the absolute impact one would expect to observe in the U.S. population as a whole.

(Comment 36) One comment raised a concern that lack of adequate pretesting of the proposed required warnings evaluated in FDA's research study could compromise the overall effectiveness of the pool of images tested. The comment stated that it would have been more helpful to conduct pilot testing with a very large group of images (at least 20 per textual warning statement) to ensure testing and selection of the most effective graphic warnings.

(Response) We agree that more extensive pretesting may have been useful. However, we disagree with the suggestion that the overall effectiveness of the required warnings could be compromised by the inability to conduct additional pretesting prior to the research study. The results of the research study as well as research submitted by others during this rulemaking proceeding indicate that the overall efficacy of the pool of proposed required warnings is quite strong. Based on those data, as well as the overall scientific literature, we conclude that the required warnings will effectively communicate the negative health consequences of smoking to smokers and nonsmokers.

(Comment 37) A comment submitted by tobacco product manufacturers asserted that selection bias is a serious methodological flaw of the study. The comment stated that participants were recruited from an Internet panel and offered the opportunity to participate in the research study, creating a selection bias that was compounded by the fact that the invitation to participate stated that the study was funded by FDA. The comment noted that there is no indication that the study corrected for the selection bias and opines that one would not expect the selection bias to be neutral given the identification of FDA as the sponsor of the study.

(Response) We disagree that selection bias is a serious methodological flaw of the study. Although we acknowledge the potential for selection bias, we disagree that this potential bias was likely to significantly affect the results of the study. Even if participants who approve (or disapprove) of FDA were more likely to participate in the study, one would expect that bias would affect all of the experimental conditions, including the text-only control warnings. A bias of this sort would affect the absolute effects of the warnings in general, but not the pattern of relative effectiveness of individual warnings. As a result, selection bias does not invalidate the results of the study, which provides insight on the relative effectiveness of the various warnings under consideration.

(Comment 38) A comment from tobacco product manufacturers stated that FDA's research study is seriously flawed because 32 percent of the participants dropped out of the study before completing the questionnaire. The comment stated that quitting the survey was not likely to be a random event and may have been a result of smokers who are not receptive to graphic health warnings dropping out. If so, the comment suggested that this would have significantly overstated the results of the study.

(Response) We disagree that the drop-out rate observed in the study undermines the validity of the results of the study. Table 3-1 from the methodology report displays the total number of individuals entering the study. However, these values represent the total number of individuals who entered the study's “landing page,” which is the site to which invitees link from the e-mail invitation. The invitation from e-Rewards, as well as the landing page, refers to the study as a “Study about Consumer Products.” There were no references to FDA, smoking, or tobacco in either the invitation or the landing page. Though it is true that a number of invitees chose not to continue after seeing the invitation or the landing page, their decision not to participate cannot be attributed to a bias for or against FDA or the implementation of graphic health warnings on cigarettes.

In addition, the number of individuals identified as “Quits” in table 3-1 of the methodology report includes individuals who quit after viewing the landing page and those who quit after having been informed of FDA's involvement and that the survey concerned smoking or tobacco. Of those individuals identified as “Quits”, only a very small number were in the latter group (
i.e.,
quit after being informed of FDA's involvement and that the survey concerned smoking or tobacco). For

example, of the 13,673 respondents who entered the adult pack survey (the point in time when they viewed the study's landing page), 2,179 chose at some point to discontinue. Of these, only 148 individuals, or about 1.1 percent of those entering the study, chose to discontinue the survey after being informed of FDA's involvement and that the survey concerned smoking or tobacco. A similar pattern exists for all of the study samples: In the adult pack follow-up sample 23 individuals, or 0.6 percent, chose to discontinue after being informed; in the adult ad study sample 193 individuals, or 2.1 percent, chose to discontinue after being informed; in the adult ad follow-up sample 26 individuals, or 0.7 percent, chose to discontinue after being informed; in the young adult study sample 152 individuals, or 1.3 percent, chose to discontinue after being informed; in the young adult follow-up sample 11 individuals, or 0.3 percent, chose to discontinue after being informed; in the youth study sample 104 individuals, or 0.3 percent, chose to discontinue after being informed; and in the youth follow-up sample 13 individuals, or 0.5 percent, chose to discontinue after being informed. The drop-out rate, as calculated here, varies across the study samples but never exceeds 2.1 percent. Therefore, we do not agree that the drop-out rate invalidates the results of the study.

(Comment 39) A comment from tobacco product manufacturers stated that the youth component of FDA's research study is subject to a response bias. The comment stated that the study failed to address the risk that the youth participants might alter their responses due to a conc

[Text truncated at 120,000 characters. The full text is on the page linked above.]

---

Source: Frix Law Library, https://www.frixlaw.com/law-library/documents/fr%3A2011-15337. Public record. Not legal advice.