TABLE 13.-Trends in public knowledge about the health risks of passive smoking Smoking is haLardous lo nonsmokers' health (percentage who agree bv smoking status) Survey YetiI Reference Current smokers Former smokers Never smokers All nonsmokers All adults I. Roper 1974 Roper 1978 30 57 46 2. Roper 1976 Roper 1978 38 61 52 3. Roper 1978 Roper 1978 40 69 58 4. AUTSa 1986 US DHHS, in press 69 82 87 85 81 5. NHISb 1987 68 85 88 81 6. Gallup 1987 ACS 1988b 64 86 89 81 `Percentages presented here are slightly lower than those previously published (CDC 1988) because the lauer did not include "don't know" respomeh in the denommator. %eliminary first-quarter data (unpublished). Year-end percentage for all adults is 81 percent. NOTE: Actual questions: l-3. Is smoking hazardous tononsmokers' health? (probably is hazardous, probably doesn't have any real effect, don't know) 4. Think now for a moment about a nonsmoker who lives or works with smokers Do you think that exposure to tobacco smoke is harmful or not harmful to the nonsmoker's health? 5. The smoke from someone else's cigarette is harmful to you. (strongly agree, agree. disagree, strongly disagree)' 6. If people smoke. do you think that it is harmful or is nor harmful to people who are near them? (yes, harmful; no, not harmful; can't ray/no op~mon) `Percentages include those who"strongly agree" or"agree." like a drug addiction." Of current smokers, 79.6 answered "yes" to the question, "Do you think you are addicted to cigarettes?" (Canadian Gallup 1986) Interaction Between Smoking and Other Exposures The 1985 Surgeons General's Report (US DHHS 1985) reviewed evidence regard- ing the interaction between smoking and a variety of occupational exposures in caus- ing disease. With respect to the interaction between smoking and asbestos, the Report concluded that these two exposures act synergistically to increase the risk of lung can- cer. The risk of lung cancer in cigarette-smoking asbestos workers is more than fif- tyfold the risk in nonsmokers who have not been exposed to asbestos. Few data are available on public knowledge of these interactions. The 1980 Roper survey (unpublished data, FTC) asked respondents about their belief concerning the following statement: "If you smoke and have worked with asbestos you are at least 50 times more likely to get lung cancer than if you have done neither." Seventy-four per- cent of respondents (and 69 percent of smokers) said that they "know it's true" or "think it's true." Smokeless Tobacco Smokeless tobacco (ST) use leads to increased risk of oral cancer and nicotine ad- diction (US DHHS 1986~). No data are available to assess trends in public knowledge of the health risks of ST use. In the 1986 AUTS, 78 percent of adults thought that the use of chewing tobacco is harmful in any way to a person's health. Similarly, 73 percent thought that the use of snuff is harmful to a person's health. Current smokers were less likely to know about the health effects of using chewing tobacco and snuff (7 1 and 66 percent, respective- ly) compared with former smokers (79 and 75 percent, respectively) and never smokers (81 and 76 percent, respectively). According to the 1987 NHIS (preliminary first-quarter estimates), 82 percent of adults thought that a relationship exists between chewing tobacco use and mouth and throat cancers. Seventy-seven percent thought that snuff use is related to these cancers (unpublished data, National Cancer Institute). Personal Health Risks for Smokers There have been few attempts to determine smokers' beliefs regarding their own per- sonal risk. Several Gallup surveys conducted between 1977 and 1987 asked respon- dents, "Do you think cigarette smoking is or is not harmful to your health?" (Table 14). Data are available for current smokers for the years 198 1 and 1985. The proportion of current smokers answering in the affirmative increased from 80 percent in 198 1 to 90 percent in 1985. These data, at first glance, suggest that a high percentage of smokers 202 TABLE 14.Trends in public, beliefs about one's personal risk from smoking Cigarette smoking is harmful to YOUR health (percentage who agree by smoking status) Survey YGU Reference Current smokers Former smokers Never smokers All nonsmokers All adults I. Gallup 1977 Gallup 1985 90 2. Gallup 1978 Gallup 1978 83 95 90 3. Gallup 1981 Gallup 1985 80 96 90 4. Gallup 1983 Gallup 1985 92 5. Gallup 1985 Gallup 1985 90 96 96 94 6. Gallup 1987 ALA 1987 94 7. NHIS' 1987 55 `Preliminary first-quarter data (unpublished). Year-end percentage is 55 percent. NOTE: Actual questions: l-6. Do you think cigarette smoking is or is not harmful to your health? 7. Do you believe your smoking has affected your health in any way? perceive a personalized risk from smoking. However, nonsmokers were asked to respond to the question, implying that the wording may not be understood by some respondents as referring to truly personalized health risks. Wording such as, "Do you think that your cigarette smoking is or is not harmful to your health?' might elicit dif- ferent responses. The 1987 NHIS (unpublished data. National Cancer Institute) showed that 55 per- cent of current smokers answered "yes" to the question, "Do you believe your smok- ing has affected your health in any way ?" The principal reason this percentage is sub- stantially lower than that obtained by the 1985 Gallup survey (90 percent) is probably that the former was likely to be understood as referring to overt symptoms or disease, while the latter was likely to be understood as referring to the risk of harm. Another approach to measure perceptions of personalized risk has been to ask smokers whether they are "concerned" about the effects of smoking on their health. It appears that smokers are more likely today to be concerned that smoking is harmful to their own health. In 1964.50 percent of current smokers were concerned about the pos- sible effects of smoking on their own health (Table 15); this proportion increased to 75 percent by 1986. However, in 1986, only 18 percent of smokers were very concerned about the effects of smoking on their health: 56 percent of smokers were only fairly or slightly concerned; and 24 percent were not at all concerned. From 1970-86, the percentage of smokers who were very concerned about the pos- sible effects of smoking on their health decreased from 29 to 18 percent, while the per- centage who were only slightly concerned increased from 19 to 34 percent. This redistribution within the population of smokers having any concern may have occurred because a much greater proportion of those who were very concerned may have quit smoking during this period; therefore, they would not have been included in subsequent surveys. A third approach to assess personalized risk, or more correctly, the absence of per- sonalized risk, is to ask smokers if they believe themselves to be at lower risk than other smokers. In 1986, 21 percent of adults thought that the cigarettes they smoked were less hazardous than other cigarettes (Table 3). Other data pertaining to perceptions of personalized risk from ETS and from smok- ing among adolescents appear in the sections on Involuntary Smoking (above) and Adolescent Knowledge (below). How Harmful Is Smoking? The data presented above reveal that a vast majority of adults agree that smoking is hazardous to health and correctly recognize the conditions that are associated with smoking. However, these data do not address the depth of the public's understanding regarding the absolute risk of smoking, the relative risks of smoking, the population- attributable risk of smoking, and the risk of smoking in comparison with other risks. A more in-depth understanding of the risks of smoking may be much more important in promoting behavioral change than the more superficial beliefs measured by the data presented above. Unfortunately, only limited data are available to address the public's in-depth understanding of the risks of smoking. 204 TABLE 15.-Trends in smokers' concern about the effects of smoking on their own health Concern about the possible effects of cigarette smoking on your health (percentage who responded by level of concern) Survey Year Very concerned Fairly concerned Only slightly concerned Not Any concerned concerna I. AUTS 1964 13 18 19 50 50 2. AUTS 1966 12 17 18 53 47 3. AUTS 1970 29 22 19 31 69 4. AUTS 1975 25 23 19 32 68 5. AUTS 1986 `Very. fairly, or only slightly concerned. NOTE: Actual questions: 18 22 34 24 75 l-5. Are you in any way concerned about the possible effects of cigarette smoking on your health? SOURCE: US DHEW ( 1969. 1973, 1976a): US DHHS, in press. Absolute Risk Absolute risks can be described by the proportion of those exposed to a given risk factor who will actually die or develop the particular condition, or by the reduction in life expectancy caused by exposure. As many as one-third of heavy smokers aged 35 years will die before age 85 of diseases caused by their smoking (Mattson, Pollack, Cul- len 1987), and 30-year-old smokers will shorten their lives an average of 6 to 8 years if they smoke a pack a day (US DHEW 1979a). From 1970-78, the proportion of adults who believed that smoking a pack of ciga- rettes a day made a great deal of difference in longevity increased slightly from 42 to 50 percent (FTC 1981). However, most adults underestimate the impact of smoking on longevity, according to a 1980 Roper survey. In this survey, 30 percent of the population and 41 percent of smokers did not know that a typical 30-year-old smoker shortened his life expectancy at all by smoking (FTC 198 1). Among those who did know that smoking reduces one's life expectancy, many underestimated the degree to which this is true. On average, nonsmokers underestimated the loss in life expectancy by about 2 years and smokers underestimated it by more than 4 years. Relative Risk Relative risk describes the risk of dying or developing disease for a person exposed to a particular risk factor compared with someone not exposed. For example, male smokers are 22 times more likely and female smokers are 12 times more likely to develop lung cancer compared with nonsmokers of the same sex (Chapter 3). In the 1980 Roper study, respondents were asked if smokers were specifically 10 times more likely to die from lung cancer (the estimated relative risk derived from the data available at that time); 23 percent of the general population and 39 percent of smokers did not believe this statement. Some of this lack of belief may be due to the use of a specific figure. However, using more general terms, 16 percent of adults and 25 percent of smokers did not think that smokers were "many times" more likely than nonsmokers to develop lung cancer (FTC 198 1). Attributable Risk and Smoking-Attributable Mortality Attributable risk refers to that proportion of a disease that can be "attributed" to (or is caused by) a particular risk factor, such as smoking. For example, smoking accounts for about 80 to 90 percent of lung cancer deaths and 80 to 85 percent of deaths from COPD (Chapter 3). Much of the information regarding the public's understanding of the magnitude of the risks of smoking comes from the Roper survey conducted in 1980. In this survey, 43 percent of adults and 49 percent of smokers did not know that smoking causes most of the cases of lung cancer and 22 percent of adults and 27 percent of smokers did not know that smoking even causes many cases of lung cancer (FTC 1981). In the 1987 NHIS (unpublished data, National Cancer Institute), 28 percent (preliminary first- quarter estimate) of smokers and 16 percent (year-end figure) of the general population 206 disagreed with the statement, "Most deaths from lung cancer are caused by cigarette smoking." Attributable risk figures can be used to calculate smoking-attributable mortality. The 1979 Surgeon General's Report (US DHEW 1979a. p. ii) attributed approximately 350,000 deaths each year to cigarette smoking. In 1985, an estimated 390,000 deaths in the United States were attributable to smoking (Chapter 3). In the 1979 Chilton sur- vey, adults aged 29 to 3 1 years were asked: "In the United States, two million people die each year. About how many of these deaths are probably related to cigarette smok- ing?" The responses offered by the interviewer, along with the percentages chosen, were: lO,OOOdeaths, 22 percent; 50.000, 16 percent; 100,000, 16 percent; 300,000, 17 percent; don't know, 31 percent (Chilton 1980). Comparative Risk The risk of dying from smoking can be compared with the risk of dying from other behavioral risk factors, such as living under stress, eating high-cholesterol foods, or drinking heavily. The public's perception of these comparative risks was assessed by Roper surveys from 1970-78 (Table 16). In 1970. living under a lot of tension and stress and not getting regular exercise were considered by more adults to make a great deal of difference in longevity than was smoking a pack of cigarettes daily. In contrast, fewer adults considered regularly eating food high in cholesterol, consuming three or four drinks of liquor a day, or being 20 lb overweight to have an effect on longevity. In 1978, only stress was considered by more adults to make a great deal of difference on longevity. In 1983, Louis Harris and Associates conducted a national telephone survey of 1,254 randomly selected adults for P reCention magazine (Harris 1983). Respondents were asked to rank 24 health and safety factors on a l-to-10 (low-to-high) scale of impor- tance. A sample of 103 health experts (medical school chairmen of preventive medicine, public health school deans, government officials, journal editors, and others) was also interviewed and was asked to make the same rankings. All of the public's mean rankings were in the top half of the scale; thus, none of the factors were seen as trivial in importance. "Not smoking" was ranked near the middle, below "keeping water quality acceptable," "having smoke detectors in the home," "taking steps to con- trol stress," and "getting enough vitamins and minerals" (Figure 1). In contrast, the panel of experts ranked "not smoking" at the top of the list (Figure 2). The 1986 AUTS asked five questions comparing the perceived risk of cigarette smok- ing with the perceived risk of drinking alcoholic beverages, smoking marijuana, being exposed to air pollution, driving without a seat belt, and being 20 lb overweight (Table 17). In each of the comparisons, never smokers were more likely to disagree than to agree that cigarette smoking is less harmful than the other risks. Only in the case of marijuana smoking are the percentages of those agreeing and disagreeing similar. On the other hand, current smokers were more likely to agree than to disagree that cigarette smoking is less dangerous than marijuana smoking and air pollution. Dolecek and coworkers (1986) surveyed 973 adults in Chicago from a sample of family members of students who participated in AHA's Chicago Heart Health Cur- 207 TABLE 16.-Trends in public knowledge about the health risks of smoking compared to other risks, 1970-78 It makes a great deal of difference in longevity if a person (percentage who agree by year) Question 1970 1972 1974 1976 1978 lives under a lot of tension and stress 69 72 74 76 74 doesn't get regular exercise 49 38 38 33 34 smokes a pack of cigarettes a day 42 42 44 45 50 regularly eats a lot of food with high cholesterol 31 34 38 39 43 drinks 3 or 4 highballs a day 29 34 35 37 39 is 20 pounds overweight 23 26 25 24 24 SOURCE: Roper ( 1978). U.: In helping people m general to IIVB a long ano nealrny me, now woulu you rate the importance of . . . M Utmost 01 Low ,. Impodancs 9 8 7 6 5 4 1 I I I I I I w Never dnvmg after drmking 9.25 (.05) HI Keeping air quality acceptable 9 11 (.05) W Keeping water quality acceptable 8.95 (.OS) ct( Havmg smoke detectors in home 8.89 ( 06) i-w Keeping close to recommended weight 8.54 ( 05) l-w Havmg blood pressure readmg annually 8 51 ( 06) ccl Takmg steps to control stress 8.38 (.06) l-o-i Gettmg enough wtamms, minerals 8.37 (.06) Y Exerclsmg regularly 8.32 (.06) I+-f Not smoking 8.25 ( 08) H( Having fnends. relabves. neighbors 8 18 ( 06) +o-f lnherthng genes from parents for long life 8 16 ( 06) m Recetvlng advice from doctor on health habits 8 13 ( 06) l-o-4 Not eatmg too much sodium 8.10 (.06) l-u Gettmg 7-8 hours sleep 8.04 (.06) I+A Eahng enough fiber 7 98 (.06) I-VI Wearmg seatbelts all the bme in front seat 7 89 ( 07) c-c( Not eating too much fat 7.88 (.07) w Gettmg enough calcium 7.84 (.06) &o-f Not eating too much sugar 7 81 ( 07) w Ealmg breakfast daily 7.61 ( 08) M Not getting loo much cholesterol 7 42 ( 07) m Dnnking alcohol moderately 6 53 ( 09) w Dnnkmg no alcohol 6 42 ( 09) FIGURE l.-Adult public's rating of 24 health and safety factors NOTE: Shown &we is the mean importance raring f~,r each factor given by l-2.54 adulta using a I to IO scale. Given III parentheses is the standard emor of the mean. The 95.percent confidence inlerval around each mean is graphically displayed as a band or range consisting of k two standard errc~r values. SOURCE: Harris (1983). 0.: Thinking about the uverall health of the general populatfon, how important is it for adults to . . . M Utmost OfLOW l-v Not smoke 9.78 ( .09) Wear seatbelts all the bme m front seat 9.16 (.12) t - i Never drive after drinking 9.03 (.18) I z I Have smoke detectors in home 8.53 ( .17) 1 = 4 Live where drmking water IS of acceptable quality 8.41 (. 17) t = 1 Have friends, relabves. neIghbars 8.31 (.16) t = I Exercise regularly 8.20 (.16) I = i Drink alcohol moderately 6.15 (.19) t - I Not eat tw much fat 7.82 (.15) .' Keep close to recommended weight 7.71 (.15) Receive advlce from doctor on health habits 1 71.67 (.22) r = I Have blood pressure reading annually 71.62 (.21) I I Inherit genes from parents for long life 7.62 (.28) t c 4 Take steps to control stress 7.58 (. 18) I = I Eat enough fiber 7.41 (.17) t c 4 Get enough Cal&urn (for women) 7.28 (.19) 1 ; 4 Not get too much cholesterol 7.15 (.19) I = I Live where air is acceptable 7.12 (.22) 1 - I Get enough vitamins and minerals 7.12 (.22) I I Not eat too much sodium 7.04 (. 19) i : I Noteat toomuchsugar 6X1(.19) 1-t Get 7-8 hours sleep 6.71 (.20) l-1 Eat breakfast daily 6.16 (.25) Drink no alcohol 3.15 (.23) t-1 FIGURE 2.-Experts' rating of 24 health and safety factors NOTE: Shown above is the mean importance ratmg for each factor given by 103 experts using a I m IO scale. Given in parentheses is the standarderror of the mean. An indicator of the variability of individual ratings around each mean is grapbi- tally displayed as a band or range consisting off two standard error values. SOURCE: Harris (1983). TABLE 17.-Public knowledge about the harmfulness of cigarette smoking compared with other risks, 1986 Percentage who agree Percentage who disagree Current smokers Former smokers Never smokers Current smokers Former smokera Never smokers Moderate use of cigarettes is less harmful to health than 32 21 20 54 63 63 moderate use of alcoholic beverages. Smoking cigarettes is less harmful to health than smoking 48 38 37 33 34 40 marijuana. Air pollution is a greater health risk than cigarettes. 4x 30 28 41 54 57 Smoking cigarettes is less dangerous than driving without a 36 25 26 52 58 68 seat belt. Smoking is less harmful than being 20 pounds overweight. 31 19 IX 59 69 71 NOTE: Percentages of those who agree include those who "strongly agree" or "somewhat agree." Percentages of those who diwgree include those who Wrongly disagree" or "somewhat disagree." SOURCE: AUTS I986 (US DHHS. in press). riculum Program during the 1980-S 1 school year. Respondents were asked to select the three major risk factors for CVD from a list of nine. The percentage responses for these risk factors were: high blood pressure, 25 percent; overweight, 22 percent; stress/tension/worry, 14 percent; cigarettesmoking, 13 percent; heredity/family history, 7 percent; eating too much cholesterol (fat), 7 percent; not enough rest/working too hard, 6 percent; not enough exercise, 4 percent; and diabetes, 2 percent. From 1982-86, Becker and Levine (1987) surveyed 90 adults with no known CHD who were siblings of patients hospitalized for recently documented CHD. Patients and siblings were all less than 60 years old. The siblings were randomized into an assess- ment group (interviewed within 2 weeks of the index patients' discharge and again 4 months later) and a control group (received only one interview at 4-month followup). Participants were asked in an open-ended question to name factors thought to cause or be associated with CHD. Smoking was identified by 81 percent of the control group (after stress, 91 percent) and was the risk factor most often cited by the assessment group (97 percent). Folsom and others (1988) conducted two surveys in the metropolitan Minneapolis/St. Paul area during 1985-86. One survey sampled blacks aged 35 to 74 years, while the other sampled a primarily white population. Subjects were asked the open-ended ques- tion, "What do you think are the most important causes of cardiovascular diseases (heart attack or stroke)?" The percentage of blacks (total sample size= 1,252) who identified smoking as one of the most important causes of CVD was 32 percent; stress/worry (54 percent) and improper diet (45 percent) ranked higher. Among whites (total sample size=1,870), smoking and improper diet were both ranked highest (54 percent). In a survey conducted in 1987 by the Gallup Organization for ACS, 90 percent of adults reported that smoking cigarettes contributes to a higher risk of cancer. Lower percentages reported that a higher cancer risk is associated with suntan and sunburn (73 percent), alcohol (34 percent), high-fat diet (33 percent), and smoked and nitrite-cured meats (31 percent) (ACS 1988b). For the studies reviewed above on comparative risk, data stratified by smoking status were available only from the 1986 AUTS. Knowledge Among Adolescents About the Health Risks of Smoking Because most regular cigarette smokers begin to smoke before age 21 (Chapter 5), it is important to consider teenagers' knowledge about the health effects of smoking. This knowledge can be addressed in the following categories: (I) general health effects of smoking, (2) personalized risk of smoking-related diseases, (3) risks of smoking com- pared with other health risks, (4) beliefs about addiction, and (5) health effects of ST use. General Health Effects Since 1975, beliefs among adolescents that cigarette smoking is harmful have in- creased. National data on knowledge of high school seniors about the health risks of smoking are available from the Monitoring the Future Project (sponsored by the Na- 212 TABLE l&-Knowledge about the health risks of smoking among high school seniors, 1975-86, Monitoring the Future Project, National Institute on Drug Abuse How much do you think people risk harming themselves (physically or in other ways), if they smoke one or more packs of cigarettes per day`? (percentage responding in each category) Survey year Don't know No risk Slight risk Moderate risk Great risk Any risk a 1975 2 3 9 35 51 95 1976 2 2 9 31 56 96 1977 2 2 9 29 58 96 1978 2 2 8 30 59 97 1979 1 2 7 27 63 97 1980 1 I 7 27 64 98 1981 I I 6 28 63 98 1982 2 2 7 30 61 97 1983 1 2 7 29 61 97 1984 I 2 6 27 64 97 1985 2 2 6 24 67 97 1986 I 1 5 26 66 97 "Slight. moderate. or great risk of harm combined. SOURCE: Bachman, Johnston, O'Malley ( I9XOa.b. 1981. 19X4. 198% 1987): Johnston and Bachman (19X0): Johnston. Bachman, O'Malley ( I9XOa.b. 19x2. 19X4. 19%) TABLE 19.-Perceived harmfulness of drugs among high school seniors, 1986; Monitoring the Future Project, National Institute on Drug Abuse How much do you think people risk harming themselves (physically or in other ways), if they... (percentage of people responding) Great risk try one or two drinks of an alcoholic beverage (beer, wine, liquor)? 5 try marijuana (pot, grass) once or twice? 15 take one or two drinks nearly every day? 25 smoke marijuana occasionally? 25 try amphetamines (uppers, pep pills, bennies, speed) once or twice? 25 try barbiturates (downers, goofballs, reds, yellows, etc.) once or twice? 25 use smokeless tobacco regularly (chewing tobacco, plug, dipping tobacco, snuff)? 26 try cocaine once or twice? 34 have five or more drinks once or twice each weekend? 39 try LSD once or twice? 42 try heroin (smack, horse) once or twice? 46 take cocaine occasionally 54 smoke one or more packs of cigarettes per day? 66 take amphetamines regularly? 67 take barbiturates regularly? 67 take four or five drinks nearly every day? 67 take heroin occasionally? 68 smoke marijuana regularly? 71 take cocaine regularly? 82 take LSD regularly? 83 take heroin regularly? 87 NOTE: Possible responses included gnat risk, moderate risk, sltght risk. no risk, don't know. SOURCE: Bachman. Johnston, O'Malley (1987). tional Institute on Drug Abuse) forevery year since 1975. Although nearly all teenagers recognize some risk of harm from smoking, the proportion who think that smoking a pack or more a day causes great risk of harm increased from 51 percent in 1975 to 67 percent by 1985 (Table 18). A 1975 survey (US DHEW 1975a) of teenagers who smoked revealed that many thought that the dangers of smoking were exaggerated for their age group (52 percent of girls; 54 percent of boys); that there was too much talk about things that were bad for them (43 percent of girls; 48 percent of boys); and that air pollution was just as im- portant a cause of lung cancer as cigarettes (67 percent of girls; 5 1 percent of boys). In 1986, only 16 percent of high school seniors agreed with the statement, "The harmful 214 effects of cigarettes have been exaggerated" (see Table 24; Bachman, Johnston, O'Malley 1987) (data stratified by smoking status were not published). Personalized Risk In a survey of 895 students in grades 2 through 12 in 134 public schools in Milwaukee, WI, during the 1979-80 academic year, Leventhal, Glynn, and Fleming (1987) assessed the degree to which the students personalized the health risk from smoking. When asked, "Do you think that smoking can injure or hurt the body?" 98 percent answered affirmatively and were able to accurately name one or more body parts that are adversely affected by smoking. A subsample of 622 subjects (smokers and non- smokers) was asked whether they "would be less likely, about as likely, or more like- ly to get sick from smoking than other people." Those answering "less likely" ac- counted for 47 percent of the smokers but only 36 percent of the nonsmokers, 47 percent of those who intended to become adult smokers versus 36 percent of those who did not intend to become adult smokers, and 41 percent of those from smoking families versus 28 percent of those from nonsmoking families. These findings suggest that although children and adolescents recognize smoking as harmful, they may not personalize the risk. This failure to personalize the perception of risk may play a role in the initiation of smoking. Some teenagers may minimize or deny their personal risk because of a belief that cer- tain smoking patterns are safe. In the 1974 and 1979 Teenage Smoking Surveys con- ducted by the Department of Health, Education, and Welfare (US DHEW 1976b. 1979b), about one-quarter of teenagers agreed with the statement, "There's nothing wrong with smoking cigarettes if you don't smoke too many." About one-third agreed with the statement, "Cigarette smoking is harmful only if a person inhales." Comparative Risk In the 1979 Chilton Survey (Chilton 1980), teenagers were asked which of the fol- lowing caused the most deaths during the past year: traffic accidents, fires, cigarette smoking, or drug overdose. Traffic accidents were cited by 44 percent of teenagers, followed by drug overdose (21 percent), cigarette smoking (19 percent), and fires (6 percent). The High School Seniors Survey includes questions about the risks associated with using a variety of licit and illicit drugs at different levels of intake. In 1986,66 percent of high school seniors thought that smoking one or more packs of cigarettes per day causes great risk of harming oneself. More students saw great risk in the regular use of marijuana, cocaine, LSD, and heroin (Table 19). In contrast, more teenagers saw great risk in regular smoking compared with trying amphetamines, barbiturates, Cocaine, or LSD; in trying or using occasionally marijuana or cocaine; or in trying al- cohol, having one to two drinks per day, or having five or more drinks one or two times per week. The Weekly Reader magazine includes a survey twice a year in the periodical, which is distributed throughout the country to more than 10 million children in grades 2 215 through 9. Surveys are filled out in class by students under a teacher's supervision. The topics addressed are rotated so that the same survey is repeated every 4 years. The Spring 1986 survey covered safety and health (Weekly Reader 1986). Of an estimated 400,000 student responses for grades 2 through 6, 128,000 were randomly chosen for analysis. Although the respondents do not represent a randomly selected sample, results pertaining to tobacco are presented here because of the large sample size and the paucity of data available for young children. The survey included the following question: "Many people say the following things are harmful for kids to do. How harmful do you think each is for kids your age? (very harmful, somewhat harmful, not harmful) , . . overeating, eating junk food, listening to very loud music, smoking, chewing tobacco, staying up late, failing to get enough ex- ercise." Grade-specific results for students in grades 4 through 6 showed that smoking (90 to 95 percent) and chewing tobacco (80 to 90 percent) were much more likely to be perceived as "very harmful" compared with the other choices, all of which were con- sidered to be "very harmful" by less than 40 percent of respondents (except for loud music, among fourth graders-70 percent). However, these results should be inter- preted with caution because of the possibility of sampling bias and the leading nature of the question. Addiction Of particular concern are teenagers who are unaware of the addictive nature of cigarette smoking, and who, therefore, may be tempted to "experiment" with smoking. In the 1974 and 1979 DHEW Teenage Smoking Surveys (US DHEW 1976b, 1979b), about one-quarter of the teenagers agreed with the statement, "Teenagers who smoke regularly can quit for good any time they like." About 60 percent agreed that "It's okay for teenagers to experiment with cigarettes if they quit before it becomes a habit." In the 1979 survey, teenagers were asked, "What would you say is the possibility that 5 years from now you will be a cigarette smoker?" Fifty percent of the current regular smokers (48 percent of boys and 52 percent of girls) answered "definitely not" or "probably not." These findings suggest that a large proportion of new smokers are un- aware of or underestimate the addictive nature of smoking. In 1975,56 percent of girls aged 13 to 17 years and 62 percent of young women aged 18 to 35 years thought that smoking was as addictive as illegal drugs (US DHEW 1975a). In the study by Leventhal, Glynn, and Fleming (1987) of 895 students in grades 2 through 12 in Milwaukee, WI, subjects were asked how hard it is for heavy smokers and for light smokers to quit smoking, and how heavy and light smokers feel when they quit. Answers were used to construct a "knowledge of addiction" scale. The inves- tigators found that young people who smoke or who have smoking family members have lower "knowledge of addiction" scores. The authors speculate that these in- dividuals may be "defending against the thought that either they or a parent has an un- controllable problem." Information on teenage beliefs concerning the addictiveness of ST use is discussed below. 216 Smokeless Tobacco Use In 1985, the Office of the Inspector General, Department of Health and Human Ser- vices, surveyed a nonrandom sample of 399 students in 11 junior high or middle schools and 20 high schools in 16 States regarding ST use (US DHHS 1986d). ST users were oversampled based on identification of users and nonusers by school officials. The sample was composed of 290 current ST users (73 percent) and 109 nonusers (27 per- cent). Eighty percent of junior high school users and 92 percent of high school users acknowledged that dipping snuff and chewing tobacco can be harmful to a person's health (Table 20). When asked about the extent of physical harm that may result from ST use, however, about half of users believed that there is no risk or only slight risk from regular use. One-third of junior high school users and only 5 percent of high school users thought that ST use may lead to mouth cancer. There was poor under- standing of the effects of ST use on gum and dental conditions. One-quarter of junior high school users believed that regular ST use is not addictive, and more than one-third did not know that snuff contains nicotine. In summary, these findings suggest that users are substantially uninformed about the health effects and addictiveness of smokeless tobacco use. However, the degree to which these results can be generalized national- ly is limited by the nonrepresentative nature of the sample. Data from the Monitoring the Future Project showed that in 1986, a total of 59 per- cent of high school seniors believed that regular ST use poses a great (26 percent) or moderate (33 percent) risk of harm, compared with 36 percent who believed that ST use poses slight (28 percent) or no (8 percent) risk (Bachman, Johnston, O'Malley 1987). Constituents of Tobacco Smoke The estimated number of known compounds in tobacco smoke exceeds 4,000, in- cluding some that are pharmacologically active, toxic, mutagenic, carcinogenic, and antigenic (Chapter 2). One of these is carbon monoxide, whose presence in cigarette smoke is cited in one of the four health warnings rotated on cigarette packages and ad- vertisements since 1985 (Chapter 7). In a 1979 survey conducted by Chilton Research Services for the Federal Trade Com- mission (FTC 1981), respondents were asked, "Does cigarette smoke contain carbon monoxide?" Fifty-one percent of teenagers (aged 13-l 8) either did not know (21 per- cent) or said "no" (29 percent); 45 percent of adults (aged 29-3 1) either did not know (26 percent) or said "no" (19 percent). In a 1980 Roper survey (FTC 1981), 53 percent of all respondents and 56 percent of smokers did not know or believe that "Cigarette smoke contains carbon monoxide, which is a dangerous gas." In the 1986 AUTS, 62 percent of current smokers answered "yes" to the question, "As far as you know, does cigarette smoke contain carbon monoxide?" Thirteen per- cent said "no," and 25 percent did not know. Former and never smokers were not asked this question. 217 TABLE 20.-Beliefs about the health effects of smokeless tobacco (ST) use among 399 junior and senior high school students (percentage who agree) in 16 States, 1986 Users Nonusers Junior high school High school (N = 76) (N = 214) TN = 109) ST use con be harmful 80 92 97 Risk from ST use None or slight 57 42 32 Moderate to great 43 58 68 Regular ST use may lead to 33 5 5 mouth cancer Gum and mouth problems among 64 41 33 users are very rare ST use increases risk of tooth 24 II 16 stains, wear, and loss Snuff does not contain nicotine 38 20 32 Regular ST use is not addictive 25 15 10 ST use is much more safe than 81 81 59 cigarettes NOTE: ST user defined as follows: has dipped or chewed more than 100 times, currently uses daily or at least 3 days per week, dipping at least three times on days of use. Nonuser detined as follows: has never dipped or chewed. or has only tried tt a few times or twxe than a few times but fewer than 100 times. SOURCE: US DHHS (1986d). 218 Health Benefits of Smoking Cessation The overall mortality ratio of former smokers (compared with never smokers) declines with increasing years of abstinence. According to data reviewed in the 1979 Surgeon General's Report (US DHEW 1979a) from the U.S. Veterans Study and the British Doctors Study, overall mortality rates of former smokers are similar to those of never smokers 15 years after quitting (US DHEW 1979a). With respect to lung cancer mortality, the increased risk diminishes substantially by 5 to 9 years after quitting, but remains above the risk of never smokers for many more years except for those with fewer than 30 years of cigarette smoking (Chapter 2). A reduction in CHD mortality occurs within the first few years after cessation (US DHHS 1983). The risk of COPD mortality decreases eventually after smoking cessation but does not decline to equal that of never smokers, even after 20 years of cessation (US DHHS 1984). In the 1986 AUTS, respondents were asked how long it takes before former smokers' chances of developing a disease return to normal. Slightly more than half believed that the risks return to normal within 5 years (Table 2 1). Results were similar when stratified by smoking status. The 1987 NHIS included questions regarding the health benefits of quitting in terms of specific disease risks. These data were not available for inclusion in this Report. Discussion It has been 25 years since the release of the first Surgeon General's Report on smok- ing and health. During that time, a major public health effort has been made to educate the public regarding the health consequences of smoking (see Chapters 6-8). Public knowledge of the health risks of smoking has improved as a result of this mas- sive public health education campaign. The belief that smoking is harmful to health has increased since 1964. In 1964, a majority of adults acknowledged the general health risk of smoking and believed that smoking is a major cause of lung cancer, but a minority believed that smoking increases the risk of COPD, heart disease, and premature birth. By the mid-1980s a substantial majority of adults (including nonsmokers and smokers) recognized the general health risks of smoking and believed that smoking increases the risk of lung cancer, COPD, and heart disease, and prematurity, low birthweight, mis- carriage, and stillbirths. Knowledge of the risks of exposure to ETS has also increased markedly since 1974; in fact, this high level of belief preceded the release of the 1986 Surgeon General's Report on the health consequences of involuntary smoking. Current Gaps in Public Beliefs About the Health Effects of Smoking Despite the growing level of public knowledge noted above, a substantial numberof Americans are still uninformed about or do not believe the health risks of smoking. These gaps in knowledge or beliefs are more evident when one considers the propor- tion of adults who do not acknowledge certain health risks rather than the proportion who do. For example, among smokers-for whom this information is particularly 219 h) t-2 0 TABLE tl.-Public knowledge about the health benefits of smoking cessation in relation to years of abstinence, 1986