SHIFFMAN, S.M. The tobacco withdrawal syndrome. In: Krasnegor. N.A. (ed.) Ci\ ......................... Office on Smoking and Health Survelh .............................. Measures of Quitting Behavior ...................................... Percentage of Former Smokers in the Entire Population Percentage of Ever Smoher\ Who Are Former Smokers ("Quit Ratio") The Smoking Continuum ......................................... Other Measures ................................................. Trends in the Proportion of Ever Smoker\ Who .,\rc F'~~wI~ Smoker\ ("Quit Ratio") .................................................. Trends by Gender ............................................... Trends by Race.. ............................................... Trends by Age ................................................. Trends by Level of Educational Attainment .......................... Long-Term Abstinence and Relapse .................................. National Health and Nutrition Examination Sur\,e! Epidemlolo~ic FollowupStudy ............................................... The Smoking Continuum ........................................... Percentage of Ever Smokers Who Have Never Tried to Quit Percentage of Those Smohing at I? Month\ Prior to ;I Sur\,c! Inter\ ieu Who Quit for at Least I Day During Those 12 blonth~ ................ Percentage of Ever Smokers Who Had Been Ah\tincnt for Lc\\ Than I YLYII Percentage of Ever Smokers Who Had Been Ah\tinent for I to 1 Year\ Percentage of Ever Smokers Who Had Been Ah\tincnt for at Least 5 Year\ Interpretation of Continuum Findings ............................... 5x3 5x3 5x3 5x-t 5x-l 5x5 sxs sxs sxx 606 606 606 607 A07 Other Measures Related to Smoking Cessation . 60X Intention to Smoke in 5 Years .-. .................................. Receipt of Advice to Quit From a Doctor ............................ Conclusions ..................................................... References ...................................................... 60X hOC) 610 63 INTRODUCTION This volume appendix discusses national trends in smoking cessation over the last 25 vears. specifically updating and expanding descriptions of the national trends in quilting activity presented in previous Surgeon General's reports (US DHHS 1980. 1983, 1988, 1989a). This Section does not provide a detailed discussion of psychoso- cial, pharmacologic, and behavioral factors known to be related to cessation, because this information is available from other sources (US DHEW 1979; US DHHS 1980, 1988. 1989a). Data are utilized from 5 national cross-sectional surveys on adult tobacco use that were performed by the Office on Smoking and Health (OSH) (formerly the National Clearinghouse for Smoking and Health) and the I2 National Health Interview Survey (NHIS) supplements and the National Health and Nutrition Examination Survey (NHANES) Epidemiologic Followup Study (NHEFS), both performed by the National Center for Health Statistics (NCHS). The surveys were conducted between 1965 and 1987. The national surveys and the measures of quitting activity are described below, followed by a discussion of the data. Information on smoking cessation during pregnancy is also included in Chapter 8. Information on smoking behavior was obtained from these surveys by means of self-report(i.e., without biochemical validation). Asdiscussed in Chapter 2. self-report is considered a valid measure of smoking status in cross-sectional surveys. although some underreporting of daily cigarette consumption likely occurs. SOURCES OF DATA National Center for Health Statistics Surveys Survey data collected by NCHS and available for inclusion in this Report were derived from the 1965, 1966.1970, 1974, 1976, 1977, 1978, 1979. 1980,1983. 1985. and 1987 supplements to NHIS and the 1982 to 1984 NHEFS. Cigarette smoking status (current, former, and never) is assessed in the same manner in all surveys. The constructs assessed on the NHIS supplements vary from survey year to survey year. Variables assessed include attempts to quit smoking among current smokers, duration of abstinence among former smokers, and receipt of advice to quit from a doctor. NHIS, a cross-sectional household interview survey, samples the civilian, noninstitu- tionalized population of the United States (NCHS 1958, 1985. 1989). Weighting procedures are used to provide national estimates. Sample sizes for the smoking supplements (ages 20+) vary from approximately 9,700 in 1980 to over 80,000 in 1966. NHEFS was a followup study of persons enrolled in NHANES-I, which assessed lifetime patterns of cigarette smoking behavior among current and former smokers. Whereas NHANES-I participants were drawn from a national probability sample of the civilian, noninstitutionalized population, NHEFS participants included only those who underwent the medical examination in NHANES-I. Personal interviews with each participant or a proxy (for deceased NHANES-I participants) were completed for 12,200 of the 14,407 original examinees. Proxy interviews were conducted with I.697 583 representatives of deceased NHANES-I examinees. The interval between NHANES-I and NHEFS was about 10 years (Madans et al. 1986: NCHS 1987). Office on Smoking and Health Surveys OSH has commissioned five national surveys of tobacco use among adults in this country, referred to as the Adult Use of Tobacco Surveys (AUTSs). The surveys ask detailed questions designed to assess the knowledge. attitudes, and practices of adults regarding all forms of tobacco use. These cross-sectional surveys were conducted in 1964,1966, 1970. 1975, and 1986 (US DHEW 1969,1973,1976; US DHHS 1989b). The similar or identical wording of several standard questions for all five surveys facilitates comparisons. Constructs assessed included tobacco use behavior, intentions regarding future smoking behavior among ever smokers. and receipt of a doctor's advice to quit smoking. Some differences in the conduct and design of the studies occurred. The mode of interviewing changed with time. The 1964 survey obtained data solely from personal household interviews. Whereas personal household interviews were the major mode of data collection in the 1966 survey. telephone interviews and mailed questionnaires were also used to collect data from eligible household members not available when the interviewer was present in the house. The 1970 and 1975 surveys conducted telephone interviews when possible and personal household interviews in nontelephone households. The 1986 survey was conducted entirely by telephone. The I964 and 1966 surveys drew samples only from the contiguous United States. Other AUTSs collected data from residents of all SO States. The actual number of respondents for each survey was 4,635 in 1964,4,06 I in 1966, 5.191 in 1970, 12,029 in 1975, and 13.031 in 1986(US DHEW 1969. 1973, 1976: US DHHS 1989b). In each survey, weighting procedures were used to adjust for an oversampling of ever smokers in the original study population. Comparisons between the 1986 AUTS and the others will not be exact. because the 1986 AUTS weights to an estimate of the adult U.S. population. whereas the other surveys weight to their respective sample sizes. MEASURES OF QUITTING BEHAVIOR As documented in several previous Surgeon General's reports (US DHEW 1979; US DHHS 1988. 1989a) and discussed in Chapter 2 of this Report. smoking cessation is a multifactorial process for overcoming an addictive behavior. One model characterizes this process as having several stages-precontemplation. contemplation, action, and maintenance (Prochaska and DiClemente 1983: Chapter 2). People frequently cycle and recycle through the various stages (marked by frequent relapse episodes) on their way to becoming long-term ex-smokers (Prochaska and DiClemente 1983; Cohen et al. 1989). This analysis of national trends in smoking cessation will use several measures to describe the quitting process. The 1989 Surgeon General's Report (US DHHS l989a) discusses three measures of quitting behavior. These interrelated parameters are discussed below. 584 Percentage of Former Smokers in the Entire Population This measure of quitting behavior has been used to calculate the number of former smokers in the population. Based on data from the 1987 NHIS. for example, 23.6 percent of the 162.6 million civilian, noninstitutionalized adults 20 years of age and older were former cigarette smokers. There were. therefore, approximately 38.5 million former smokers 20 years old or older in the United States in 1987. The percentage of former smokers in the entire population is limited as a measure of quitting activity primarily because it does not take into account the percentage of the population that has ever smoked (and thus is "at risk" of quitting). It also does not differentiate between people who have been abstinent for a short period and people who have maintained abstinence for several years (US DHHS 1989a). Percentage of Ever Smokers Who Are Former Smokers ("Quit Ratio") By dividing the number of ever smokers into the number of former smokers. perspective is given to the magnitude of quitting in a population. The term "quit ratio" has been used to describe this measure (CDC 1986: Pierce. Aldrich et al. 1987; US DHHS 1988. 1989a: Fiore et al. 1989) and is the term used below; this measure has also been termed the "quit rate" (Kabat and Wynder 1987) or the "cessation rate" (Jarvis 1984). The term "ratio" is mostly used in sciences when the numerator and the denominator are two separate and distinct quantities (Elandt-Johnson 1975). "Quit ratio" is used here, even though the numerator is included in the denominator. because of its repeated use in the literature as well as in previous Surgeon General's reports. The percentage of ever smokers who have discontinued smoking indicates the prevalence of abstinence (Ossip-Klein et al. 1986). In 1987, 23.6 percent of the population were former cigarette smokers and 29.1 percent of the population were current smokers. The quit ratio among ever smokers was 44.8 percent; that is, nearly one-half of all living adults who ever smoked cigarettes had quit. Quit ratios by gender and age were recently published for 36 States and the District of Columbia based on 1988 data from the Behavioral Risk Factor Surveillance System (Anda et al. 1990) (Table I). The measure is limited because it treats all former smokers equally, regardless of duration of abstinence. It also classifies current smokers who had never tried to stop smoking in the same manner as it does current smokers who had been abstinent for a long period of time and relapsed shortly before the time of the survey (US DHHS 1989a). The Smoking Continuum The I989 Surgeon General's Report defined a 1 O-category smoking continuum based on data from the 1986 AUTS. This continuum expanded on the smoking status variable (current, former, and never) to incorporate the timing and duration of quit attempts (US DHHS 1989a). Respondents were asked whether they had ever made a serious attempt to quit, and if the response was affirmative, they were then asked about the timing of TABLE I.-Quit ratio" in selected States, by age group and gender-BRFSS, 1988 I x-34 3544, S(L64 25s MW Wonlrn Total SlLlk (h (fYS% Clhl % l+Ys% (`I) % (+Ys% Cl, % (+Ys% Cl) % (*OS% CI) `k (WS'/i <`I) % (It').S'ir (`I) Ald%l~~~;l Arimm California Connecticut District of Columbia Florida Georgia Hawaii Idaho Illinoi\ Indiana Iowa Kentucky Maine Maryland Massachusetts Michigan Minnesota Misouri Montana 143 4.5,s 40. I 3X.6 37.2 43.7 3Y.S 33.7 43.6 2s. I 23.1 30.0 22.6 3x.3 41.1 35.2 37.4 41.2 3.5.x 45.7 (7.0) (7.Y) (5.7) (7.4) (8.X) (7.4) (S.6) (7.1) (7.1) (6.7) (5.5) (Y.2) (S.b) (7.2) (Y.0) (6.6) (7.0) (4.4) (7.1) (9.6) `II.3 54.2 54.3 40.0 45.4 45.3 40.3 45.0 4Y.2 42.6 41.2 503 33.h so.4 45.6 50.x s0.s ss.2 43.`) 5X.X (7.3) (X.0) (5.h) (7.3) (Y.9) (7.2) (X.7) (7.6) (6.`)) (6.6) (S.X) (10.5) (6.9) (7.4) (X.3) (7.0) (7.3) (4.3) (73) (7.7) 54.3 (X.3) 61.2 (9.`)) 64.3 (6.Y) 60.7 (Y.3) 56.5 ( I I .h) S2..? (7.7) S2.1 (12.7) 61.5 (Y. I ) sx.3 (X.3) S3.h (X.0) 51.`) (6.7) 55.4 (12.3) 4x.2 (7.0) 65.3 (X.4) S7.Y (Y.9) X).X (Y.5) 5.5.0 (10.7) 64.6 (5.4) 57.5 (X.41 5x.x (X.2) 65.X 67.0 65.6 74.`) SY.4 77.6 70.2 7O.Y 77. I 64.1 79. I 71.1 63.6 72.7 70.`) 75.5 b.5 .Y 76.0 79.1 7Y.O ( IO. I ) (X.8) (7.X) (X.0) (12.5) (5.7) (17.2) (9.Y) (6.2) (X.3) (5.`)) ( IO.`)) (7.0) (X.7) (Y.3) (7.1) ( 13.4) (S.0) (7.0) (7.1) 47.4 (5.6) 37.3 60. I (6.2) 40.x 56.4, (4.4) 49.3 ss. I (6.4) so.4 52.0 (X.0) 43.4 5x.5 (4.X) so.4 44.2 (7.2) 4h.6 49.6 (5.X) 44.3 61.3 (S.5) 44.2 4x.9 (5.3) 42.4 51.6 (4.4) 41.2 60.2 (X.1) 3Y.l 42.6 (4.`)) 31.1 h0.S (5.Y) 44.5 s3.0 (7.1) 4x.0 56.2 (6.3) 47.6 52.0 (6.4) 45.5 60.4 (3.4) 49.2 54.7 (6.1) 43,s 62.0 (6.4) s4. I (S.9) (6.0) (45) (.5.X) (7.3) (S.2) (6.X) (6.3) (5.5) (5.2) (4.6) (7.1) (4.5) (5.5) (h.5) (5.2) (5.X) (3.X) (5.4) (6.2) 43.2 5.5.4 53 7 52.x 47.6 s4.x 4S.3 47.3 s4.0 3.5.x 47.0 so.7 27.x 53.2 SO.6 51.`) 4X.Y 5.5.4 4Y.b SX.6 (4.1 ) (4.3) (3.1) (4.3, (5.5) (3.5) (5.2) (3.3) (7.Y) (3.h) (7.2) (S.X) (3.4, (4. I ) IS 0) (4.2) (43) l3.5) (4. I ) (4.S) TABLE l.-Continued state 1 X-34 3549 so-64 265 Men Women Total % (2995% CI) % w50/, CI) % (+9s%cI) % (H58 Cl) % (k9sx CI) % ( f')S%, CI % (FWE Cl) Nebraska New Hampshire New Mexico New York North Carolina North Dakota Ohio Oklahoma Rhode Island South Carolina South Dakota Tennessee Texas Utah Washington West Virginia Wisconsin Median prevalence 39.6 (7.4) 57.6 35.2 (7.0) 53.0 3x.4 (X.7) 47.0 29.8 (7.4) 42.7 37.0 (6.X) 47.5 3x.7 (7.1) so.9 30.6 (6.5) 42.3 37.1 (9.1) 43.6 34.7 (6.4) 44.x 2x.9 (5.X) 41.5 37.4 (9.2) 52.3 29.0 (4.X) 40.9 3X.X (X.6) 45.7 33.5 (7.5) 50.9 37.7 (7.3) 54.X 3X.S (7.1) 43.4 35.0 (6.9) 92.0 37.2 47.0 (8.3) 57.6 (7.4) 60.3 (9.2) 53.4 (8.1) 66.3 (6.7) 47.9 (6.X) 62.2 (7.6) 57.4 (8.6) 55.9 (6.X) 55.6 (6.6) 58.7 (8.6) 60.8 (5.7) 49.5 (8.4) 53.7 (X.1) 6X.3 (7.5) 53.7 (6.5) 49. I (7.3) 62.2 57.5 (9.0) (9.9) (9.X) (X.9) (8.5) (7.9) (9.0) (10.2) (X.5) (7.6) (9.2) (6.X) (10.5) (I 1.3) (X.7) (7.2) (X.5) 74.5 (7.X) 74.2 (9.2) 64.7 (I 1.7) x0.3 (X.2) 72.3 (7.0) 73.3 (7.4) 67.3 (X.4) 62.9 (13.4) 60.X (7.2) 72.2 (X.X) 71.1 (X.7) 67.2 (7.9) 69.3 (12.5) x0. I (Y.2) x2.5 (7.9) 69.4 (7.2) 76.0 (9.1) 71.1 59.0 (9.9) s3.x (6.1) 40.4 (7.3) 54.2 (6.X) so.4 (5.7) 5x.3 (5.1) s-2.0 (h.2) 53.4 (7.0) 51.1 (5.h) 46.0 (5.1) 55.6 (5.X) 43.4 (4.S) 52.5 (6.7) 65.2 (5.X) 5x.2 (5.X) 54.5 (5.2) 63.X (S.6) 94.2 47.2 (6.1) 54.0 4X.6 (6.0) 51.5 46.3 (7.0) 4x.0 46.X (6.5) SO.5 43.4 (5.2) 47.3 45,s (5.6) 53.1 31.4 (5.7) 44.9 3Y.b (6.7) 47.5 44.6 (51) 47.x 42. I (S.4) 44.4 SO.5 (6.7) 53.4 3Y.4 (4.1) 4 I .x 41.4 (6.X) 47.9 40.5 (X.6) 5h.h 46. I (5.X) 53.0 3X.4 (5.0) 47.6 46.3 (6.7) 56.5 44.6 so.5 (4.3) (4.6) (5.2) (4.7) (3.9) (3.X) (4.2) (S.1) (3.X) (3.8) (4.3) (3.2) (4.X) (5.0) (4.2) (3.7) (4.4) NOTE: BRFSS=Behavioral Rl\k Factor Survedlance System. "Ddmed as the percentage ofever vnokcn who were former amoken at the time ot the survey hConfidrncr interval. SOURCE: BRFSS IYXX (Anda et al. IYYO) z -4 their most recent quit attempt. This measure provides information on the recent quitting history of the population (Pierce, Giovino et al. 1989: US DHHS 1989a). The trend analyses presented below will use an eight-category continuum (Table 2) among ever smokers to incorporate data from the 1978, 1979. 1980, and 1987 NHISs. As opposed to the 1986 AUTS, the questions asked in these NHISs do not permit a dichotomous classification of current smokers who had never tried to quit according to interest in quitting. In addition to a description of the overall smoking continuum, several segments of the continuum, or measures derived from the continuum, will be described separately. These measures include the following: o The percentage of ever smokers who have never tried to quit; o The percentage of people smoking at I2 months prior to a survey interview who had been abstinent for at least I day during those 12 months: o The percentage of ever smokers who had stopped smoking for less than 1 year: o The percentage of ever smokers who had stopped smoking for I to 4 years: and o The percentage of ever smokers who had stopped smoking for at least 5 years. Other Measures Respondents to AUTSs were asked to estimate the possibility that they would be smoking 5 years after the survey. This question gives a measure of intention to smoke. Finally, respondents to several NHISs and to all OSH tobacco use surveys were asked if a physician had ever advised them to stop smoking. TRENDS IN THE PROPORTION OF EVER SMOKERS WHO ARE FORMER SMOKERS ("QUIT RATIO") Using data from NHISs for I965 to 1987. trends in the proportion of ever cigarette smokers in the U.S. adult population who have stopped smoking cigarettes (quit ratio) are presented by gender and by race in Figures 1 and 2. respectively. Trends for the total adult population. as well as trends by age and by education, are shown in Table 3. These data, with the exception of the age-specific estimates. are age-adjusted to the 1985 population. In thes,e analyses. the quit ratio was regressed on the calendar year of data collection. The R- statistic. supplied for each trend analysis. is a measure of the strength of the linear relationship. R' values may range from 0 (no linear trend) to 1 .O (a perfect positive or negative linear relationship). Trends by Gender As shown in Figure 1. the quit ratio for both genders has been increasing in an approximately linear fashion (R-=0.94 for males and 0.97 for females) since 1965, and 5X8 TABLE d.-Cigarette smoking continuum by year, percentage of ever cigarette smokers, NHISs, United States, 1978-87, adults aged 20 and older Cigarette smoking continuum I. Current smokers who had never tried to quit 2. Current smokers who had quit previously but not in past year 3. Current smokers who had quit for <7 days in past year 4. Current smoker5 who had quit for >7 days in past year 5. Former smoker\ who had quit within past 3 mo 6. Former smokers who had been abstinent for 3-12 mu 7. Former smokers who had been abstinent for I-S yr X. Former smokers who had quit 25 yr earlier Percentage of those smoking during the year prior to the survey who tried to quit during that year (Categories 3+4+5+6 divided by 1+2+3+4+.5+6) Percentage of those smoking during the year prior to the survey who quit during that year and were still abstinent at the time of the survey (Categories S+6 divided by 1+2+3+4+5+6) 197x I979 IYXO 10x7 25.9 26. I 25.4 1X.Y 22.7 2 I .4 23.1 20.0 6.6 6.0 5.9 7.0 KS X.6 7.x x.4 I.3 1.6 I .4 1.x 2.7 2.6 2.7 2.X Y.0 IO.0 Y.5 IO.4 23.3 23.6 24.1 20.7 -___ .- 2X.2 2x.4 26.X 34.0 6. I 6.3 6.7 7.x NOTE: NHIS=NAonal Health Interview Survey SOURCE: NHlS\ IY7X. 197'). 19X0. 1987. z 0 60 MALES FEMALES Slope of Regression Line 0.70 0.76 -........_._..__________________________--------------.------.-.--....-..-..---- !f -----....____ !E! __________ !!L?! _________ I 1 I I I I I I I I I I I I I I I I I I I I I 1964 1969 1974 1979 1984 YEAR FIGURE I.-Trends in the quit ratio, United States, 1965-87, bv gender NOTE: Quit ratio ib the proportion of ever \mohers who are former \moher\. NH`IS=N;it ionnl Health Intreview Survey; OSH=Office on Smoking ond Health. SOURCE: NHIS!, IOhS. IYhh. lY70. 1974, lY7h. 1077. lY7X. 1979. IYXO. lYX3. IYXS. IYX7: OSH. 1. 1 1. 40 30 20 3 10 s E z E 0 Whites Blacks Slope of Regression Line 0.72 0.45 I32 0.96 0.86 ______________---__-____________________~--~--------------------~----~--------------------~..~~~~-~~~~~-~~-~.~~~~~~-~-~.~~~-~--- I I I I I I I I I I I I I I I I I I I I I I I I 964 1969 1974 1979 1984 YEAR FIGURE 2.Trends in the quit ratio, United States, 1965-87, by race NOTE: NHlS=National Health Interview Survey; OSH=Office on Smoking and Health. SOURCE: NHISs 1965. 1966, 1970, 1974, 1976, 1977, 1978, 1979, 1980. 1983, 1985, IYX7: OSH, 22 unpublished data. TABLE .X-Trends in quit ratio ( %) (percentage of ever cigarette smokers who are former cigarette smokers), by age and by education, NHISs. United States, 1965-87, adults aged 20 and older Year I Y65" I YM 20-2-l 17.x 17.0 Age (yr) Educational level Lr\s than Hngh school Some h!gh school (`allege 25-u 4SA4 xl5 graduate praduatc college graduate 22.6 30.0 1x.7 `3.4 30.0 SO.5 33.3 2x.0 28.7 30.1 1070 3s.1 20.x 3.X 36.1 S6.Y 3x. I 33.6 34.9 3x.2 1974 7h.J 2O.Y 3.3 3Y.7 57.x 3X.0 35.2 36.6 47.') I Y7h 37.1 22.0 X4 JO.4 5Y.h 3Y.S 35.0 37.' 36. I 1077 36.X 22.0 ?Y.fl 3.5 5x.7 3x.3 34.0 36.X 4X.6 197x 3X.5 22.x 7 I .Y 40. I 62.4 3x.7 36.3 4 I .o 4Y.7 I Y7Y 34.0 27.h 3 I .x 32.4 61.7 30.x 36.7 37.5 SO.6 I YXO NO 73 7 --.- 33.0 40.0 hl .o 3Y.3 36.5 40.6 4x.7 3x.7 `II.2 S4.Y 40.5 46.0 61.1 `Il.1 4.5.5 SC). I 0.5s 0.74 0 xx 0.0s 0.0x 0. I3 0.92 O.YO 0 x3 the rates of increase for both are also similar (0.70 percentage points/year for males and 0.76 percentage points/year for females). The quit ratio has been consistently higher for males than for females. Using data from the 1970 and I975 AUTSs. Jarvis (1984) reclassified as current smokers males who gave up smoking cigarettes but who con- tinued to smoke cigars and/or pipes. When the use of other forms of smoking tobacco was considered, the difference between males and females in the quit ratio (termed as the "cessation rate" by Jarvis) was reduced by more than two-thirds. Data from the 1987 NHIS Cancer Epidemiology and Control supplement (Schoen- born and Boyd 1989) were analyzed to update the work of Jarvis (Table 4). The weighted percentage of ever cigarette smokers who were former cigarette smokers among males was 48.7 percent. The corresponding number among females was 40. I percent. When former cigarette smokers who smoked cigars and/or pipes were reclas- sified as current smokers (without changing the denominator). the prevalence of cessation among ever smokers became 45 percent for males and 40 percent for females. Furthermore, when former cigarette smokers who used any other form of tobacco (cigars, pipes, snuff, and/or chewing tobacco) at the time of the survey were classified as current tobacco users, the figures became 42. I percent for males and 39.9 percent for females (OSH, unpublished data). Thus, reclassification of former cigarette smokers who were smoking cigars and/or pipes as current smokers reduced the difference in the quit ratio between males and females from 8.6 to 5.0 percentage points. Former cigarette smokers who were using any other form of tobacco were reclassified as current tobacco users, and this reclassification further reduced the difference to 2.2 percentage points. Trends by Race Trends by race are presented in Figure 2; The quit ratio among both whites and blacks has been increasing steadily since 1965 (R&=0.96 for whites and 0.86 for blachs). While the change per year since 1965 is higher for whites (0.72 percentage points/year) than it is for blacks (0.45 percentage points/year), the lines have been essentially parallel since 1974 (Fiore et al. 1989). Use of the 1987 NHIS data to reclassify as current smokers all former cigarette smokers who were smoking cigars or pipes reduced the quit ratio from 46.4 to 44.2 percent among whites and from 3 I .S to 30.2 percent among blacks. Further reclassification, as current tobacco users, of former cigarette smokers who were using any other form of tobacco reduced the numbers to 42.5 percent for whites and 29.1 percent for blacks (OSH, unpublished data). Trends by Age Table 3 provides information on the quit ratio stratified by age. For all age categories. the quit ratio increased from 1965 to 1987. The rate of change was highest in the age categories of 45-64 years and 65 years and older. Reclassification of the I987 data to account for cigar and pipe smoking and for any other tobacco use lowered the numbers from 23.8 percent to 23.4 and 22.2 percent. respectively. among the 20-24-year-olds: from 37.2 percent to 35.6 and 34.3 percent. respectively.among 25-G-year-olds; from 4Y.2 percent to 46.4 and 35.0 percent. respectively. among 4%64-year-olds; and from 69.1 percent to 66.2 percent and 62.8 percent, respectively, among those 65years-old and older (Table 4) (OSH. unpublished data). A detailed analysis of trends in the quit ratio by age for the period 197-I through I087 has been completed (Novotny et al., in press). Differences in quit ratios between age groups may reflect actual differences in quitting activity by age-that is. older persons may be more prone to quit and maintain abstinence than younger smokers, perhaps because of the occurrence of smoking- related symptoms or illness. However, continuing smokers are less likely than former smokers to survive to old age (Chapter 3): this selective mortality will artifactually increase the quit ratio among older age groups. TABLE 4.-Effect of adjusting for use of other tobacco products on quit ratio (percentage of ever cigarette smokers who are former cigarette smokers), 1987, NHIS, United States Quit ratio (%) Adjusting for UnadyNed" Adjustmg for cigar\/pipe\h ciE"r"/pipe\/~nuff/~hewin@ tobacco` Gender Males Females Race White5 Black> .Age (yr) xb-21 7.5-4-I 35-64 265 Education (yr I