PREVENTIVE
MEDICINE
21, 311-319 (1992)
Characteristics DAVID
WILSON,
M.P.H.,
of Heavy Smokers *J MELANIE
NEVILLE OWEN, PH.D.,?
WAKEFIELD,
M.A.,*
AND LYN ROBERTS,PH.D.S
*Behavioural Epidemiology Unit, South Australian Health Commission, P.O. Box 6, Rundle Mall, South Australia 5000, Australia; tDepartment of Communi@ Medicine, University of Adelaide. G.P.O. Box 498, Adelaide, South Australia 5001, Australia; and #South Australian Smoking and Health Project, P.O. Box 160, North Adelaide, South Australia 5006, Australia Background. Heavy smokers(thosewho smoke 225 or more cigarettes a day) are a subgroup whoplacethemselves and others at risk for harmful health consequences and also
are those least likely to achieve cessation. Despite this, heavy smokers are not well described as a segment of the smoking population. Merhods. We used representative population data on 1,048 smokers to examine differences between heavy and lighter smokers (~24 cigarettes per day). Results. Heavy smokers constituted 26.7% of all cigarette smokers. Compared with lighter smokers, heavy smokers were significantly more likely to be male, to be age 30 years or older, to smoke their first cigarette of the day within 30 min of waking, to perceive quitting as very difftcult, to have little confidence in their ability to quit, to be less likely to report variation in their rate of smoking between work and leisure days, and to be less likely to be employed. Conclusions. Public health strategies which may particularly assist heavy smokers include stronger restrictions on smoking in public places, nicotine replacement therapies, and the use of segmentation research to more carefully target campaign messages to influence quit attempts and confidence. 0 1992 Academic Press, Inc.
INTRODUCTION
Heavy smokers (those who smoke ~25 cigarettes per day) typically find it hardest to quit smoking (1). Among smokers, this is the group whose health is most at risk from the effects of smoking and who most expose family, friends, and workmates to the risks of passive smoking. In spite of these concerns, little is known about this group of smokers that would assist in the targeting of public health campaigns. Most studies of heavy smokers have not used representative population data, but have relied upon convenience samples drawn from clinical, workplace, or student populations, or have recruited study subjects through newspaper advertisements or community service announcements (2-5). Much of the research on heavy smokers has involved trials of interventions to manipulate nicotine intake (6,7) or to test the effects of drugs on alleviating nicotine withdrawal (8,9). These studies have suggested that heavy smokers are more driven to regulate their nicotine intake and therefore find it more difficult to quit (1). In addition, some studies have generally concluded that other drug use is common in this group (3, 4). Other research has found that heavy smokers have more illness episodes, a lower commitment to health, greater anxiety levels, and a weaker response to ’ To whom reprint requests should be addressed. 311 0091-7435192 $5.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
312
WILSON
ET
AL.
external cues (2, 5, 10-12). Compared with lighter smokers, heavy smokers have reported higher levels of environmental stressors, fewer supportive social resources, and poorer psychological functioning (13). Heavy smokers have also been found more likely to be in the precontemplation, rather than the contemplation or preparation stages of quitting (14). Heavy smoking has additionally been associated with heavy drinking habits (15). Most smokers who quit do so on their own, without recourse to formal treatment programs (16, 17). Even though heavy smokers are more likely to attend smoking cessation programs (17) and are more likely to express a preference for some form of assistance with cessation (18), the majority, around SO%, do not make use of such programs (17). Studies based on smokers participating in treatment programs therefore draw conclusions based on the attitudes and behaviors of a self-selected minority of heavy smokers, who are motivated to quit. In contrast, public health campaigns are aimed at the majority of smokers, where the motivation to quit may vary. Recent years have seen a proliferation of mass communication campaigns, which have the potential to reach the majority of smokers in a population. To maximize their efficacy, such campaigns need to be informed about the characteristics and needs of subgroups of smokers, and to plan their intervention strategies using representative segmentation data. Using data from a representative population survey, we compared the characteristics of heavy smokers (a25 cigarettes a day) to those of lighter smokers (~25 per day). METHODS The Australian Bureau of Statistics conducted a survey on behalf of the South Australian Smoking and Health Project in late 1989. This project is a statewide multimedia education program, the aims of which are to encourage and support smoking cessation and prevent the initiation of smoking. To fulfill the research and evaluation requirements of the Project, annual cross-sectional surveys of a representative population sample are undertaken. Apart from monitoring smoking prevalence, these data are used to describe different smoker segments and assist in the development of appropriate services and resources. The 1989 survey was a multistage systematic clustered area sample of 0.44% of South Australian dwellings. All persons in each selected household age 15 years and over were eligible for interview. There was no replacement for nonresidents. Up to five call-back visits for urban residents and three for rural dwellers were undertaken, giving a final response rate of 89%, or a final sample size of 3,767 persons age 15 years and over, weighted according to age and sex. The survey took the form of a personal interview in the home of respondents by trained interviewers. A person was classified as a smoker if they reported smoking one or more cigarettes per day. The smoking rate was derived from the reported number of cigarettes smoked on work and leisure days. Cigarette smokers were asked “Does the number of cigarettes you smoke vary between work days and leisure days?” Those who responded affirmatively to this question were then asked to report how many cigarettes they smoked on a typical work day and how many were smoked
HEAVY
SMOKERS
313
on a typical leisure day. For smokers whose consumption varied between work and leisure days, a weighted average of their work and leisure day cigarette consumption was calculated. Those whose consumption did not vary or who did not work were asked how many cigarettes they smoked on average each day. Heavy smokers were defined according to the criterion adopted by the United States Surgeon General-smoking 325 cigarettes per day (19). Confidence (self-efficacy) was assessed by asking “How confident are you that you could quit smoking in the next 3 months?” Response options were “very confident; fairly confident; not at all confident; not interested in quitting.” Perceived difficulty of cessation was assessed by the question “How difficult do you think it would be to stop smoking ?” Response options were “very difficult; fairly difficult; not at all difficult; never intend to stop.” An index of the perceived health risks of smoking was constructed with a range from 0 to 9, based on responses to nine health problems which were related to smoking. Stage of change was assessed using a IO-point scale developed by Biener and Abrams (20), with the zero point labeled “no thought of quitting,” point 2 labeled “think I need to consider quitting someday,” point 5 labeled “think I should quit, but not quite ready,” point 8 labeled “starting to think about how to change my smoking patterns,” and point 10 labeled “taking action to quit (e.g., cutting down, enrolling in a program).” It was also possible for respondents to indicate numbers between these main scale points. Three categories were derived which corresponded to the stage-of-change model, these being precontemplation (point zero), contemplation (points 1 to 6) and preparation (points 7 to 10) (14). An item from the Fagerstrom Tolerance Questionnaire (2 1) that asked how soon after waking the first cigarette was smoked and that was identified as a powerful predictor of nicotine addiction was used as a measure of dependence. Respondents were also asked whether they had ever tried to quit smoking. Finally, interviewers obtained information on respondents’ age, gender, maritaI status, employment status, educational attainment, and, if employed, occupational status. Data were analyzed using the SPSS-PC statistical package. Conventional x2 tests were used to test the association between cigarette consumption and variables of interest at the univariate stage. Variables which were significantly related to cigarette consumption were then entered into an unconditional logistic regression, to examine which of the attributes of smokers best jointly predicted being in the heavy smoker category. Associations between predictor variables were examined to ensure that the model was not subject to multicollinearity. Also, stratified analyses were used to check for homogeneity of associations across different levels of the demographic variables. RESULTS
Of the 3,767 respondents, 1,048 (27.8%) were current smokers. This corresponds closely with the national prevalence estimate for 1989 (22). Among the smokers, 26.7% smoked 225 cigarettes a day and thus were classified as heavy smokers. In the univariate analyses (x2 tests with 1 d’ 12 = 1,028), a number of variables
314
WILSON
ET AL.
were found to be associated with being a heavy smoker (see Table 1). Compared with lighter smokers, heavy smokers were more likely to be male; to be age 30 years or older; to smoke the first cigarette of the day within 30 min of waking; to perceive quitting as very difficult; and to have little confidence in their ability to quit. Furthermore, heavy smokers were significantly less likely than lighter smokers to be employed; to have attained a tertiary level of education; to vary cigarette consumption between work and leisure days; and to be in the preparation stage of quitting smoking. Cigarette consumption was not associated with marital status, occupational status, perception of health risks associated with smoking, or the likelihood of a past attempt to quit. Prior to undertaking multivariate analyses, associations between predictor variables were checked to ensure the logistic regression model was not subject to multicollinearity. The highest measure of association found was between perceived difficulty in quitting and confidence in the ability to quit smoking in the following 3 months (4 = 0.56). The data were also stratified by age and sex to check for homogeneity of associations across different values of these demographic variables. A test of interaction was used to assess interaction effect when the data were stratified (23). The strength of association was found to vary by sex for both confidence in the ability to quit smoking and employment status. Those who lacked confidence in their ability to quit smoking were more likely to be heavy smokers; however, this relationship was more pronounced among females (x2 = 5.06, P < 0.05). The probability of nonemployed people being heavy smokers, already greater than that for those who were employed, was increased if they were female (x2 = 4.75, P < 0.05). TABLE ASSOCIATIONS
BETWEEN
1
CIGARETTE CONSUMPTION VARIABLES OF INTEREST
Lighter (l-24 cpd) Variable Sex (% male) Age (% 30+ years) Marital status (% married or de facto) Employment status (% employed) Occupational prestige (% low) Educational attainment (% tertiary) Work/leisure day variation (% no) Time to first cigarette (% less than 30 min) Health knowledge (% low) Ever tried to quit (% no) Perceived difficulty of quitting (% very) Confidence can quit (% not at all) Quit stage (% action) a Occupational
(HEAVY
vs LIGHTER)
AND
n = 154
Heavy (25+ cpd) n = 274
P
53.3 53.5 55.4 69.0 36.1 25.4 42.4
64.4 71.2 60.0 54.1 32.7 14.7 63.6
0.002
MEDICINE
21, 311-319 (1992)
Characteristics DAVID
WILSON,
M.P.H.,
of Heavy Smokers *J MELANIE
NEVILLE OWEN, PH.D.,?
WAKEFIELD,
M.A.,*
AND LYN ROBERTS,PH.D.S
*Behavioural Epidemiology Unit, South Australian Health Commission, P.O. Box 6, Rundle Mall, South Australia 5000, Australia; tDepartment of Communi@ Medicine, University of Adelaide. G.P.O. Box 498, Adelaide, South Australia 5001, Australia; and #South Australian Smoking and Health Project, P.O. Box 160, North Adelaide, South Australia 5006, Australia Background. Heavy smokers(thosewho smoke 225 or more cigarettes a day) are a subgroup whoplacethemselves and others at risk for harmful health consequences and also
are those least likely to achieve cessation. Despite this, heavy smokers are not well described as a segment of the smoking population. Merhods. We used representative population data on 1,048 smokers to examine differences between heavy and lighter smokers (~24 cigarettes per day). Results. Heavy smokers constituted 26.7% of all cigarette smokers. Compared with lighter smokers, heavy smokers were significantly more likely to be male, to be age 30 years or older, to smoke their first cigarette of the day within 30 min of waking, to perceive quitting as very difftcult, to have little confidence in their ability to quit, to be less likely to report variation in their rate of smoking between work and leisure days, and to be less likely to be employed. Conclusions. Public health strategies which may particularly assist heavy smokers include stronger restrictions on smoking in public places, nicotine replacement therapies, and the use of segmentation research to more carefully target campaign messages to influence quit attempts and confidence. 0 1992 Academic Press, Inc.
INTRODUCTION
Heavy smokers (those who smoke ~25 cigarettes per day) typically find it hardest to quit smoking (1). Among smokers, this is the group whose health is most at risk from the effects of smoking and who most expose family, friends, and workmates to the risks of passive smoking. In spite of these concerns, little is known about this group of smokers that would assist in the targeting of public health campaigns. Most studies of heavy smokers have not used representative population data, but have relied upon convenience samples drawn from clinical, workplace, or student populations, or have recruited study subjects through newspaper advertisements or community service announcements (2-5). Much of the research on heavy smokers has involved trials of interventions to manipulate nicotine intake (6,7) or to test the effects of drugs on alleviating nicotine withdrawal (8,9). These studies have suggested that heavy smokers are more driven to regulate their nicotine intake and therefore find it more difficult to quit (1). In addition, some studies have generally concluded that other drug use is common in this group (3, 4). Other research has found that heavy smokers have more illness episodes, a lower commitment to health, greater anxiety levels, and a weaker response to ’ To whom reprint requests should be addressed. 311 0091-7435192 $5.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
312
WILSON
ET
AL.
external cues (2, 5, 10-12). Compared with lighter smokers, heavy smokers have reported higher levels of environmental stressors, fewer supportive social resources, and poorer psychological functioning (13). Heavy smokers have also been found more likely to be in the precontemplation, rather than the contemplation or preparation stages of quitting (14). Heavy smoking has additionally been associated with heavy drinking habits (15). Most smokers who quit do so on their own, without recourse to formal treatment programs (16, 17). Even though heavy smokers are more likely to attend smoking cessation programs (17) and are more likely to express a preference for some form of assistance with cessation (18), the majority, around SO%, do not make use of such programs (17). Studies based on smokers participating in treatment programs therefore draw conclusions based on the attitudes and behaviors of a self-selected minority of heavy smokers, who are motivated to quit. In contrast, public health campaigns are aimed at the majority of smokers, where the motivation to quit may vary. Recent years have seen a proliferation of mass communication campaigns, which have the potential to reach the majority of smokers in a population. To maximize their efficacy, such campaigns need to be informed about the characteristics and needs of subgroups of smokers, and to plan their intervention strategies using representative segmentation data. Using data from a representative population survey, we compared the characteristics of heavy smokers (a25 cigarettes a day) to those of lighter smokers (~25 per day). METHODS The Australian Bureau of Statistics conducted a survey on behalf of the South Australian Smoking and Health Project in late 1989. This project is a statewide multimedia education program, the aims of which are to encourage and support smoking cessation and prevent the initiation of smoking. To fulfill the research and evaluation requirements of the Project, annual cross-sectional surveys of a representative population sample are undertaken. Apart from monitoring smoking prevalence, these data are used to describe different smoker segments and assist in the development of appropriate services and resources. The 1989 survey was a multistage systematic clustered area sample of 0.44% of South Australian dwellings. All persons in each selected household age 15 years and over were eligible for interview. There was no replacement for nonresidents. Up to five call-back visits for urban residents and three for rural dwellers were undertaken, giving a final response rate of 89%, or a final sample size of 3,767 persons age 15 years and over, weighted according to age and sex. The survey took the form of a personal interview in the home of respondents by trained interviewers. A person was classified as a smoker if they reported smoking one or more cigarettes per day. The smoking rate was derived from the reported number of cigarettes smoked on work and leisure days. Cigarette smokers were asked “Does the number of cigarettes you smoke vary between work days and leisure days?” Those who responded affirmatively to this question were then asked to report how many cigarettes they smoked on a typical work day and how many were smoked
HEAVY
SMOKERS
313
on a typical leisure day. For smokers whose consumption varied between work and leisure days, a weighted average of their work and leisure day cigarette consumption was calculated. Those whose consumption did not vary or who did not work were asked how many cigarettes they smoked on average each day. Heavy smokers were defined according to the criterion adopted by the United States Surgeon General-smoking 325 cigarettes per day (19). Confidence (self-efficacy) was assessed by asking “How confident are you that you could quit smoking in the next 3 months?” Response options were “very confident; fairly confident; not at all confident; not interested in quitting.” Perceived difficulty of cessation was assessed by the question “How difficult do you think it would be to stop smoking ?” Response options were “very difficult; fairly difficult; not at all difficult; never intend to stop.” An index of the perceived health risks of smoking was constructed with a range from 0 to 9, based on responses to nine health problems which were related to smoking. Stage of change was assessed using a IO-point scale developed by Biener and Abrams (20), with the zero point labeled “no thought of quitting,” point 2 labeled “think I need to consider quitting someday,” point 5 labeled “think I should quit, but not quite ready,” point 8 labeled “starting to think about how to change my smoking patterns,” and point 10 labeled “taking action to quit (e.g., cutting down, enrolling in a program).” It was also possible for respondents to indicate numbers between these main scale points. Three categories were derived which corresponded to the stage-of-change model, these being precontemplation (point zero), contemplation (points 1 to 6) and preparation (points 7 to 10) (14). An item from the Fagerstrom Tolerance Questionnaire (2 1) that asked how soon after waking the first cigarette was smoked and that was identified as a powerful predictor of nicotine addiction was used as a measure of dependence. Respondents were also asked whether they had ever tried to quit smoking. Finally, interviewers obtained information on respondents’ age, gender, maritaI status, employment status, educational attainment, and, if employed, occupational status. Data were analyzed using the SPSS-PC statistical package. Conventional x2 tests were used to test the association between cigarette consumption and variables of interest at the univariate stage. Variables which were significantly related to cigarette consumption were then entered into an unconditional logistic regression, to examine which of the attributes of smokers best jointly predicted being in the heavy smoker category. Associations between predictor variables were examined to ensure that the model was not subject to multicollinearity. Also, stratified analyses were used to check for homogeneity of associations across different levels of the demographic variables. RESULTS
Of the 3,767 respondents, 1,048 (27.8%) were current smokers. This corresponds closely with the national prevalence estimate for 1989 (22). Among the smokers, 26.7% smoked 225 cigarettes a day and thus were classified as heavy smokers. In the univariate analyses (x2 tests with 1 d’ 12 = 1,028), a number of variables
314
WILSON
ET AL.
were found to be associated with being a heavy smoker (see Table 1). Compared with lighter smokers, heavy smokers were more likely to be male; to be age 30 years or older; to smoke the first cigarette of the day within 30 min of waking; to perceive quitting as very difficult; and to have little confidence in their ability to quit. Furthermore, heavy smokers were significantly less likely than lighter smokers to be employed; to have attained a tertiary level of education; to vary cigarette consumption between work and leisure days; and to be in the preparation stage of quitting smoking. Cigarette consumption was not associated with marital status, occupational status, perception of health risks associated with smoking, or the likelihood of a past attempt to quit. Prior to undertaking multivariate analyses, associations between predictor variables were checked to ensure the logistic regression model was not subject to multicollinearity. The highest measure of association found was between perceived difficulty in quitting and confidence in the ability to quit smoking in the following 3 months (4 = 0.56). The data were also stratified by age and sex to check for homogeneity of associations across different values of these demographic variables. A test of interaction was used to assess interaction effect when the data were stratified (23). The strength of association was found to vary by sex for both confidence in the ability to quit smoking and employment status. Those who lacked confidence in their ability to quit smoking were more likely to be heavy smokers; however, this relationship was more pronounced among females (x2 = 5.06, P < 0.05). The probability of nonemployed people being heavy smokers, already greater than that for those who were employed, was increased if they were female (x2 = 4.75, P < 0.05). TABLE ASSOCIATIONS
BETWEEN
1
CIGARETTE CONSUMPTION VARIABLES OF INTEREST
Lighter (l-24 cpd) Variable Sex (% male) Age (% 30+ years) Marital status (% married or de facto) Employment status (% employed) Occupational prestige (% low) Educational attainment (% tertiary) Work/leisure day variation (% no) Time to first cigarette (% less than 30 min) Health knowledge (% low) Ever tried to quit (% no) Perceived difficulty of quitting (% very) Confidence can quit (% not at all) Quit stage (% action) a Occupational
(HEAVY
vs LIGHTER)
AND
n = 154
Heavy (25+ cpd) n = 274
P
53.3 53.5 55.4 69.0 36.1 25.4 42.4
64.4 71.2 60.0 54.1 32.7 14.7 63.6
0.002
