Article Text

Dispelling myths about gender differences in smoking cessation: population data from the USA, Canada and Britain
  1. Martin J Jarvis1,
  2. Joanna E Cohen2,3,
  3. Cristine D Delnevo4,5,
  4. Gary A Giovino6
  1. 1Department of Epidemiology and Public Health, University College London, London, UK
  2. 2Ontario Tobacco Research Unit, University of Toronto, Toronto, Canada
  3. 3Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, USA
  4. 4Department of Health Education and Behavioral Science, University of Medicine and Dentistry of New Jersey, New Brunswick, New Jersey, USA
  5. 5Center for Tobacco Surveillance and Evaluation Research, University of Medicine and Dentistry of New Jersey, New Brunswick, New Jersey, USA
  6. 6Department of Community Health and Health Behavior, School of Public Health and Health Professions, The State University of New York, Buffalo, New York, USA
  1. Correspondence to Professor Martin J Jarvis, Emeritus Professor of Health Psychology, Department of Epidemiology and Public Health, University College London, 1-19 Torrington Place, London WC1E 6BT, UK; martin.jarvis{at}ucl.ac.uk

Abstract

Objectives Based mainly on findings from clinical settings, it has been claimed that women are less likely than men to quit smoking successfully. If true, this would have important implications for tobacco control interventions. The authors aimed to test this possibility using data from general population surveys.

Methods The authors used data from major national surveys conducted in 2006–2007 in the USA (Tobacco Use Supplement to the Current Population Survey), Canada (Canadian Tobacco Use Monitoring Survey) and the UK (General Household Survey) to estimate rates of smoking cessation by age in men and women.

Results The authors found a pattern of gender differences in smoking cessation which was consistent across countries. Below age 50, women were more likely to have given up smoking completely than men, while among older age groups, men were more likely to have quit than women. Across all age groups, there was relatively little difference in cessation between the sexes.

Conclusions Conclusions about gender differences in smoking cessation should be based on evidence from the general population rather than from atypical clinical samples. This study has found convincing evidence that men in general are not more likely to quit smoking successfully than women.

  • Smoking cessation
  • gender
  • quit ratios
  • harm reduction
  • prevalence
  • environmental tobacco smoke
  • cessation
  • ethics
  • policymakers
  • ideology
  • evaluation
  • advertising and promotion
  • non-cigarette tobacco products
  • surveillance and monitoring
  • smoking-caused disease
  • surveillance
  • environment

Statistics from Altmetric.com

Introduction

There is continuing debate over whether there are consistent gender differences in smoking cessation. Higher rates of successful quitting in men than in women have been observed in trials of nicotine replacement therapy1 ,2 and of bupropion,3 as well as in studies of behavioural counselling.4 The proportion of ever-smokers who have quit cigarettes (cigarette quit ratio5) has been reported to be higher in men than in women in the USA.6 Women attending NHS Stop Smoking Services in the UK were less successful in quitting than men.7 Observations such as these have led some authors to conclude that women are generally less successful at quitting smoking than men.8–10 Others have questioned this interpretation,11 ,12 arguing first that gender differences observed in clinical trials may be attributable to differences between male and female help-seekers and therefore not necessarily generalisable to the general population of smokers; second, use of the crude cigarette quit ratio is criticised as introducing a gender bias, as it fails to account for the continuing use of cigars, pipes and other tobacco products in ex-cigarette smokers, behaviours which are typically observed in men but not in women.

Analyses of gender differences in smoking cessation in Britain, which adjusted the cigarette quit ratio for current pipe and cigar smoking in ex-cigarette smokers, have found no clear evidence for overall sex differences in successful smoking cessation.11 ,12 Women tended to be more likely than men to quit in young adulthood, while among those in their 50s and older, higher quit ratios were observed in men than in women. An analysis of US data also found that allowing for use of tobacco products other than cigarettes greatly attenuated the apparent male advantage in cessation5 However, these studies were conducted 20 years or more ago. It is possible that different patterns may be observed in different countries and that patterns of quitting by gender may be changing over time. We here compare quit ratios in men and women in three countries, the USA, Canada and Britain, using data from major national surveys conducted in 2006–2007 in each country. We use the cigarette quit ratio adjusted for current pipe/cigar use as our measure of successful cessation and also explore the effect of adjusting for current use of snuff and chewing tobacco where data are available. Since most misreporting of ex-smoking status is found among smokers who reported giving up in the previous 12 months, we test the sensitivity of our findings to defining as ex-smokers only those who report having given up more than a year ago.

Methods

Data sources

USA

We used data from the 2006–2007 Tobacco Use Supplement to the Current Population Survey, combining data from fieldwork conducted in May and August 2006 and February 2007 and generating a total sample of 70 189 ever-smokers of 100+ cigarettes who were interviewed in person. The Tobacco Use Supplement to the Current Population Survey is co-sponsored by the National Cancer Institute and the Centers for Disease Control and Prevention and is a key source of national- and state-level data on smoking and other tobacco use in the USA. It is a household survey of representative samples of the civilian non-institutionalised population aged 15 years and older. Full methodological details are available on the US Census Bureau website (http://www.census.gov/apsd/techdoc/cps/cps-main.html).13

Canada

Data from the 2006 and 2007 annual Canadian Tobacco Use Monitoring Survey (CTUMS) conducted by Statistics Canada were combined, giving a total sample of 17 504 ever-smokers of 100+ cigarettes. CTUMS targets a nationally representative sample of respondents aged 15 years and older, using computer-assisted telephone interviewing. Full methodological details are available on the Statistics Canada website (http://www.census.gov/apsd/techdoc/cps/cpsmayaug06.pdf).

Great Britain

Data from the 2006 and 2007 General Household Survey (GHS) were combined, yielding a sample of 14 874 ever-regular cigarette smokers aged 16 years and older. The GHS is a continuous cross-sectional survey based each year on a sample of the general population living in private households who are interviewed in their homes. It is the primary source of national smoking prevalence estimates. Full details on sample design and response rates of the GHS are given in published reports (http://www.ons.gov.uk/ons/rel/ghs/general-household-survey/2007-report/index.html).

Adjusted quit ratio

In the US and Canadian surveys, an ex-cigarette smoker was someone who had smoked over 100 cigarettes in their lifetime and who reported no current cigarette use. The cigarette quit ratio was calculated as ex-cigarette smokers as a percentage of current daily and less than daily cigarette smokers + ex-smokers, and the adjusted quit ratio reassigned current (every day or some days) smokers of pipes and cigars to the current smoking category. In analyses to further adjust for use of snuff and chewing tobacco, we reassigned ex-cigarette smokers who reported every day or some day use of snuff/chew to the current use category before calculating the quit ratio. In the GHS, current cigarette smokers are defined as those responding ‘yes’ to the question ‘Do you smoke cigarettes at all nowadays?’ and ex-cigarette smokers as those who respond ‘no’ to this question but ‘yes’ to the question ‘Have you ever smoked cigarettes regularly?’ Further items enquire about use of pipes (‘at all nowadays’) and cigars (‘at least one of any kind per month nowadays’), but use of smokeless tobacco is not assessed. Cigar use is ascertained in both sexes, but only men are asked about pipe smoking. The adjusted quit ratio was defined as ex-cigarette smokers (less those who were current pipe/cigar smokers) as a percentage of current + ex-smokers.

Statistical analysis

We calculated adjusted quit ratios by age group and by sex applying supplied weights for each survey using SPSS V.18. We further calculated adjusted quit ratios for a definition of ex-smoking that reassigned past-year quitters from the ex-smoking to the current smoking category.

We estimated 95% CIs using methods appropriate for the complex design of each survey. For the US data, we used SUDAAN survey software. Sample weights, which correct for the varying probabilities of selection, were applied, and variance was computed using replicate weights provided by the National Cancer Institute. The UK data were analysed with the complex survey module in SPSS, which adjusted for stratification and clustering in the sample design. For the Canadian data, we followed the methods recommended in the CTUMS 2006 and 2007 user guide and the supplied coefficient of variation tables.14

Results

Table 1 presents adjusted quit ratios and their 95% CIs by age group in each country and also shows estimated quit ratios when those who reported giving up smoking in the past year are reclassified as continuing smokers. Across all ages combined, the quit ratio was somewhat higher in women than in men in the USA (50.7% in women vs 48.8% in men) and Canada (57.5% vs 56.9%) and slightly lower in Britain (50.7% vs 52.4%). The female advantage in cessation was statistically significant in the USA, but the difference by gender in Canada and the UK did not reach significance. However, despite these modest overall differences, there were marked gender differences in particular age groups. Among younger adults, there was a clear tendency, evident in all three countries, for women to be more likely to have quit smoking than men. This was especially evident in those in their 20s and 30s but was also seen in those in their 40s. Among older age groups, this pattern was reversed, with men aged 60 years and older being more likely to have quit than women. As shown by the lack of overlap between 95% CIs, in the USA, women were significantly more likely than men to have quit in all age groups between 20 and 60 years, and the male advantage above age 60 was not significant. In the UK and Canada, the higher rate of cessation in women than in men was significant only in the age group 20–29 years. These age-related patterns of gender differences are shown in figure 1.

Table 1

Quit ratios (%) (95% CI) adjusted for cigar/pipe smoking by age and sex in the USA, Canada and Great Britain in 2006–2007

Figure 1

Gender differences in quitting (female quit ratio–male quit ratio) by age group in the USA, Canada and Britain. This figure is produced in colour in the online journal—please visit the website to view the colour figure.

Re-classifying ex-smokers of <1 year as continuing smokers disproportionately reduced estimated quit ratios in younger age groups, as relatively more of their quitting had occurred recently. However, the higher rate of cessation in younger women than men was still evident in all three countries after this adjustment. Across all age groups combined, there were only small gender differences in cessation when past-year quitters were counted as continuing smokers (46.4% in women vs 45% in men in the USA; 52.9% vs 53.8% in Canada; and 45.5% vs 47.7% in Britain). The female advantage in cessation remained statistically significant in the USA.

Table 2 shows the effects of further adjusting the quit ratio to exclude current users of smokeless tobacco from the category of quitters. This adjustment, which was possible only for the US and Canadian surveys, made little difference to Canadian estimates. In the USA, it reduced the observed quit ratio more in men than in women, as would be expected given men's higher use of oral snuff and chewing tobacco. The all-ages quit ratio in the USA adjusted for smokeless use was 50.6% in women vs 46.6% in men, and 46.3% vs 43.1% considering only those quitting more than 1 year ago as successful quitters.

Table 2

Quit ratios (%) (95% CI) additionally adjusted for smokeless tobacco use by age and sex in the USA and Canada in 2006–2007

Discussion

Our findings show a pattern of gender differences in smoking cessation which is consistent across three Western countries. Female ever-smokers in their 20s, 30s and 40s are more likely to have given up smoking completely than are their male counterparts, while among older age groups, men appear more likely to have quit than women. Across all age groups combined, there is relatively little difference in cessation between the sexes, with women significantly more likely to have quit than men in the USA and slightly but not significantly in Canada, and men slightly more likely than women in Britain. The findings largely replicate previous observations from British data gathered 20 years or more ago.11 ,12

Our study has a number of strengths. We used recent data from three countries gathered at the same point in time in major national surveys. The design of these surveys, which in each country are the source of national estimates of smoking prevalence, ensures that findings can be generalised to the general population of smokers. We conducted sensitivity analyses to demonstrate that our findings are robust to adjustment which treats only those who report quitting more than 1 year ago as ex-smokers.

There are also some potential limitations. The quit ratio, which is derived from analysis of cross-sectional survey data, is subject to a number of potential biases. It relies on self-reports of individuals' smoking status and history. To the extent that these are inaccurate, the quit ratio will also be inaccurate. It is also influenced by the uptake rate. For example, if a group has a much higher uptake rate than another group in recent years, the quit ratio for that group will be lower even if the annual cessation rate for current smokers is the same for the two groups. Since smoking uptake is essentially complete by the mid-20s, this is likely to be relevant only for the youngest age group in our analyses. People who have been ex-smokers for many years may forget that they ever smoked or cease to think of themselves as having smoked. There may also be some inaccuracy in reports of having quit: this is most likely in those who report having quit most recently. However, these acknowledged sources of imprecision will only bias gender comparisons of smoking cessation if they apply differentially to men and women, and there appears to be no reason to think that this might be the case. Comparisons of patterns of smoking cessation derived from longitudinal follow-up of a cohort of smokers and from estimating quit ratios from cross-sectional survey data have shown very similar pictures.15

Age differences in cessation may reflect the natural history of smoking cessation across the life course but might also reflect differing patterns in cohorts whose life experience has been very different: it is not necessarily the case that cessation by age 60 will be the same in those who are now in their 20s as in the current generation of 60-year-olds. Differences between the sexes in older adults are particularly difficult to interpret, especially since older female smokers typically started to smoke several years later than their male counterparts.12 Those who have died of smoking-related disease are not available for survey, and since men in Western countries are more likely to have died of smoking, ever-smoking men may be systematically under-represented in survey samples. These are samples of survivors, rather than all those born in particular years. This differential loss of male smokers could inflate the estimated quit ratio for older men and be an important factor explaining the apparent higher male cessation rate in older groups.

We used the quit ratio adjusted for current use of pipes or cigars rather than the simple cigarette quit ratio. The argument for this is, we think, compelling. Just as in a trial of smoking cessation, cigarette smokers who quit cigarettes but smoke cigars or pipes are treated as treatment failures rather than successful quitters,16 so in cross-sectional surveys, ex-cigarette smokers who use cigars or pipes should be regarded as continuing smokers, not ex-smokers. To use the simple cigarette quit ratio would introduce an important gender bias as men are much more likely than women to use these alternative tobacco products.17

How can our findings be reconciled with the frequent observation from clinical trials of higher rates of successful cessation in men than in women? It is important to recognise that most smokers do not seek formal help with stopping smoking, and most successful cessation is unaided. Men are less likely than women to enrol in treatment programmes, raising the possibility that help-seekers are not representative of men and women generally. Young adult smokers, among whom the female advantage in cessation is most evident, tend to be under-represented in clinical trials. It might also be the case that female smokers could make more cessation attempts, but be relatively less likely to succeed on any given attempt, with the net outcome being overall parity in successful cessation with men. However, a recent meta-analysis of smoking cessation in general population samples has found no gender differences in either quit attempts or their likelihood of success.18

Whatever the explanation for the diversity of findings from clinical settings and from population samples, it is important to recognise that conclusions about smoking cessation in men and women in general should be based on evidence from the general population rather than from the atypical samples found in clinic studies. Our study has found convincing evidence that men in general are not more likely to quit smoking successfully than women. The myth of female disadvantage at quitting smoking is bad first and foremost for women. If women, particularly young women, are misled into believing that they are less likely to give up smoking successfully than men, it could undermine their confidence in quitting and foster a self-fulfilling prophecy. It is also bad for men, as they may unjustifiably feel that, hard though quitting is, at least they are better at it than women. It is bad for gender stereotypes, in a world where inaccurate stereotypes are rife. It is bad for cessation therapists and counsellors, who need to be armed with accurate knowledge as they seek to help smokers. It is bad, finally, for policymakers, as they may be misled into designing interventions for an imaginary problem rather than the many real challenges of increasing rates of cessation. It is time to put aside the idea that women are less successful than men at giving up smoking.

What this paper adds

  • The idea that women are less successful than men at quitting smoking successfully continues to be expressed by some researchers.

  • This study, which uses data from major national surveys from the USA, Canada and the UK, finds no support for this.

  • In all three countries, younger women were more likely than men to have quit smoking, and across all ages, there was little difference in cessation by gender.

  • Policymakers should not be misled that women are at a disadvantage to men when it comes to quitting success.

References

Supplementary materials

Footnotes

  • Competing interests None.

  • Ethics approval Appropriate national bodies in each country.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.