Alcohol and Alcoholism Advance Access originally published online on October 12, 2007
Alcohol and Alcoholism 2008 43(1):91-96; doi:10.1093/alcalc/agm138
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Self-reported never-drinkers in England 1994–2003: Characteristics and trends in adults aged 18–54 years
1 London School of Hygiene & Tropical Medicine
2 University College, London
3 Middlesex University, UK
* Author to whom correspondence should be addressed at: PEHRU, LSHTM, Keppel Street, London WC1E 7HT, UK. Tel: 0207 927 2316/01223 244 869; E-mail: sam.pattenden{at}lshtm.ac.uk
Received 19 June 2007; first review notified 12 July 2007; in revised form 30 July 2007; accepted 23 August 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT IntroductionMethodsResultsDiscussionReferences |
|---|
Aims: This paper describes prevalences, time-trends and characteristics of self-reported never-drinkers, during the period 1994–2003, focussing particularly on white adults aged 18–54. Methods: Data on 122,809 adults (18 + ) were obtained from the Health Survey for England (HSfE). Logistic regressions were used to estimate time trends in self-reported never-drinking, and associations between never-drinking and living alone, and educational qualification. Analyses were stratified by gender, age group and period. Results: The overall proportion of white, female never-drinkers was 5.5%, rising monotonically with age. Proportions among men were much lower, with the lowest proportion (1.1%) in the 30–54 age group. Odds of never-drinking increased by 3% per year in those aged 30–54, a trend not explained by any covariates. Smaller increases were seen among those aged 18–29. Never-drinking was strongly associated with living with another adult and with lower qualification. The association with qualification increased over time among young women, and the association with living with another adult increased among men aged 30–54. Conclusions: Never-drinkers are a significant minority in England, whose prevalence rose, between 1994 and 2003, among adults aged under 55 years. The prevalence varies considerably by age, sex, and social characteristics, and the social discrepancies in never-drinking appear to be widening.
|
Introduction |
|---|
|
TOPABSTRACTIntroductionMethodsResultsDiscussionReferences |
|---|
Alcohol consumption is a topic of current concern in the UK, especially in relation to young people, and has come under considerable scrutiny. By contrast, lifetime abstainers, comprising a small but not insignificant fraction of the population, are infrequently studied. The literature does provide some information on abstainers, a selection of which is reviewed below. However, abstainers are often considered a homogenous group, and reported figures frequently include ex-drinkers (Goldman and Najman, 1984
; Fillmore et al., 2006).
Abstinence has, with some consistency, been reported as higher among women than among men (Academy of Medical Sciences, 2004; Pollack et al., 2005; Rahav et al., 2006; Silveira et al., 2007). A recent European study found women to be more likely than men to abstain in 10 out of 13 countries, though with no gender differences in Finland, Iceland or Norway (Makela et al., 2006). Rahav et al. interestingly noted that ‘differences between countries in the gender gap in drinking [are] strongly associated with women's position in society, as well as with modernization.’ (Rahav et al., 2006). Reasons for gender differences are also discussed by Wilsnack et al. (2000). In a large international WHO review, the greatest gender discrepancies were seen in Greece, the Philippines, and Armenia with, respectively, current abstention rates among women 13, 7 and 6 times those in men (Rehm et al., 2004). The lowest female/male ratios were seen in Fiji, India, and Sri Lanka, with abstention very high for both sexes; in Russia, Finland, and Germany, with fairly low abstention rates; and in Nigeria, where about 50% of both genders were abstainers. An unusual pattern was reported for Namibia, of 61% current male abstainers and 47% female. (We refer only to non-imputed proportions in this WHO report.) In a study looking at drinking patterns in eight low- and middle-income countries, some results differed from those in the Rehm report (Obot and Room, 2005). Gender ratios for abstention ranged from close to 1 in Uganda, Brazil, and Nigeria, to 2.6 (female/male), 4.3, and 4.5 in Argentina, Mexico, and Sri Lanka.
Non-drinking has been associated with increasing age, particularly in women (Emberson et al., 2005; Makela et al., 2006). A steep rise in abstention with age was reported among women in the UK, with a less clear pattern for men (Makela et al., 2006). The same was found in Argentina, and in Costa Rica current abstention was linked to age in both sexes (Bejarano–Orozco, 2005; Munne, 2005). However, in a Brazilian survey, age was not a strong predictor for either sex, while other surveys have shown abstention rates dropping with age for one or both sexes (Benegal et al., 2005; Ibanga et al., 2005; Kerr–Correa et al., 2005; Mendoza et al., 2005; Tumwesigye and Kasirye, 2005).
A number of studies, using various measures of socioeconomic status, have reported drinking to be more prevalent (though not necessarily more harmful) in higher socioeconomic groups (Marmot, 1997; van Oers et al., 1999; Pollack et al., 2005). An American study showed alcohol consumption to be greatest in the least-deprived areas, despite alcohol outlets being most common in the most-deprived areas, and in Brazil and Costa Rica, abstinence was higher among the lowest educated (Bejarano–Orozco, 2005; Kerr–Correa et al., 2005; Pollack et al., 2005). However, a recent international study showed mixed patterns across 15 countries of abstention by educational level (Bloomfield et al., 2006). For women, although the low-qualified group in all countries contained more abstainers than the middle or highly qualified, the proportions of abstainers did not always follow a monotonic trend across categories of qualification. For men in the UK and in several other countries, no significant inequalities were seen in abstention rates. Nor, in a separate survey, was education related to abstinence among women in Argentina (Munne, 2005). In India and Sri Lanka, abstention was higher among the more educated for one or both sexes, though, paradoxically, abstention in India was highest in the lowest income quartile (Benegal et al., 2005; Hettige and Paranagama, 2005).
Personal relationships (and lack of relationships) are greatly influential and may affect individuals' alcohol consumption. Borsari, focussing upon US college students, identified three main pathways through which relationships might act: the ‘lack or breakdown’ of quality relationships, alcohol use being integral to relationships, and, conversely, disapproval of, or abstinence from alcohol by peers (Borsari and Carey, 2006). In the Argentina survey, proportions of abstainers (male and female) were higher among married/cohabiting than among single adults, but different patterns have been seen elsewhere (Munne, 2005). In Brazil, single people were more likely to be current abstainers than married people. (Kerr–Correa et al., 2005). In Sri Lanka and India this was true only for men, with no difference between single and married Sri Lankan women, and higher abstention rates among married than single Indian women (Benegal et al., 2005; Hettige and Paranagama, 2005).
In this paper, we describe prevalences, time-trends, and characteristics of English never-drinkers aged 18–54. We assess associations between abstention and age, gender, education and whether or not living with another adult. For reasons described below, our analyses are restricted to white respondents.
|
Methods |
|---|
|
TOPABSTRACTIntroductionMethodsResultsDiscussionReferences |
|---|
Data
Health Survey for England (HSfE) data were obtained for years 1994–2003, for 122,809 adults (aged 18 + years). The HSfE is a series of annual surveys, incorporating questionnaires and nurse visits, designed to be representative of the general population. In some years, surveys are boosted to enhance participation of population sub-groups (e.g. ethnic minorities, the elderly), but in these analyses only core surveys from the general population were included. Non-white respondents made up only 6% of these core samples, yet had very different patterns of alcohol consumption. Current analyses were therefore restricted to white respondents (114,850), of whom 72,394 were aged 18–54 years.
HSfE participants were categorized into three groups, according to self-reported drinking behaviour: (1) current drinkers, including ‘occasional’ drinkers, (2) ex-drinkers, and (3) never-drinkers. Three survey questions were used to derive these. Precise wording varied slightly between years. Those for 1994 were:
- A: Do you ever drink alcohol nowadays, including drinks you brew or make at home?
- B: Could I just check, does that mean you never have an alcoholic drink nowadays, or do you have an alcoholic drink very occasionally, perhaps for medicinal purposes or on special occasions like Christmas and New Year?
- C: Have you always been a non-drinker, or did you stop drinking for some reason?
- B: Could I just check, does that mean you never have an alcoholic drink nowadays, or do you have an alcoholic drink very occasionally, perhaps for medicinal purposes or on special occasions like Christmas and New Year?
Our focus is on category (3), the self-reported lifetime abstainers.
Four age-groups were defined: 18–29, 30–54, 55–64 and 65+ years. Proportions of current-, ex-, and never-drinkers are given for all four age-groups. The main analyses are restricted to those aged under 55 years.
Educational qualification was categorized into four groups: above ‘A’ level, GCSE grades a–c, and below GCSE grade c (or their equivalents). This variable was chosen as a socioeconomic measure, which would be comparable with several previous studies. Rather than use it as a socioeconomic ‘catch-all’, we controlled for several other socioeconomic or related factors (tenure, car access, also smoking, area of residence, and region; see below).
Questions regarding marital or relationship status could not easily be standardized across the years. The nearest equivalent was to construct a binary variable coded 1 if the respondent lived with someone else aged 16+ years, and 0 if not. This variable is sometimes referred to below, for brevity, as ‘living alone’, though the individual may be living with children aged 15 years or less.
Statistical analyses
All results are presented stratified by gender and age group, and some are presented combined as well. Numbers of self-reported never-drinkers, ex-drinkers, and current drinkers were tabulated. All other analyses used a multivariable model. This simultaneously included age (except where stratifying by age), sex (except where stratifying by sex), living alone, and qualification. These analyses also controlled for tenure (owner-occupier, local authority/housing association, privately renting), car access, region, smoking habit (never regularly smokes, ex-smoker, smokes < 10 cigarettes/day, 10–19/day, or 20+ per day), and, as a random effect, sampling-area.
Percentages of never-drinkers were estimated for each category of the variables ‘lives alone’, and ‘qualification’, mutually adjusted, from the multivariable model. To see whether these characteristics changed over time, interactions were tested between the explanatory variables and a binary ‘period’ variable (1994–98, 1999–2003).
Time-trends in abstinence were estimated by including the year as a continuous variable, first in a model containing only the year as the explanatory variable, then including it in the multivariable model described above.
Analyses were conducted in STATA v7 SE. All results except the unadjusted prevalences were obtained from logistic regressions with robust standard errors at the area (primary sampling unit) level.
|
Results |
|---|
|
TOPABSTRACTIntroductionMethodsResultsDiscussionReferences |
|---|
Of the total 114,850 white adult (18+ years) respondents, 92.0% defined themselves as current drinkers, 4.1% as ex-drinkers, and 3.8% as never-drinkers. Age and gender were strongly associated with drinking status (Table 1). Nearly three times as many women reported being never-drinkers than men (5.5 vs 1.9%). For women, the proportion of never-drinkers increased monotonically with age, from 2.4% among those aged 18–29 years, to 12.4% in those aged 65+ years. For men, those aged 65+ years were also the most likely to be never-drinkers, at 3.5%, and the least likely were those aged 30–54 years, at 1.1%. The remainder of the results are restricted to respondents aged under 55 years of age.
|
Other characteristics of never-drinkers
Table 2 shows, by gender and age group, percentages of never-drinkers by categories of the characteristics examined, mutually adjusted (and controlling for region, tenure, car access, and smoking). All tested variables were strongly associated with reported drinking status, though not always for both genders or both age groups.
|
Reported never-drinking was associated with: living with another adult, particularly for women aged 30–55 years, and men aged 18–29 years; lower qualifications, showing stronger gradients for men than women, and for younger than older respondents.
There were no interactions between living alone or qualification and either age or gender.
Changes in characteristics of never-drinkers over time
Some of the characteristics of never-drinkers changed over time (Table 3). For men aged 30–54 years, in 1994–98 the proportion of never-drinkers was the same whether or not the respondent lived with another adult, but in the period 1999–2003, those living alone had less than half the probability of being a never-drinker than other men (P = 0.015). The association between qualification and never-drinking did not change significantly over time among men (P > 0.10).
|
Among young women (18–29 years), the association with educational qualification became more marked over time. In 1994–98, the percentage of never-drinkers in the lowest-qualified group was 2.1 times that in the most highly qualified, while in 1999–2003 this ratio increased to 4.7 times, due to changes at both ends of the qualification spectrum. The association between living alone and never-drinking did not change significantly over time among women (P > 0.10).
Trends over time
The prevalence of never-drinking rose over time in the 30–54 years age group, with the odds of reporting never-drinking increasing by 3% each year overall, and by 5% for men (Table 4). These rises were not explained by any of the potential confounders, with observed trends increasing slightly in a fully adjusted model. Increases in never-drinking over time in the younger age-group were smaller and did not reach statistical significance.
|
|
Discussion |
|---|
|
TOPABSTRACTIntroductionMethodsResultsDiscussionReferences |
|---|
About 4% of white English adults defined themselves as never-drinkers, with the proportion varying considerably by age and gender. Among female respondents, the proportion of never-drinkers was lowest in the youngest age group, and highest in those aged 65+ years. Men were least likely to classify themselves as never-drinkers in the 30–54 years age-group, and were less likely to be never-drinkers than women in every age-group.
While teetotallers can take up drinking, individuals who drink cannot later become never-drinkers, only ex-drinkers. We might therefore expect fewer, not more, never-drinkers at older ages. However, the result partly reflects a cohort effect, particularly among women, with a steady rise in drinking over the last 60 years taking place disproportionately among the young (Academy of Medical Sciences, 2004; Hastings et al., 2005; Foster and Marriott, 2006; Plant and Plant, 2006). Selective mortality may also play a role in these findings.
There was evidence that the proportion of never-drinkers was rising over the period (1994–2003), with the odds of never-drinking increasing by 3% per year among those aged 30–54 years. This appeared unrelated to changes in smoking, socioeconomic measures such as qualification, tenure or car access, or any other tested variable. A smaller increase was seen among the younger age group. A rise in abstention may seem surprising when, nationally, alcohol consumption is known to be increasing (Academy of Medical Sciences, 2004; Plant and Plant, 2006). A widening gap between drinking behaviours is suggested. Among the white HSfE respondents who identified as drinkers (including occasional drinkers), average weekly alcohol consumption increased between 1994 and 2002 by 6.7 units among men aged 18–29 years, and by 4.4 and 1.5 units among women aged 18–29 and 30–54 years respectively (results not shown). There was a non-significant decrease in consumption among men aged 30–54 years. We must consider too that, unlike current trends in drinking, a rise observed now in lifetime abstainers reflects not just current behaviour of individuals but their behaviour over the preceding years (decades, even) as well.
Never-drinking was most prevalent among those living with another adult and, even after controlling for other socioeconomic factors, among the lowest qualified. There was some evidence that both these associations became stronger over the period. In men aged 30–54 years, the proportion of never-drinkers dropped sharply over time among those living alone, while changing little among those living with another adult. Among young women, the proportion of never-drinkers halved among those with ‘A’ level or higher, while rising slightly among the low qualified.
It is difficult to compare absolute rates of abstinence with findings from previous studies, since most reported figures include ex-drinkers, with rates correspondingly higher (e.g. Academy of Medical Sciences, 2004; Makela et al., 2006). However, our figures are broadly in line with (current) abstinence rates in France, Germany, and some Scandinavian and Central European countries, reported by Rehm et al. (2004). Our results match, too, the finding in many countries of higher abstinence rates among women (Academy of Medical Sciences, 2004; Rehm et al., 2004; Obot and Room, 2005; Pollack et al., 2005; Rahav et al., 2006; Silveira et al., 2007). We found the rate in women (all ages combined) to be 2.9 times higher than in men. Makela reported rates around 2–3 times higher in 10 European countries, though with rate ratios around one in Finland, Iceland, and Norway. Our results agree, too, with Makela's previous UK findings, and results from Argentina, that abstinence increases sharply with age in women, and less clearly in men (Munne, 2005; Makela et al., 2006).
We should mention some caveats and short-comings. Where we refer in the text to ‘never-drinking’, this has been a short-hand for self-reported never-drinking. We do not wish to imply that the two are completely synonymous, and presume that for most readers true abstention is of greater interest. The degree of validity of self-reported never-drinking is unclear. In a small UK longitudinal study, reporting on 143 self-reported never-drinkers, 67% of these defined themselves differently in previous surveys (mostly as ‘occasional drinkers’) (Caldwell et al., 2006). The discrepancy between current and previously reported status increased with time, a factor which may contribute to increases in self-reported never-drinking with age. However, a small German study found that questionnaire responses agreed, in a repeated survey, in 91% of self-reported abstainers (Huth et al., 2007).
Another factor which may be seen as a shortcoming is the restriction of analyses to white respondents. Preliminary data tabulations showed that non-white respondents, who made up only 6% of the sample, had very different drinking patterns. For example, odds of self-reported never-drinking were 4.8 times higher among black respondents than white, and 50.2 times higher among Asians, and these differences varied according to age and gender. All results, then, would have to be stratified by ethnicity. It was felt that this would reduce overall clarity, and also that it was undesirable to present complex results for a 6% sub-sample, without sufficient ethnic minority HSfE ‘boost’ surveys available to better assess time trends. With further boost samples, comparable analyses could be done.
Lifetime abstainers form a particular group who, avoiding the considerable burdens suffered by heavy-drinkers, may also miss some health benefits, as well as the pleasure, of moderate drinking. Many studies report a J- or U-shaped association between alcohol intake and cardiovascular health problems (Academy of Medical Sciences, 2004; Doll et al., 2005; Baglietto et al., 2006; Foster and Marriott, 2006), although it has been suggested that some reported risks of abstinence have been overestimates or spurious (Emberson et al., 2005).
The Academy of Medical Sciences has called for extensive debate and research to inform policy, at a time when concerns over alcohol use, particularly among young people, are high (Academy of Medical Sciences, 2004). The Advisory Council on the Misuse of Drugs notes that drinking and drunkenness among UK children are among the highest in Europe, and that alcohol, remaining accessible and increasingly affordable to young people, causes them more problems than any other drug (Advisory Council on the Misuse of Drugs, 2006). To gauge and engage with the culture of consumption in which we live, we need to understand all parts of it. Abstainers form a special group, a small but growing minority, some of whose characteristics are changing in ways which suggest an increasing polarization within this culture.
|
ACKNOWLEDGEMENTS |
|---|
This research was funded by the Alcohol Education and Research Council. We are grateful to Ian White and Neil Poulter for their help and encouragement as advisors to the project.
|
References |
|---|
|
TOPABSTRACTIntroductionMethodsResultsDiscussionReferences |
|---|
Academy of Medical Sciences. Calling Time: The Nation's Drinking as a Major Health Issue. (2004) London: Academy of Medical Sciences.
Advisory Council on the Misuse of Drugs. Pathways to Problems: Hazardous use of Tobacco, Alcohol and other Drugs by Young People in the UK and its Implications for Policy. (2006) London: HMSO.
Baglietto L., English D. R., Hopper J. L., et al. Average volume of alcohol consumed, type of beverage, drinking pattern and the risk of death from all causes. Alcohol and Alcoholism (2006) 41(6):664–671.
Bejarano-Orozco J. Alcohol, Gender and Drinking Problems: Perspectives from Low and Middle Income Countries. (2005) Geneva: WHO. Alcohol, Gender and partner aggression: a study in the greater metropolitan area of Costa Rica.
Benegal V., Nayak M., Murthy P., et al. Alcohol, Gender and Drinking Problems: Perspectives from Low and Middle Income Countries. (2005) Geneva: WHO. Women and alcohol use in India.
Bloomfield K., Grittner U., Kramer S., et al. Social inequalities in alcohol consumption and alcohol-related problems in the study countries of the EU Concerted Action ‘Gender, Culture and Alcohol Problems: a multi-national study’. Alcohol and Alcoholism (2006) 41(S1):i26–i36.
Borsari B., Carey K. B. How the quality of peer relationships influences college alcohol use. Drug and Alcohol Review (2006) 25(4):361–370.[CrossRef][Web of Science][Medline]
Caldwell T. M., Rodgers B., Power C., et al. Drinking histories of self-identified lifetime abstainers and occasional drinkers: findings from the 1958 British Birth Cohort Study. Alcohol and Alcoholism (2006) 41(6):650–654.
Doll R., Peto R., Boreham J., et al. Mortality in relation to alcohol consumption: a prospective study among male British doctors. International Journal of Epidemiology (2005) 34:199–204.
Emberson J. R., Shaper A. G., Wannamethee S. G., et al. Alcohol intake in middle age and risk of cardiovascular disease and mortality: accounting for intake variation over time. American Journal of Epidemiology (2005) 161:856–863.
Fillmore K., Kerr W., Stockwell T., et al. Moderate alcohol use and reduced mortality risk: systematic error in prospective studies. Addiction Research and Theory (2006) 14:101–132.[CrossRef]
Foster R. K., Marriott H. E. Alcohol consumption in the new millennium—weighing up the risks and benefits for our health. Nutrition Bulletin (2006) 31(4):286–331.[CrossRef]
Goldman E., Najman J. M. Lifetime abstainers, current abstainers and imbibers: a methodological note. Addiction (1984) 79(3):309–314.
Hastings G., Anderson S., Cooke E., et al. Alcohol marketing and young people's drinking: a review of the research. Journal of Public Health Policy (2005) 26(3):296–311.[CrossRef][Web of Science][Medline]
Hettige S., Paranagama D. Alcohol, Gender and Drinking Problems: Perspectives from Low and Middle Income Countries. (2005) Geneva: WHO. Gender and alcohol in Sri Lanka.
Huth C., Siegert N., Meisinger C., et al. Individuals with very low alcohol consumption: a heterogeneous group. Journal of Studies on Alcohol (2007) 68(1):6–10.[Web of Science][Medline]
Ibanga A. J., Adetula A. V., Dagona Z. Alcohol, Gender and Drinking Problems: Perspectives from Low and Middle Income Countries. (2005) Geneva: WHO. The contexts of alcohol consumption in Nigeria.
Kerr-Correa F., Hegedus A. M., Trinca L. A., et al. Alcohol, Gender and Drinking Problems: Perspectives from Low and Middle Income Countries. (2005) Geneva: WHO. Differences in drinking patterns between men and women in Brazil.
Makela P., Gmel G., Grittner U., et al. Drinking patterns and their gender differences in Europe. Alcohol and Alcoholism (2006) 41(S1):i8–i18.
Marmot M. Inequality, deprivation and alcohol use. Addiction (1997) 92(S):13–20.
Mendoza M. R., Medina-Mora M. E., Villatoro J., et al. Alcohol, Gender and Drinking Problems: Perspectives from Low and Middle Income Countries. (2005) Geneva: WHO. Alcohol consumption in Mexican women: implications in a syncretic culture.
Munne M. I. Alcohol, Gender and Drinking Problems: Perspectives from Low and Middle Income Countries. (2005) Geneva: WHO. Social consequences of alcohol consumption in Argentina.
Obot I. S., Room R., eds. Alcohol, Gender and Drinking Problems: Perspectives from Low and Middle Income Countries. (2005) Geneva: WHO.
Plant M., Plant M. Binge Britain. (2006) Oxford University Press.
Pollack C. E., Cubbin C., Ahn D., et al. Neighbourhood deprivation and alcohol consumption: does the availability of alcohol play a role? International Journal of Epidemiology (2005) 34:772–780.
Rahav G., Wilsnack R., Bloomfield K., et al. The influence of societal level factors in men's and women's alcohol consumption and alcohol problems. Alcohol and Alcoholism (2006) 41(S1):i47–i55.
Rehm J., Room R., Monteiro M., et al. Comparative Quantification of Health Risks. (2004) Geneva: WHO.
Silveira C. M., Wang Y.-P., Andrade A. G., et al. Heavy episodic drinking in the Sao Paulo epidemiologic catchment area study in Brazil: Gender and sociodemographic correlates. Journal of Studies on Alcohol (2007) 68(1):18–27.[Web of Science][Medline]
Tumwesigye N. M., Kasirye R. Alcohol, Gender and Drinking Problems: Perspectives from Low and Middle Income Countries. (2005) Geneva: WHO. Gender and the major consequences of alcohol consumption in Uganda.
van Oers J. A. M., Bongers I. M. B., Van de Goor L. A. M., et al. Alcohol consumption, alcohol-related problems, problem drinking, and socioeconomic status. Alcohol and Alcoholism (1999) 34:78–88.
Wilsnack R. W., Vogeltanz N. D., Wilsnack S. C., et al. Gender differences in alcohol consumption and adverse drinking consequences: cross-cultural patterns. Addiction (2000) 95(2):251–265.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||