The Department of Health estimates that the harmful use of alcohol costs the National Health Service in England around £3.5 billion a year and 8% of all hospital admissions involved an alcohol-related condition. Drinking can lead to over 40 medical conditions, including cancer, stroke, hypertension, liver disease and heart disease. Reducing the harm caused by alcohol is therefore a priority for the Government and the devolved administrations. Excessive consumption of alcohol is a major preventable cause of premature mortality with alcohol-related deaths accounting for 5.3% of all deaths in England and Wales in 2005. The Opinions and Lifestyle Survey (OPN), General Lifestyle Survey (GLF) and General Household Survey (GHS) are important sources for monitoring trends in alcohol consumption.
These drinking data are widely used by universities and health organisations. The School of Health and Related Research (ScHARR) at the University of Sheffield has used the data to carry out alcohol-related public health research. Public Health England also uses the data on drinking to estimate the proportion of deaths and illnesses that can be attributed to alcohol consumption.
Information on the proportion of men drinking more than eight units of alcohol in a day and women drinking more than six units in a day is used as a proxy measure of binge drinking (known as heavy episodic drinking). In particular it is used as an outcome measure for the Government's Alcohol Strategy.
The OPN, GLF and GHS have measured drinking behaviours for over 30 years. This release presents findings from these surveys on recent trends in drinking habits in Great Britain, and compares the drinking habits of different groups of adults in 2012.
In 2012 58% of adults1 in Great Britain (GB) drank alcohol in the week before being interviewed. There was a two percentage point fall for both men and women compared with 2011. This continues the general decline in the proportion of adults drinking in GB since 2005. The decline between 2011 and 2012 was driven by the 25-44 age group. The proportion of men aged 25-44 who drank in the last week fell from 67% to 63% between 2011 and 2012. In comparison there was a fall from 56% to 50% in this proportion amongst women of the same age.
Over the eight year period between 2005 and 2012, men were consistently more likely than women to have drunk alcohol in the last week, Figure 1.
11% of adults in GB were frequent drinkers (drank alcohol on at least five days in the week before being interviewed) in 2012. Men were consistently more likely than women to be frequent drinkers.
There was a fall from 16% to 14% between 2011 and 2012 in the proportion of men who were frequent drinkers. There was no change in the proportion of women who were frequent drinkers (9%).
Those aged 16-44 were the main cause of the decline in the proportion of adults who drank in the last week, and in the proportion of adults who were frequent drinkers, between 2005 and 2012.
People were asked how much alcohol they drank on their heaviest drinking day in the last week (HDD). There was a decline between 2005 and 2012 in the proportion who exceeded 4/3 units on their HDD (that is, four units for men and three units for women), Figure 2. There was also a decline between 2005 and 2012 in the proportion who exceeded 8/6 units on their HDD (eight units is the the same as four pints of normal strength lager, and six units the same as two large glasses of wine). These falls were present in all age groups except those aged 65 and over.
In 2012 58% of adults1 in Great Britain (GB) drank alcohol in the last week (the week before interview). tTis proportion varied by different groups. Men (64%) were more likely than women (52%) to have drunk in the last week. This is true for all age groups apart from those aged 16-24. Men and women aged 16-24 were almost equally likely to have drunk in the last week, Figure 3.
The proportion who drank in the last week also varied by age. Those aged 16-24 were least likely to have drunk in the last week (49%), and those aged 45-64 (65%) were most likely. However, when 16 and 17 year olds are excluded, there is no difference between the proportions of 18-24 year olds and those aged 65 and over who drank in the last week. The proportion of 16 and 17 year olds who said that they drank in the last week was low. This is most likely because in GB the legal age for alcohol purchases is 18.
A higher proportion of working adults (65%) drank in the last week than those who were not working (48%). This is in contrast to findings from the 'Opinions and Lifestyle Survey - Smoking Habits Amongst Adults, 2012' publication, which found that the level of smoking prevalence in 2012 was highest amongst those who were unemployed (not working but looking for work).
There was variation in the proportion of adults who drank in the last week between most ethnic groups, Figure 5. Six out of 10 White adults drank in the last week, compared to aproximately two out of 10 Asian and Asian British adults.
The proportion of people who don't drink at all (abstinence) is known to vary by ethnic group. Abstinence is highest amongst South Asians, particularly those from Pakistani, Bangladeshi and Muslim backgrounds. Abstinence is often influenced by religion which is closely associated with ethnicity.
Adults living in London (51%) and the West Midlands (52%) were least likely to have drunk in the last week when compared with other English regions. However in these regions a higher proportion of people are from non-white ethnic groups. Our analysis indicates that once other factors such as ethnicity have been accounted for, adults in London were no less likely than those from other English regions to have drunk in the last week.
Adults who drank on five or more days in the week before interview have been classed as frequent drinkers. In 2012 11% of all adults were frequent drinkers. The likelihood of being a frequent drinker increased with age, Figure 6. 3% of 16-24 year olds were frequent drinkers, increasing to 18% of those aged 65 and over. Men (14%) were more likely to be frequent drinkers than women (9%), and this gap increased with age.
Single adults were less likely than other adults to be frequent drinkers. However our analysis found that single people were more likely to be younger, and that this reduced the likelihood of a single adult being a frequent drinker. When other factors, such as age, were accounted for, single adults were no less likely than others to be frequent drinkers.
White adults (12%) were four times as likely as those from non-white ethnic groups (3%) to be frequent drinkers. This is consistent with drinking in the last week.
Respondents were asked about their heaviest drinking day in the last week (HDD). They were asked about the types and measures of alcoholic drinks consumed on that day. These drinks have been converted into units of alcohol based on assumed alcohol levels for each measure/type of drink (more information on this can be found in background note 3).
For the purposes of this release, those who drank in the week before interview and who drank more than 8/6 units (that is, eight units for men and six units for women) on their HDD have been classed as heavy drinkers. Eight units is roughly equivalent to four pints of normal strength lager, and six units roughly equivalent to two large glasses of wine. Those who drank more than 12/9 units on their HDD have been classed as very heavy drinkers. 12 units of alcohol is roughly equivalent to six pints of normal strength lager, and nine units roughly equivalent to three large glasses of wine.
In 2012 a quarter of adult drinkers were heavy drinkers. Male drinkers (29%) were more likely to be heavy drinkers than female drinkers (21%), regardless of age. The likelihood of being a heavy drinker decreased with age, Figure 7.
Younger adults were less likely to have drunk frequently, but younger drinkers were more likely to have drunk heavily. The opposite was true for older adults, who were more likely to have drunk frequently, but older drinkers were less likely to have drunk heavily.
Economically inactive drinkers (such as students and retired people) were half as likely to be heavy drinkers (15%) when compared with drinkers who were in or looking for work (30%). Further analysis found that economically inactive people tended to be older. When other factors such as age are accounted for, drinkers who were economically inactive were no less likely than others to be heavy drinkers.
Single drinkers (39%) and cohabiting drinkers (36%) were more likely to be heavy drinkers than married drinkers (19%) or widowed/divorced/separated drinkers (20%).
White drinkers were almost three times as likely to have been very heavy drinkers (14%) as non-white drinkers (5%).
In 2012, 64% of female drinkers drank wine on their heaviest drinking day in the last week (HDD), making it the most popular type of drink amongst women. The exception was the 16-24 age group, where female drinkers showed a preference for spirits/liqueurs over wine. 59% of male drinkers drank normal strength beer/lager/cider/shandy on their HDD. Normal strength beer/lager/cider/shandy was the most popular type of drink amongst all ages of male drinker. Amongst older male drinkers the popularity of normal strength beer/lager/cider/shandy decreased and the popularity of wine increased.
A small proportion of drinkers drank sherry/martini or alcopops on their HDD. 9% of female drinkers aged 65 and over drank sherry/martini on their HDD; a higher proportion than any other age/gender group. In contrast drinkers aged 16-24 (8%) were more likely than those in other age groups to have drank alcopops.
Very heavy drinkers were five times more likely than other drinkers to have drunk strong beer/stout/lager/cider, and over twice as likely to have drunk spirits or liqueurs, Figure 8. The difference in the proportion who drank strong beer/stout/lager/cider was even greater for male drinkers, where the proportion was nearly six times higher amongst men who drank very heavily (28%) than it was amongst other male drinkers (5%).
In 2012 four out of five adults rated their general health positively (as either good or very good). However there was a relationship between self-reported health status and drinking habits. 83% of adults who said that they drink alcohol nowadays rated their general health positively compared with 68% of adults who said that they do not. A contributing factor may be that people who rated their health as bad or very bad may be unable to drink alcohol because of health issues.
Those who didn't drink at all and those who drank every day in the last week were least likely to rate their health positively, Figure 9.
Adults' perceptions of heavy drinking levels are higher than the official measures. Adults who drink more than 8/6 units on their HDD are considered binge (or heavy episodic) drinkers in the Government's Alcohol Strategy. Eight units is the same as four pints of normal strength lager, and six units is the same as two large glasses of wine. Whilst 14% of all adults in 2012 were classified as binge drinkers using this measure, less than 1% felt that they were heavy drinkers.
Adults' self reported health was also related to their self reported level of drinking. Adults who felt that they drink heavily were less likely than other adult drinkers to rate their general health positively, Figure 10.
Only 56% of adults who felt that they drink heavily rated their general health positively, compared with 82% of other adults who said that they do drink, Figure 10.
Smokers have been grouped into light smokers (smoke on average up to 10 cigarettes per day), moderate smokers (10 up to 20 cigarettes per day) and heavy smokers (20 or more cigarettes per day).
Drinkers who smoked (25%) were more than twice as likely as those who did not smoke (11%) to be very heavy drinkers, regardless of how much they smoked, Figure 11. However drinkers who were ex-smokers (13%) were more likely to be very heavy drinkers than those who had never smoked (9%). It is not clear whether people who have given up smoking reduce their drinking frequency, or whether smokers who reduce their drinking frequency then go on to give up smoking.
Heavy smokers (21%) were most likely to have been frequent drinkers. They were almost twice as likely as light smokers (11%) or moderate smokers (11%) to be frequent drinkers, Figure 12. Ex-smokers (16%) were also more likely than light or moderate smokers to have been frequent drinkers, with those who have never smoked least likely (7%).
Pregnant women aged 16-49 were far less likely than women who were not pregnant to have drunk in the last week. One in 10 pregnant women drank in the last week compared with five in 10 of all other women.
None of the pregnant women interviewed were frequent drinkers, compared with 5% of all other women aged 16-49.
Measuring Alcohol Consumption
Obtaining reliable information about drinking behaviour is difficult, and social surveys consistently record lower levels of consumption than would be expected from data on alcohol sales. This is partly because people may consciously or unconsciously underestimate how much alcohol they drink. Drinking at home is particularly likely to be underestimated because the quantities consumed are less likely to have been measured and also more likely to be larger than those dispensed in licensed premises.
There are different methods for obtaining survey information on drinking behaviour. One approach is to ask people to recall all episodes of drinking during a set period (See Goddard E (2001) 'Obtaining information about drinking through surveys of the general population’ (275.9 Kb Pdf) , National Statistics Methodology Series NSM 24). However, this is time-consuming and is not suitable for the Opinions and Lifestyle Survey (OPN), where drinking is only one of a number of subjects covered.
Since 1998 respondents have been asked questions about their drinking behaviour in the seven days before interview. Specifically, people responding to the OPN are asked on how many days they drank alcohol during the previous week. They are then asked how much of each of six different types of drink (normal strength beer; strong beer; wine; spirits; fortified wines; and alcopops) they drank on their heaviest drinking day during the previous week. These amounts are converted to units of alcohol and summed to give an estimate of the number of units the respondent consumed on their heaviest drinking day (see background note 3 for further details of the number of units assumed to be in each measure of each type of drink).
Change in survey vehicle
In 2012, the survey vehicle for collecting drinking data changed from the General Lifestyle Survey (GLF) to the Opinions and Lifestyle Survey (OPN).
The OPN uses the same initial approach to sampling as the GLF. That is, a random sample of addresses is drawn from the Postcode Address File (PAF). Initially, a sample of postcode sectors is drawn, and from within those, a list of addresses is chosen. The design means that every address and every person in Great Britain has an equal chance of selection. The PAF is ordered by region and other socio-demographic indicators provided by the census. Ordering the PAF helps to ensure the sample represents the general population of Great Britain. GLF analysis was based on four waves of longitudinal data, weighted to produce cross-sectional estimates. The OPN only produces cross-sectional estimates.
Both the OPN and GLF surveys use Computer Aided Personal Interviewing (CAPI). Interviewers visit sampled addresses and interview respondents with the aid of a laptop. Face to face interviewing helps to maximise participation in the survey, which is around 60% of sampled addresses for both surveys.
All of the trends reported on in this release are based on the same questions previously included on GLF. Therefore, no trend changes will have been introduced as a result of changing question wording.
After data collection, the results are 'weighted' to ensure they represent the general population of Great Britain. The approach to weighting is the same for both surveys. Firstly, an adjustment is made to account for those respondents who chose not to take part in the surveys (or, for OPN, the chances of selection within a household - see 'What are the differences?', below). Secondly, the samples are grossed up to total the GB population, taking account of the age and gender profile of the population, as well as the distribution of people across Local Authorities.
What are the differences?
The primary difference between these two surveys is the approach to interviewing residents of sampled addresses. The GLF aimed to interview all residents in the household aged 16 or above. The OPN only interviews one person aged 16 or above in each sampled household. The OPN randomly selects a resident for interview, thus reducing the potential for selection bias. Despite a different approach to within household sampling selection, the achieved number of interviews is broadly the comparable between the two surveys. The GLF achieved approximately 13,500 adult interviews per year compared to approximately 13,000 adult interviews per year on the OPN.
There was a reduction in the number and detail of questions migrated from the GLF to the OPN. Details of the questions that were moved from the GLF to the OPN are provided in Appendix A of the Future of the GLF (108.9 Kb Pdf) update.
As part of the OPN, respondents have also been given the option to record normal strength beer/lager/cider/shandy and strong beer/stout/lager/cider in measures of schooners. A schooner is a glass that contains two-thirds of a pint. When estimating the number of units of alcohol that a person drank, the units for schooners have been pro-rated to account for this difference in size.
What do these differences mean for comparability?
Every effort has been made to minimise the potential for discontinuity between the results from the two surveys. Our analysis shows that the results of the OPN are comparable with those of the GLF.
The reduction in achieved sample size has not led to significant loss in precision. The confidence intervals around key estimates such as the proportion of adults in Great Britain who drank alcohol in the last week are comparable between the GLF and OPN. Results have been found to follow the same patterns by various socio-demographic characteristics such as age, gender and employment status. The comparability of these estimates is deemed sufficient to make comparisons between 2012 and previously reported results.
However, the reduction in questions asked on topics covered by the GLF has led to a discontinuation of some outputs. For example, in the case of smoking, the age at which people started smoking is no longer reported on, as the source data is not collected on the OPN. Similarly, questions on tobacco dependency were no longer asked from January 2012.
Recent changes in methodology
The conversion of volumes of alcoholic drinks to units of alcohol is based on assumptions about the size of a given measure (for example, a glass of wine) and the alcohol content of the type of drink, that is, the percentage of alcohol by volume (ABV). In recent years there have been changes to both of these factors and these have been reflected in revisions to the conversion method. The survey does not ask about the specific ABV of every alcoholic drink consumed but assumes an average for each type of drink. The revised method changed the number of units assumed to be in drinks in the ‘normal strength beer, lager and cider’ and ‘strong beer, lager and cider’ categories but the main impact was on drinks in the ‘wine’ category. Background note 3 gives details of the number of units assumed to be in each measure of each alcoholic drink type.
The revised method had a large impact on the estimates of units of alcohol consumed from wine because it changed both the assumed ABV of wine (from 9 to 12%) and the size of a glass of wine. Until 2006 a glass of wine was assumed to be 125 ml. Respondents are now asked whether they have consumed small (125 ml), standard (175 ml) or large (250 ml) glasses of wine. It is now assumed that a small glass contains 1.5 units, a standard glass contains 2 units and a large glass contains 3 units. Discussion of the impact of these changes on the estimates of consumption can be found in the report on the 2009 data.
Assumed alcohol by volume (ABV) percentages for different types and measures of alcoholic drink
Table 1, below, shows the assumed number of units of alcohol present in each measure of each drink type.
|Normal strength beer/lager/cider/shandy||Half Pint||1.0|
|Normal strength beer/lager/cider/shandy||Pint||2.0|
|Normal strength beer/lager/cider/shandy||Small Can||1.5|
|Normal strength beer/lager/cider/shandy||Large Can||2.0|
|Normal strength beer/lager/cider/shandy||Bottles||1.5|
|Normal strength beer/lager/cider/shandy||Schooners 1||1.3|
|Strong beer/stout/lager/cider||Half Pint||2.0|
|Strong beer/stout/lager/cider||Small Can||2.0|
|Strong beer/stout/lager/cider||Large Can||3.0|
|Strong beer/stout/lager/cider||Schooners 1||2.7|
|Spirits/liqueurs||Standard single (25ml)||1.0|
|Spirits/liqueurs||Standard double (50ml)||2.0|
|Sherry/martini, including port, vermouth, Cinzano and Dubonnet||Glass||1.0|
|Wine, including champagne or babycham||Small Glass (125ml)||1.5|
|Wine, including champagne or babycham||Medium Glass (175ml)||2.0|
|Wine, including champagne or babycham||Large Glass (250ml)||3.0|
|Wine, including champagne or babycham||Bottle||9.0|
|Wine, including champagne or babycham||Unknown||2.0|
Details of the policy governing the release of new data are available by visiting www.statisticsauthority.gov.uk/assessment/code-of-practice/index.html or from the Media Relations Office email: email@example.com
|Tom Howe||+44 (0)1633 456330||Office for National Statisticsfirstname.lastname@example.org|