Introduction

Trends towards increased consumption of alcohol even at levels well short of dependence raise concerns for public health, particularly among those in later life. Alcohol consumption and alcohol-related deaths or problems have recently increased among older age groups in many developed countries, including the USA, Australia and many countries of the European Union [1]–[5]. The US National Institute on Alcohol Abuse and Alcoholism recommends that healthy adults aged over 65 years who do not take medications should not drink more than 3 drinks (42 g alcohol) per day or 7 drinks (92 g alcohol) per week [6]. Yet 10.8% of men and 2.9% of women in a large US sample aged over 65 reported exceeding these limits, compared with 28.6% of men and 10.3% of women in a UK sample the same age [7] and just over a fifth of samples of comparable age in Finland and Belgium [8]. In the USA, 23% of men and 9% of women aged 50–64, and 14% of men and 3% of women aged 65 and above, report binge drinking (consuming five or more drinks) on at least one occasion in the preceding month [9], while 27% of Europeans aged over 54 years report doing so on a weekly basis [10]. Frequent alcohol consumption can also have a negative impact on health, and is positively associated with age. In England, 28% of men over 65 years and 14% of women over 65 drink alcohol more than 5 times per week, while in Australia 15% of those aged over 65 report daily consumption compared with 4% of those aged under 50 [3], [11]. Other work has however found that frequency of consumption may decrease across the older age group [12], which suggests there is still much important research to be conducted in order to gain a detailed picture of alcohol consumption in older people.

Risk of mortality over 20 years is increased among individuals aged 55–65 years consuming a daily average of 3 standard drinks (42 g alcohol) [13]; in a sample of male general practitioners aged 48–74 years this risk was increased at an average daily consumption of 2 drinks or more [14]. Decreasing consumption below this level may not continue to reduce risk of mortality, however. Older adults whose average daily consumption fell between 1 and 2 standard drinks per day showed no increase in combined risk of disability and mortality compared with those consuming an average of 0 to 1 standard drinks per day [7]. Older people are more likely than younger adults to suffer from chronic health problems, bereavement or isolation and to use medications to a greater extent [15], [16]. These medications may interact with any alcohol consumed, for instance increasing their likelihood of falls or problems with the medication [12]. Encroaching health or social problems can lead to increased alcohol consumption either to self-medicate or to cope. Consuming more than 14 units of alcohol per week in this age group has been found to be positively associated with symptoms of depression and anxiety and less social support, though the direction of causality in this relationship is unclear [17]. For these reasons, regular and heavy drinking (in excess of recommended guidelines) can have a greater impact on physical and mental health in older individuals than in younger ones, and can have an effect at lower levels of consumption than would be the case in middle age or early adulthood [18]–[20].

Heavy drinking is more widespread than physiological or psychological dependence on alcohol, and is often targeted by national policy, for instance through taxation, public health messages about its impacts and support for preventive non-specialist approaches such as brief interventions [21]. Given these efforts, it is important to understand why older age groups are drinking more heavily than previously. Interventions to reduce alcohol consumption aim to encourage reasoned decisions about health [22]. However, responses to public health messages may be weak where the risk to health is from a consumable substance rather than an organism (such as influenza), where that substance can be perceived as desirable, and where health risks are deferred [23]; all qualities characteristic of alcohol consumption. In such circumstances, individuals weighing up the immediate and long-term consequences of a risky behaviour may, through engagement with cultural channels or humour, favour contradictory theories or findings around health risks, and also speculate on the motives or partiality of sources of advice [24]. This can build a rationale privileging the perceived benefits of heavy drinking of alcohol over generalised understanding of its risks to health [25]. Yet this may not fully explain increasing consumption in older people; the same considerations should apply to younger adults whose average drinking levels have reduced over recent years but who are more likely to defer health considerations to a later time of life [26], [27].

Rather than viewing health risks from alcohol as being outweighed by its desirability, older people may reason around alcohol primarily in social or moral rather than health terms [24]. Heavy drinking in this age group is subject to particular stigma [18], [28]–[30]. Given the gendered patterns of alcohol consumption [31], this may be especially so for women's drinking; women in mid to later life, for instance, may perceive particular pressure to present their drinking as appropriate and moderate in a way that men of the same age do not [32], [33]. Although stigma may increase the isolation older people already suffer disproportionately, and hamper the identification and prevention of risks to their health from alcohol, there has been little elaboration of how this is maintained and exerts influence. Identity theory suggests that over successive transitions in life, older people will have rehearsed and accumulated preferred versions of self within which to frame themselves and their experience [34], [35]. These identities will be multiple, mutable and responsive to personal, temporal, social and cultural contexts [36]. An older person may for instance move between representing themselves as a matriarch or patriarch, a patient, someone winding down their working life, or someone active or ‘young for their age’. Such identities are likely to ascribe particular meanings for alcohol use, behaviour which has considerable and diverse cultural connotations. In public health terms, heavy drinking and dependence on alcohol represent differing degrees of problems and health impacts. However, being a heavy drinker may underpin important community or family identities that are challenged by significant health events or bereavement in later life. Identity as someone dependent on alcohol may carry particularly adverse moral associations for older people. Relinquishing a version of self that is cherished as normal in favour of new, healthier identities implies a greater challenge in later life than simply recognising that health is at risk.

These social theories have distinct implications for understanding why older people may be drinking more. We report a qualitative study in the UK, where substantial numbers of individuals over the age of 50 exceed governmental recommendations that adults (women/men) should not drink more than 3/4 units of 8 g of alcohol per day or 14/21 units per week [11]. This research aimed to elucidate the views of older individuals aged over 50 years about alcohol consumption, health and well-being, to inform future targeted prevention in this group. The findings provide insight into older people's constructions of drinking behaviour in both health and social terms, and highlight that stigma around problem drinking in this age group may lead to moral reasoning being prioritised over public health rationales.

Methods

This study was given favourable ethical approval by Newcastle University's Faculty of Medical Sciences Ethics Committee (application no. 000224/2009). In line with the terms of consent to which participants agreed, the data are not publicly available.

In-depth interviews (n = 24, 12 male, 12 female) were conducted between 19/11/09 and 15/03/10 by GW, a researcher with extensive experience of qualitative research on interventions for families or substance use. Purposive sampling aimed to recruit both genders and represent a broad range of age and self-reported drinking, past or current. Participants aged 50 and above were sought in order to consider views both leading up to and following the transition of retirement, and to consider how problems after the age of 65 may arise in late working life. Three local branches of a national charity aiming to improve later life (Age UK) and two regional services for alcohol problems were given research information leaflets. These explained the reasons and objectives for the research; what was involved in participation; how information would be handled and used; and who to contact for further information. The leaflets were distributed to service clients or volunteers who had experience of drinking alcohol at any level of consumption. Staff members invited the clients to consider participating in an interview, answered any questions they had about the research and asked those who were interested to complete and sign a consent form. GW contacted all potential participants by telephone to arrange a single interview; both on this call and prior to commencing interviews, the researcher checked that the individual was familiar with the research leaflet, answered any questions they had, reminded them of the voluntary and confidential nature of participation and checked that they were still happy to proceed. Initially 17 participants were recruited; although heavy drinkers were not particularly targeted in recruitment, most reported experience of dependence on alcohol. One further form was received but this participant withdrew consent without supplying a reason when contacted by GW. We sought the perspectives of older people with a range of patterns of consumption, in order to avoid over-representation of very heavy drinkers. For this reason a second wave of recruitment targeted only Age UK clients who drank but considered themselves to have had no problems with alcohol. Strategic ‘snowballing’, whereby interviewees suggested other individuals who could be invited to participate, added two further interviewees who were not using services. This increased the sample to 24, at which point the research team agreed that data saturation had been reached. Fifteen of these were recruited via Age UK, six of whom were also involved with voluntary services for mental health support. Considerable diversity was achieved (see Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Interviewee characteristics.

https://doi.org/10.1371/journal.pone.0071792.t001

Individual claims regarding stigmatised aspects of identity are likely to be qualified in the presence of peers, and recognising this influence is important to understanding how social contexts might also affect the implementation of preventive measures to address that behaviour. Focus groups with Age UK clients or volunteers were organised towards the end of data collection to allow comparison of individual accounts with socially negotiated versions of drinking in later life. Age UK staff, who also attended the groups, were asked to invite any men or women whose age made them eligible for Age UK services and who drank alcohol. One focus group was organised at each of three local branches; all those who expressed an interest attended one of these groups. The first group comprised 9 participants (1 m, 8f, ages 79–95); the second group comprised 12 participants (5 m, 7f, ages 50–85) and the third comprised 6 participants (all female, ages 51–76). To encourage participation, focus group participants were not required to disclose personal details other than age, date of birth and address; these data were gathered on consent forms. The research team prepared topic guides (available on request) to initiate discussion or return talk to the research topics. At the groups, participants were invited by the facilitator (GW) to offer views in general rather than recounting personal experience in front of others. Interviews and focus groups lasted between 40 and 150 minutes either at individual respondent's homes or the offices of Age UK or other services, as well as one interview at Newcastle University.

Interview and focus group data were audio recorded with consent, transcribed verbatim, anonymised and loaded into NVivo software. Analysis was informed by grounded theory and discursive psychology approaches. GW repeatedly read transcripts and coded for emergent themes; early analysis informed later interviews and focus groups [37]. Passages where participants accounted for drinking were examined for discourse features such as repertoires (recurring patterns of characterisation or metaphor) or contested identity claims (where descriptions of self or another are treated as questionable) [38]. Coding was refined through discussion of emerging themes amongst all researchers. Focus group data were used to triangulate findings from individual interviews; in each section below we report on findings from individual interviews, and then consider how the focus group data inform these themes. To reach a fully consistent interpretation of the data, instances diverging from the themes were considered in terms of how they might inform the interpretation. Lay reasoning, identities and stigma were considered and brought to bear during inductive analysis as theoretical constructs useful in interpreting emerging themes. We focus here on identity formation around alcohol, then on accounts of alcohol in relation to health, and finally on any influence of gender.

Results

Discussion

The aim of this research was to contribute to contemporary understandings of alcohol use in later life and its implications for the health of older people. UK adults in later life who were asked to talk about alcohol and health oriented strongly towards the opposed identities of normal or problematic drinker, the latter being highly stigmatised. These identities were defined primarily by self-control and propriety, rather than health considerations. In data from focus groups, participants could be seen to negotiate consensual views on alcohol and its implications for both health and propriety, providing support for their views being socially constructed in response to a particular context. These older people portrayed drinking less alcohol as an appropriate response if one experienced impaired health. However, continued heavy drinking could be presented as normal behaviour for someone experiencing relative wellbeing in later life, or if their experience of ill health was construed as unrelated to alcohol consumption. This normalisation was supported by treating general health advice on alcohol as questionable or individually irrelevant. Drinking patterns did not appear to be strongly defined by gender, although some gendered expectations of drinking were evident.

Absolute numbers of risky drinkers in older age groups are small compared with younger adults but will grow in an ageing population [15], [39]. It is predicted that 4.4 million US adults aged 50 or above will need treatment for substance use in 2020 compared with 1.7 million in 2000 [39]. More stringent guidelines on alcohol consumption for those aged 65 and over are often recommended by governments but not always formally instituted [15], [40]. There is a need to develop preventive approaches that specifically target older populations, for instance providing longer term in-home support, tailored information on risks to health from alcohol use in later life, or health workers with specific training on older people's needs in relation to alcohol [41]. Non-specialist preventive approaches to avoid the harmful consequences of continued heavy drinking, such as brief interventions, are effective with older people [42], [43]. Nevertheless, preventive measures rely on reasoning in terms of health risks and benefits, and so it is notable that health risks from drinking, as opposed to health impacts, did not appear meaningful to our participants.

The scepticism displayed by participants towards health advice regarding alcohol consumption is consistent with resistance to health promotion reported for other lifestyle issues, such as smoking [44], [45]. Advice on risks from alcohol must compete for older people's attention with many other warnings and may be less meaningful because of a general perception of ‘overkill’ [46]. Focus groups facilitate qualification of one participant's views by others and so represent a particularly fertile environment for such reasoning to emerge, for instance in making reference to famous instances of untrustworthy advice [47]. Crossley [48] found a similar recollection of unhelpful UK government advice about salmonella risk from eggs to that raised by one of our focus groups. However, individual reasoning may not be consistent with socially negotiated points of view. Participants' questioning of advice on alcohol when they had not experienced its adverse impact on their health was characteristic of lay understandings, which are crucially informed by lived experience or personal observation and the desirability of the health behaviour [24], [49]. Their reasoning appeared to conform to public health rationales where health risks were no longer perceived as deferred [cf 23]; thus behaviour change was warranted only within a narrative of damaged health, or once harmful or dependent drinking identity was recognised. Yet participants described experiencing adverse health and increased need for treatment and care in later life; in this respect older people might be expected to be more likely than younger populations to consider issues in health terms.

Increasing alcohol consumption among older age groups might instead reflect identity work in this population. Older adults can be relatively ill-informed about alcohol measurements on which guidelines are based [50], [51]. Participants reasoned that recommended guidelines are inconsistent with individual capacities, but also showed concern that being considered an unhealthy drinker, even at low levels of drinking, might signify alcoholism. Given that constellations of positive and negative aspects were not consistently applied by individuals to discrete levels of moderate and heavy drinking, they can be viewed as repertoires [52]: self-contained sets of characteristics used to associate an individual's drinking with a particular identity, whatever they actually consume. They characterise positive or negative identities rather than objectively distinct behaviours. The positive repertoire invites the idea of a normal and therefore acceptable drinker; the negative repertoire identifies an individual's drinking as a source of stigma. Interviewees saying they drank showed strong sensitivity to being perceived as the ‘wrong’ sort of drinker (irresponsible, improper, or wilfully abusing alcohol) and those identifying themselves as previously dependent on alcohol treated this as a negative aspect of their former selves.

Drinking behaviour did not appear to be strongly defined by gender in our study in contrast to other qualitative work. This difference may reflect the fact that gender-based issues were not the main focus of this study. Some gendered expectations of drinking were described, often with a focus on culturally stereotyped patterns of consumption. In line with findings from somewhat younger UK men [53], both men and women treated regular drinking of ‘pints’ or socialising in ‘pubs’ together as normal for men but not women. In the present study, however, gendered drinking patterns tended to be presented as more widely adhered to in earlier life. Furthermore, individuals who had experienced alcohol dependence described overruling those gendered patterns that were identified. In addition, it was noted that drinking could become less gender segregated in later life, as some couples spent more time together. Further qualitative research focused on later life drinking in males and females is needed to unravel these complex social and health issues.

Longitudinal studies of large samples of older adults in both the US and Norway have shown reduced drinking linked to the incidence of chronic or painful medical conditions [5], [54] or acute health events [20]. Such changes in wellbeing may provide a narrative framework to account for reducing one's drinking and may constitute opportunities for engaging the interests and attentions of those who drink to excess [55]. Preventive approaches should aim to target unhealthy behaviour before it impairs health; yet older adults who feel well appear to question longer term health risks from heavy drinking, or feel that the risks are outweighed by benefits of sociability and relaxation. The talk of older adults indicates that unhealthy drinking may not imply a spoiled identity [28], and therefore represent a cause for concern, unless it can be read by others as a loss of self-control. However, acknowledging that one's health might be affected by alcohol implies a loss of control over physical integrity and therefore risks incurring undesirable identity as an ‘alcoholic’. In this way, discourse practices reinforce or maintain stigma but can also have an important role in re-constructing self-identities [55]. To counter heavy drinking in older populations before it has significant impacts on health, interventions must engage and foster positive identities for older adults that can support change. Social marketing approaches would be appropriate for this and have been used to promote healthier lifestyles in this age group [56]. In particular interventions should aim to reframe the widely recognised but poorly understood concept of alcoholism. For instance, the information component of brief advice could be tailored to clarify the mechanism of cumulative impacts on health through drinking at levels short of dependence over a number of years, and how and why such impacts are strengthened in later life.

Although the overuse of alcohol in later life and its associated stigmas are clearly concerns for public health [15], [18], [30], [57] these issues have not been adequately explored through the framework of social theories to date. Qualitative methods have allowed this study to access social constructions grounded in older peoples' own frames of reference. It is a limitation that recruitment was mostly through one agency, and a larger study might have adopted a broader strategy. Nevertheless, the sample was diverse and from a UK region with an older population than average and a culture of heavy drinking [58], [59]. Stigma around drinking in later life has been observed in other developed countries besides the UK, and discourse observed here around propriety and health is likely to be analogous to that in other locations and populations. For instance, working adults in mid-life in Glasgow and the North East also identify a tradition of nights out drinking with colleagues [32], [60], though those studies found somewhat more readiness, particularly among men, to endorse inebriation as a goal than among the older adults in the present study. Although larger than many estimates of numbers needed to reach data saturation [61], the size of sample in this study is appropriate given the variation encompassed. Participants reported varied consumption and ages ranged from 50–90. As such, findings extend those based on treatment populations only [62]. The interview sample included a greater proportion of individuals identifying themselves as currently or formerly dependent on alcohol than might be expected in the overall population of older people. However individual views were triangulated with those from the focus groups, where a normative identity as a sensible drinker was adhered to by all participants. The ratio of women to men within the groups (21∶6) was considerably greater than in the UK population aged 65 and over (around 5∶4) [63]. This may reflect a greater willingness among older women compared with older men to participate in group discussions. However, the unit of analysis for focus group data should be considered the group given that not all participants may contribute or contribute equally [64], and two of the three groups were mixed.

Consideration of our data in light of theories of lay epidemiology and identities has informed an under-recognised area of public health concern. Strengthening the evidence for specific risks to health from alcohol and clarifying how these interact with behaviour might increase the impact of preventive strategies to reduce heavy drinking in later life. However, the processes by which identities can be constructed are ‘tremendously powerful’ in governing social interaction ([38], p5), and the concept of identities can also improve understanding of public health issues around older people's alcohol-related behaviour. Although older people construct positive identities for themselves by discounting stigma [65], emphasising propriety to avoid stigma around alcohol could lead them to treat unhealthy behaviour as normal, such as the drinking of eight pints of beer endorsed in Excerpt 6 on the grounds of apparent self-control. Long-term comparative data indicate that consuming alcohol at high risk in later life tends to reinforce some social resources at the expense of others [66]; older men's heavy drinking is positively associated with friends' approval for drinking, but is likely to be reduced over time if they have a partner who drinks less heavily [67]. Access to a strong identity as a moderate drinker might enable older adults to resist normative influences within relationships that support heavy drinking. This identity should retain strengths such as sociability without implying susceptibility to alcoholism. Community-level interventions to reinforce heavy drinking as unhealthy but not necessarily synonymous with dependency would facilitate older people's discussion of alcohol use with services, particularly if they clarified specifically how and why health can be at risk from heavy drinking in the absence of immediate symptoms. Future research could extend our findings by exploring identities around such interventions, and among older adults not using services.