Interventions to reduce the negative effects of alcohol consumption in older adults: a systematic review

Worldwide, alcohol-related diseases are responsible for 2.5 million deaths per year [1]. Therefore, alcohol consumption and associated negative impacts are a significant problem to public health. Projections from the World Health Organization (WHO) indicate that by 2025, alcohol consumption is expected to increase in almost half of the member states, something that will only be reversible with the implementation of “effective policy responses” [2]. The damage from alcohol use inflicts social and economic damage across many societies and this worldwide burden will increase if policy is not improved [2]. According to the WHO the highest levels of alcohol consumption are in the developed world, in particular Europe and the Americas, with intermediate levels of consumption in the Pacific and African regions with the lowest consumption in South-East Asia and the Eastern Mediterranean [2].

By 2050, around 22% of the world population will be aged 60 and over, with a significant proportion of these older individuals having a “pattern or level of drinking which places them at harm” p.656 [3]. In comparison to younger people, older adults are more susceptible to the detrimental effects of alcohol, as their tolerance to alcohol lowers with age. In addition, older people are also more likely to take prescription medications which, when taken with alcohol, can reduce effectiveness of medication, exacerbate side effects or even lead to the development of new illnesses [4].

Drinking more than five standard drinks per week and a history of an alcohol problem in men over the age of 50 has been found to quadruple the risk of developing psychiatric problems including depression and memory loss [5, 6]. Cognitive impairment can lead to an increased likelihood of falls [7] and because older people often have weaker bones, this can lead to hip fractures, which is one of the highest causes of death in the older population [8].

Significant events experienced during the life course have been associated with increased alcohol consumption. One such event is retirement which is associated with changes in drinking patterns [9]. However, the relationship between retirement and alcohol consumption is unclear. In their review of the literature, Zantinge et al. [10] found some studies reported an increase in alcohol use after retirement, while others reported a decrease, or no change. They concluded that individuals who retired involuntarily were generally more likely to increase alcohol consumption, with those choosing to leave less likely to change drinking pattern. There is also some evidence suggesting a reverse causality in this relationship between alcohol use and retirement, highlighting that men with existing alcohol problems who are eligible to retire are more likely to do so than men without such problems [11].

A recent review by Bhatia et al., [12] addressed recent advances in treatments for older people and the effectiveness of interventions for substance use problems as a whole. They concluded that older people are good at utilising interventions in this area and that they show positive outcomes. They do however note that the evidence base needs to be developed and refined. The current review is the first study to focus specifically on alcohol use in later life, as opposed to the broader topic of substance misuse.

The current study focuses on research that has evaluated alcohol interventions in older people. As alcohol is a complex problem, specific knowledge is needed to enable the development of effective alcohol interventions specifically for older alcohol users. Therefore, the aim of this study was to conduct a systematic review of interventions to reduce alcohol targeted at older individuals in order to examine the factors, conditions and motivations that contribute to successful intervention in this area.

Figure 1 (Appendix 1) contains a PRISMA flow diagram of included reviews and shows the number of articles obtained at each stage of the review process. Table 2 shows the seven trials included in this review. A narrative synthesis is provided on the characteristics of the included studies, including outcome measures and critical appraisal of the interventions. As all interventions had different outcome measures it was not appropriate to conduct a meta-analysis. All included studies were randomised controlled trials; therefore the CASP Tool for Randomised Controlled Trials was used as the appraisal tool [16]. All trials in this study were considered to have appropriately met the validity and quality appraisal in terms of CASP.

This analysis of seven published randomised controlled trials, involved a combined total of 3531 participants. Six of the trials were conducted in the United States of America and one in Denmark. Studies were carried out in primary care centres and community-based groups or provided elsewhere [17‐22]; the exact location is not specific in Hansen et al. [23]. The included studies showed that there were varying levels of success, all interventions showed improvements in at least one area of alcohol consumption or frequency of consumptions. This was not always significantly more than control groups and the potential reasons for this are explored below.

This is the first systematic review that focuses solely on the effectiveness of interventions to reduce alcohol use in older individuals and in particular individuals aged over 55. There were improvements in frequency and amount of alcohol consumption in at least one area in all included interventions. Some interventions were successful, apart from Hansen et al., [23] and Gordon et al., [19] who found that whilst interventions show significant reductions over time, they did not show significant differences in comparison to the control group individuals who received standard care. The improvements included lower amounts of frequency and consumption [17‐19] and reduced 7-day alcohol use ([20, 22]). In addition, other studies found lower rates of heavy episodic drinking and visits to physicians and emergency departments [17] a lower frequency of “binge drinking” episodes and frequency of excessive drinking [22] and harmful and hazardous drinking [18].

In a previous systematic review of interventions focusing on the wider issue of substance misuse in older adults, older adults were found to respond well to psychological treatments [12] Bhatia et al. found that although there were promising responses from participants in regards to current treatments; further examination should focus on a wider range of interventions which could be offered and may better suit older individuals. The current review begins to investigate these strengths of interventions, evidencing that individuals do appear to respond to psychological based treatments such as counselling and advice on behaviours, they also respond as well to educational tools, personalised reports that indicate their own level of risk and the use of diaries.

A common theme in this review was that studies often lacked a clear intervention description, which meant that determining which components were effective was problematic. The effective components were psychological based treatments such as counseling and advice on behaviours, educational tools, personalised risk reports and diaries. However, the information provided on these tools was not sufficiently detailed to clarify which specific elements of the interventions were effective. Information provided on the control groups that received “care as usual” was limited. One study provided some alcohol-specific information to control participants and some others (n = 3) provided a general health information sheet to participants; others explicitly stated that discussions surrounding alcohol were not discouraged (n = 3), however more information on the control groups would have been useful in assessing the effectiveness of the interventions.

Hansen et al. [23] found no significant difference between their brief motivational intervention group and the control group and they give numerous possible reasons for this. They stated that they included no pure control group for ethical reasons and that whilst control participants did not receive the brief motivational intervention, they were given leaflets on alcohol and treatment. This information alone may have resulted in the documented reductions as they concede that even the act of taking part in a health-based intervention may have led to a reduction in alcohol use. They also suggest that reductions could be due to “regression to the mean, social desirability bias and historical changes in alcohol consumption” p. 30.

All studies included were assessed as per the CASP guidelines. There are 11 questions included in the guidelines that help the researcher to assess trials systematically. By using this set of guidelines, this systematic review assessed each trial appropriately. All included trials were assessed and deemed to be valid.

Six of the seven studies included used some form of blinding outcomes, which adds to the validity of results. However Hansen et al. [23] did not use any blinding in their groups, declaring it unfeasible and Fink et al., [18] provide limited information on blinding. Fink et al. also admit that physicians in their three groups may have discussed the process, however they have no evidence that this did occur; this is a weakness of both included studies.

In terms of sample size, Gordon et al., [19] had a relatively small sample size, with no power calculation provided and the authors comment on this small sample as a potential limitation. Kuerbis et al. [9] also had a small sample for their pilot study. Sample sizes were larger in Ettner et al. (n = 1186) [17] however power was not addressed in this study, nor by either Fleming et al. [22] or Fink et al. [18] . Moore et al. [20] and Hansen et al. [23] addressed power, stating that numbers used ensured sufficient power, Hansen stating that power was sufficient providing that 75% of participants completed the full study, which they did. Using the support from other individuals, including family members and friends could be effective in reducing alcohol consumption. Only one study in this review used the support of other individuals within the intervention setting which consisted of speaking to family members about participants’ results [22]. Whilst this is useful to gain clarification on self-reported results, they do not explain how they affect alcohol-related behaviours.

Support from individuals could span beyond close friends and family members. Evidence suggests that social networks and groups to which an individual subscribes could also link to drinking behaviours and could include where an individual works or their age group. Views and behaviours in relation to drinking are linked to social background and may respond to interventions and public health messages accordingly [31]. Future research could investigate how family and social networks surrounding older individuals could contribute to a successful intervention.

All the papers included in this review contain at least some element of self-reporting. This may lead to socially desirable responding from participants, who do not want to disclose the full extent of their problems and under-report usage [32]. This can threaten the validity of trials and underestimate harm. Gordon et al. [19] suggest that there is a need for changes when measuring consumption and more reliable ways of reporting, as all reviewed studies relied on self-reporting which can lead to “conservative estimates of consumption” [18].

Another limitation of the papers included in this review is that the interventions appear to address different levels of drinking in individuals. Five of the included interventions were carried out on “at risk” or “hazardous” drinkers, whereas two focused on individuals who consumed alcohol. This could lead to disparity in the results, as it is not determinable which elements of the interventions work or for whom.

Although a positive effect was found with most of the interventions in this review, the estimate of the effect could be even higher using other measures such as reporting from family members, clinicians or research assistants on physical aspects of alcohol use, or could use a “graduated frequency” approach, which begins questioning with high levels of alcohol consumption to avoid socially desirable responding [33]. This approach should be used with caution however, as starting with higher levels risks normalising them to participants and potentially encourage higher consumption [33].

All but one of the studies included in this review were carried out in the USA, with one study conducted in Demark. Further research needs to be carried out cross-culturally as there could be some non-generalisable difference. One such difference could be the healthcare system that exists in the USA: where individuals may have to pay for treatment so may be less inclined to access support. It should also be noted that there is an underrepresentation of non-white, non-educated individuals in this field.

Some of the papers included in this review address randomisation and blinding, there was frequently only limited information presented on the techniques used. This is a limitation of the papers included and should be addressed in future work. The reliability and validity of interventions have been discussed within the original papers, however they could be demonstrated better and this could be done through better intervention description and there are tools available to assist with this. The Medical Research Council [34] offers a framework for the development and evaluation of complex interventions and we recommend that this is followed in future work. The TIDieR (Template for Intervention Description and Replication) checklist and guide would also be a useful tool in intervention description and would allow authors to provide a concise description of administered interventions and where this information can be found within the individual papers.

Some papers were excluded from this review due to their inclusion of comorbidities. However, such comorbidities may be integral when investigating excessive alcohol use in older people, for example, in their study examining inpatient treatment for alcohol dependence in a group of older people, Blow et al. [35] reported that 31% of their sample were experiencing at least one psychiatric illness. Alcohol misuse and dependence can contribute to psychiatric illness and lower levels of functioning, a risk significantly higher in older adults [5, 6]. Future work should explore how different groups may respond differently to interventions or have different motivations for drinking such as using alcohol as a coping mechanism [9].

There are several strengths to this review. This review is the first study that has conducted a systematic approach to data collection focusing on interventions aimed at reducing alcohol use in older people. Previous studies have been less specific and focus on substance abuse as a whole, or have included pharmaceutical interventions which may lead to a more reductionist approach to decreasing alcohol consumption in this field, rather than the current study which focusses solely on behaviour change interventions.

This study has numerous implications for public health. Individuals are living longer and their health is important if they are to continue to live independently and enjoy later life. The development and use of interventions to reduce alcohol use in older individuals will lead to prevention, or delay of diseases such as stroke, heart disease or cancer [36].

By using the information from this review, further work should investigate what works and for whom. Targeting interventions through public health practice, could lead to a significant reduction to the health and economic burdens of excessive alcohol use. In the UK, reductions in funding across the social care sector, are affecting people in later life [37]. By introducing and utilizing cost effective behaviour change interventions, older individuals will live longer, whilst also reducing the current financial burden on the economy.

This study has shown that while there is a growing evidence base for interventions for alcohol use in older individuals, there is still a need to conduct more research in the field to understand more about alcohol use in later life, and specifically understand which interventions work and for whom. Currently, interventions are aimed at general populations rather than focusing on older people. Older people are affected disproportionately by lifestyle changes such as bereavement, social isolation and loneliness and worklessness which may affect alcohol consumption [9]. More work is needed to establish the relationship between these factors and patterns of drinking in older people and also to look at varying levels of alcohol consumption across the life course, playing closer attention to stages of old age and factors such as retirement.