Effectiveness of alcohol brief interventions in general practice Emma Clossick and Sue Woodward
Harmful levels of alcohol consumption are a longstanding but continually growing health concern affecting individuals, and consequently populations. Through personalised advice, alcohol brief interventions (ABIs) target drinking habits, which are deemed damaging to health. They are not targeted at dependent drinkers for whom a more intensive and specialist service would be better suited. In response to recent shifts in public-health strategy away from ABIs, and the publication of new trials into their effectiveness, this article evaluates the value of ABIs through the review of eleven relevant studies. Findings suggest that hazardous drinkers did reduce their drinking patterns following ABI, but not more so than control participants who were given standard treatment. The evolution of standard treatment to include thorough screening, non-personalised advice and sometimes literature on alcohol intake delivers an intervention, that is effective in reducing drinking. Therefore, future research and practice should focus on screening methods and quality of standard care rather than on brief interventions. Key words: Alcohol brief interventions (ABIs) ■ Harmful and hazardous drinking ■ Alcohol strategy ■ Public health ■ Practice nursing
I
n 2011–2012, there were 1 220300 alcohol-related admissions to hospital. This is more than twice as many admissions as in 2002–2003 and providing estimated costs of alcohol-related harm to the NHS in England of £3.5 billion for the year (Health and Social Care Information Centre, (HSCIC) 2013). The prevalence of at-risk levels of drinking is not a new phenomenon but remains a critical problem, affecting individual, community and public health. This literature review will evaluate the effectiveness of general-practice-based alcohol brief interventions (ABIs) in reducing alcohol consumption and alcohol-related problems among harmful and hazardous drinkers. According to Lacey (2011), harmful levels of alcohol consumption lead to collateral damage, including relationship problems, financial distress and alcohol-related offending. These are in addition to direct health problems such as liver disease, mental health problems and injury, among others (Babor and HigginsBiddle, 2001). Despite the adverse effects, approximately 39%
Emma Clossick is Staff Nurse, St George’s Hospital NHS Trust, London and Sue Woodward is Lecturer, Florence Nightingale School of Nursing and Midwifery, King’s College London Accepted for publication: May 2014
574
of men and 27% of women in the UK are reported to drink above the recommended maximum consumption at least once in a week (HSCIC, 2013). Thus, encouraging healthier drinking behaviour must be a pressing concern for health professionals wishing to improve health and quality-of-life outcomes for individuals and populations. ABIs are short (5–15 minute) sessions of information and advice on reducing drinking, often accompanied by selfhelp literature. ABIs offer drinkers personalised feedback and structured advice about how to reduce alcohol-related risk (Hutchings et al, 2006). Though the content of ABIs may vary according to the patient and professional involved, NICE guidelines suggest they should accommodate the six key aspects shown in Box 1. ABIs are delivered specifically to harmful or hazardous rather than dependent drinkers (Box 2), reflecting a principle of health policy that alcohol strategy should be tailored to target drinkers according to type (Babor and Higgins-Biddle, 2001). In this review, ABIs are explored as interventions within primary care settings, delivered opportunistically and by a member of the multidisciplinary team, commonly the practice nurse, during routine patient visits to practice. Placing Box 1. FRAMES acronym for structuring an ABI Content of an Alcohol Brief Intervention ■ Feedback: feedback on the client’s risk for alcohol problems ■ Responsibility: highlight that the individual should take responsibility for change ■ Advice: explicitly advise reduction ■ Menu: outline options for change ■ Empathy: offer a warm, reflective and understanding approach ■ Self-efficacy: encourage optimism about behaviour change Adapted from NICE, 2010 Box 2. Definitions of types of drinking behaviour Hazardous drinking - a pattern of alcohol consumption over the recommended weekly limit (21 units for men, 14 for women), or which put an individual at risk of alcohol-related health problems. ■ Harmful drinking - a pattern of alcohol consumption that is already causing damage to health. ■ Dependent drinking – symptoms of physical and/or psychological addiction and actual or perceived difficulty maintaining normal function without drinking. Usually including withdrawal symptoms if intake is stopped. ■
Adapted from NHS Inform, 2013
© 2014 MA Healthcare Ltd
Abstract
British Journal of Nursing, 2014, Vol 23, No 11
British Journal of Nursing.Downloaded from magonlinelibrary.com by 147.188.128.074 on March 22, 2015. For personal use only. No other uses without permission. . All rights reserved.
LITERATURE REVIEW the intervention within routine primary care potentially yields many advantages; including that intervention before referral for secondary care enables a preventative approach (Babor and Higgins-Biddle, 2001). ABIs are also relevant to contemporary nursing roles in particular, given that health promotion responsibilities are increasingly found within a nursing remit (Whitehead, 2011). Nevertheless, the focus of the alcohol strategy in the UK has shifted in recent years, away from the input of professionals and towards the legal limits, through advertising, cost and availability. Indeed, guidelines from the National Institute for Health and Care Excellence (NICE) (2010) highlight pricing and marketing of alcohol as likely to be more important than the interventions of health professionals. There is therefore a need to appraise existing evidence on ABIs with a view to making the case for their continued inclusion within public-health strategy.
© 2014 MA Healthcare Ltd
Evidence base ABIs are already frequently practised within primary care (World Health Organization (WHO), 2006) and there is an existing research base relating to their effectiveness. A recent Cochrane review (McQueen et al, 2011) found benefits in reducing alcohol consumption but looked only at interventions delivered to inpatients. Another Cochrane review cautiously but positively assessed advice-based alcohol interventions targeted at individuals during pregnancy (Stade et al, 2009). More relevantly to this article, an earlier Cochrane review (Kaner et al, 2007), specifically into ABIs delivered to a general primary care population, found an average reduction of six drinks per week for recipients of ABIs, who usually tend to be white males (Kaner et al, 2013). However, as with all public-health interventions, there is a need to keep the evidence base current. A number of years has passed since the Cochrane review was produced in 2007 and a several more recent notable randomised controlled trials (RCTs) have emerged (Drummond et al, 2009; Butler et al, 2013; Kaner et al, 2013; Watson et al, 2013). Thus, the Cochrane database cannot currently provide up-to-date judgement on whether ABIs should remain as recommended care pathways for general practices. Moreover, the Cochrane review (Kaner et al, 2007) and other previous reviews of ABIs (e.g. Ballasteros et al, 2004) have drawn solely on RCTs to demonstrate the utility of interventions. RCTs are often described as the ‘gold standard’ in research design (Katikireddietal, 2001; Macintyre, 2011) because of their potential to minimise the impact of confounding factors. Yet there are hesitations surrounding how fully they can provide answers to questions surrounding public-health interventions. For instance,Victoria et al (2004) argue that RCT evidence must be supplemented with qualitative information to adequately assess whether an intervention can impact on the complex and long-term behavioural processes that are specific to public health. Qualitative studies allow the gathering of attitudes and opinions surrounding interventions, which can complement the information provided by statistics/ RCTs to allow a fuller assessment of the viability of a public-health intervention. Therefore, in using both qualitative and quantitative evidence, this review will benefit from a more varied evidence base than has been available previously.
British Journal of Nursing, 2014, Vol 23, No 11
Table 1. Database search results—with limits applied of English language only, published 2000-2014 Search terms
BNI
CINAHL
Embase
Medline
Total combined
‘Brief intervention’ and ‘alcohol’
66
16
1038
621
1741
‘Brief intervention’ and 17 ‘alcohol’ and ‘primary care’
3
241
152
413
‘Brief intervention’ and ‘alcohol’ and ‘general practice’
1
43
22
72
Results from database searching = 72
6
Results from reference list searching = 7
Results from contacting leading academics = 1
Excluded papers = 68 ■■ No primary research ■■ Duplicates ■■ Wrong sample ■■ Pilot study only Papers for further evaluation = 12 Excluded = 1 Wrong intervention
Included = 11
Figure 1. A summary of the search and papers included
Method Four databases (MEDLINE, Embase, British Nursing Index (BNI) and Cumulative Index to Nursing and Allied Health Literature (CINAHL)) were searched for relevant studies. The keywords ‘brief intervention’ and ‘alcohol’ were entered into database search tools, and the search limited to English language studies published between January 2000 and the January 2014. It was deemed that alcohol brief interventions do not generally or regularly go by any other description or facet. To narrow the search to the setting in question, and provide a suitably sized set of papers for screening, ‘general practice’ was introduced as an additional search term rather than ‘primary care’ (Table 1). Figure 1 details the process of finding the final eleven included papers. Inclusions needed to collect and present primary data on ABIs as opposed to offering summaries or reviews of pre-existing evidence. Those looking at interventions within specialist-alcohol treatment, or aimed at populations with mental-health or dependency diagnoses were excluded on the basis that such individuals are best suited to specialist not routine care-based interventions (Babor and Higgins-Biddle, 2001). Fernandez (2007) was included for further evaluation but then excluded on these grounds, as intervention was delivered by alcohol specialist services rather than general primary care. Studies involving
575
British Journal of Nursing.Downloaded from magonlinelibrary.com by 147.188.128.074 on March 22, 2015. For personal use only. No other uses without permission. . All rights reserved.
Results Eleven studies were found to meet the criteria described. Nine were RCTs. Two of these (Aalto et al, 2000; 2001) used an identical research design but had different samples, according to gender and inclusion eligibility. Two studies were qualitative focus-group studies. Table 2 summarises the sample, methodology, findings and strengths and limitations of each. The Critical Appraisal Skills Programme (CASP) tools for quantitative (Public Health Resource Unit (PHRU), 2006) and qualitative studies (CASP International, 2013) were used to assess the validity and relevance of the included studies. Each of the quantitative studies found that patients receiving ABIs reduced alcohol consumption or experienced fewer alcohol-related problems after intervention. Neither of the qualitative studies found any evidence of this. Furthermore, although the RCTs did see reductions in drinking following intervention, only one of the seven RCTs (Curry et al, 2003) found ABI to be more effective than the control of nontreatment or very basic advice. A dominant theme that emerges from the studies is the problem of loss to follow-up. For instance, in Kaner et al’s (2013) study, only 57% of those randomised to extended intervention cooperated with the trial for a long enough period to receive it. The concern is that those likely to drop out between baseline and follow-up are potentially unrepresentative of the overall sample. Edwards and Rollnick (1997) suggest drop-out is most common among those younger than the average and, crucially, less likely to respond to the ABI. Furthermore, in all studies except Butler et al
576
(2013), drop-out tended to be higher within the intervention cohorts than control. The fact that a frequent consequence of the intervention was non-compliance presents as evidence against the effectiveness of ABIs. Some of the stronger studies (Lock et al, 2006; Butler et al, 2013) took care to address the issues of loss to follow-up and its potential effect on outcomes, specifically through using an intention-to-treat (ITT) analysis. Within this approach, every individual who begins treatment is considered part of the trial, whether they go on to finish or not (Hollis and Campbell, 1999), and their outcomes are analysed according to their original grouping, as recommended by the CASP framework (PHRU, 2006). This avoids drop-out breaking the initial randomisation in a study, and the resulting evidence provides comment on how effectively the intervention was delivered as well as its benefits post delivery. Curry et al (2003) make an exception to this approach, using multiple imputation, whereby statistics software assigns multiple plausible outcomes for lost participants, and the presumed results are obtained by combining them. This addressed the problem of differential loss to follow-up. However, it arguably overestimated treatment effect relative to ITT analysis, by allowing some lost subjects to be assigned positive response, despite their ultimate non-engagement. A number of studies failed to provide sufficiently large or representative samples. The CASP framework highlights sample size power as a key factor in minimising the play of chance within studies. Yet, only a small number of studies used power calculations to legitimise the size of their sample (as did the Kaner et al, 2013 study which required 435 patients and had an initial sample of 756), and within those that did, the size required was more often not met. (Lock et al, 2006; Drummond et al, 2009) The qualitative studies were unsurprisingly especially small; Beich et al (2002) had 24 participants and Lock (2004) had 31. As well as size, the quality of recruitment methods is worthy of particular reflection. Notable bias in recruitment procedures was seen in Drummond et al (2009) and in Watson et al (2013), where only males and older participants respectively were recruited owing to their perceived greater response to ABIs. So, it is important to acknowledge that sample quality and follow-up were methodological issues which affected the validity of the 11 included studies although certain compensatory strategies, such as ITT analysis, were often employed. Nevertheless, the commonality of findings across the RCTs in particular (8 out of 9 suggesting ABI was not statistically or qualitatively better than control) does suggest that sufficient rigour was maintained to allow a common theme to emerge. As discussed, inclusion criteria within the search strategy allowed for considerable variation in outcome measures in order to reflect the broad nature of public-health outcomes. The corresponding difficulty is that direct comparisons between study results are not straight-forward. One common measure is units or standards drinks per week, enabling certain like-for-like comparisons. For instance, Lock et al (2006), found a drop of 7 units per week among the intervention arm at 6 weeks, whereas Aalto et al (2001) did not find any reduction in number of drinks over 12 months. However, as argued by Sanchez-Craig (1992), such a reliance
© 2014 MA Healthcare Ltd
an additional health factor, such as pregnancy or injury, were excluded on the grounds of dissimilar sample and reduced transferability of findings. However, studies in which the intervention was conducted by a non-nursing professional, i.e. a doctor or researcher, were not excluded, so long as the intervention did not require skills or specialist training beyond those routinely used by a practice nurse. The search criteria left considerable room for variation between included studies. As mentioned, qualitative and quantitative studies were permitted. Moreover, the wide spectrum of consequences associated with a public-health problem such as alcohol consumption poses difficulty in finding unified outcome measures. For instance, it is known that the physical, psychological, social and financial costs of harmful drinking lead to serious and diverse knock-on effects from impaired liver function to emotional harm, relationship breakdown and social isolation. Patterns of drinking, plus secondary consequences related to drinking, are intended targets of ABIs as well as the number of units consumed. So, objective measurements such as blood tests and more subjective measures such as self-reported wellbeing are justifiably used within relevant literature. Studies with quite different outcome measures, from gamma-glutamyl transferase levels (Aalto et al, 2000) to episodes of drink-driving (Curry et al, 2003) were therefore included in this review. The value of assessing the complete, multi-dimensional impact of a public-health intervention is felt to outweigh the benefits of identical outcome measures.
British Journal of Nursing, 2014, Vol 23, No 11
British Journal of Nursing.Downloaded from magonlinelibrary.com by 147.188.128.074 on March 22, 2015. For personal use only. No other uses without permission. . All rights reserved.
LITERATURE REVIEW Table 2. Summary of literature search studies and findings Author and year
Sample size, country and inclusion criteria
Methodology
Results
Lock et al (2006)
127 adults
Intervention group received a 5–10 minute brief intervention. Control received ‘standard advice’ and a leaflet.
No statistically significant evidence that brief intervention provides superior outcomes to control. Reduced consumption for 51% of ABI participants and 59% of control
North east of England Randomised controlled trial
AUDIT score >8 or >7 for women
Drummond et al (2009)
112 male adults
Randomised controlled trial
AUDIT score >= 8
Butler et al (2013)
1827 adults
Randomised controlled trial
AUDIT-C score >= 4 (men) or 3 (women)
Maisto et al (2001)
301 adults USA
Kaner et al (2013)
Intervention group received stepped care (extended ABI) and control group received minimal intervention (standard ABI)
South Wales
South Wales
Randomised controlled trial
Follow-up at 6 and 12 months. Outcome measures—AUDIT scores, drink units per week and Drinking Problems Index scores.
Follow-up at 6 months post intervention. Primary outcome measures total number of drinks consumed in previous 180 days, number of days abstinent and average drinks/day. Intervention group received behaviour-change counselling from trained GP or practice nurse, control group received screening and consultations from GP/ nurse not trained in behaviour change.
Intervention groups received either brief advice or motivational enhancement (extended intervention), control group received standard care (generalised advice). Follow-up at 1, 3, 6, 9 and 12 months. Outcome measures as number of drinks consumed, and broader drinking behaviour measures.
756 adults
Intervention group received either brief advice + leaflet (basic ABI) or basic life counselling + brief advice + leaflet (extended ABI). Control group received screening and leaflet only.
AUDIT >= 8 Follow up at 6 and 12 months. Outcome measures—AUDIT Score, alcohol problems questionnaire, service use questionnaire, patient satisfaction questionnaire and use of ‘Drink Line’ helpline.
Randomised controlled trial Aalto et al (2000)
118 female adults Finland
Randomised controlled trial Aalto et al (2001)
© 2014 MA Healthcare Ltd
Randomised controlled trial
Those assigned to intervention received either 7 (intervention group A) or 3 (intervention group B) sessions. Control group received advice to reduce drinking upon baseline measurement only.
CAGE test giving >= Two affirmative answers or self-reported alcohol consumption >= 190 g absolute ethanol per week
Follow-up at 3 years post intervention. Main outcome measures self-reported weekly alcohol consumption, bloodtest markers including gamma-glutamyl transferase (GGT) and mean corpuscular volume (MCV).
296 male adults
Methodology matched Aalto et al (2000)
Finland CAGE test giving >= Three affirmative answers or self-reported alcohol consumption >= 280 g absolute ethanol per week
British Journal of Nursing, 2014, Vol 23, No 11
Greater reported intention to change at 3 months from intervention group. Yet no statistically significant difference in reported behaviour at 3 or 12 months between intervention and control.
In-person follow up immediately after consultation plus questionnaire follow up at 3 and 12 months.
AUDIT score >= 8 or quantity- frequency score of >= 16 (male)/ 12 (female) standard drinks per week
Practices from across England and Wales
Basic ABI led to significantly reduced alcohol consumption, as did extended intervention. The slight difference in effect (stepped intervention produced greater results than standard ABI) was not statistically significant.
Alcohol consumption was reduced in both interventions, as well as with the standard care cohort. Found that very minimal intervention is sufficient.
All groups reduced alcohol consumption and smaller reduction in control group was not statistically significant. High drop-out from extended intervention arm implies greater difficulty of implementing counselling relative to brief intervention or screening only.
Depending on outcome variable reduction was between 27% and 75% of total participants. No statistically different outcome between groups A, B or Control regardless of outcome measure chosen.
Greater reduction in intervention groups A and B than in control group but did not reach level of clinical significance (group B = 5–13% greater reduction) Statistically significant reduction in MCV was found across all groups.
577
British Journal of Nursing.Downloaded from magonlinelibrary.com by 147.188.128.074 on March 22, 2015. For personal use only. No other uses without permission. . All rights reserved.
529 adults over 55 years old England and Scotland AUDIT score >= 8
Intervention (‘stepped care’) group received 20-minute behavioural change counselling plus motivational enhancement therapy and additional referrals if indicated. Control group received 5-minute brief intervention. Primary outcome as average drinks per day, followed up at 12 months.
Randomised controlled trial Curry et al (2003)
307 adults USA
Randomised controlled trial Lock (2004)
Four eligibility criteria relating to quantity of drinks consumed and three indicators of at-risk drinking behaviour (relating to chronic drinking, binge drinking and drink driving) 31 adults
Intervention group received advice from physician during initial visit, self-help manual, written feedback and up to three follow-up counselling calls. Control group received no intervention. Follow-up at 3 months and 12 months. Outcome measures quantity of drinks consumed and three indicators of at-risk drinking (matching screening indicators).
Focus group based, using a funnel approach to discussion surrounding attitudes to ABIs
North east of England Combination of random and purposive sampling All drinking statuses included
Beich et al (2002)
24 general practitioners
Qualitative study using focus group and individual interviewing Specific focus on the feasibility of implementing ABIs, as opposed to their theoretical effectiveness.
on quantitative outcomes risks ignoring changes in pattern of consumption. Therefore, inclusion of aspects such as drinkdriving (Curry et al, 2003) or intention to change habits (Butler et al, 2013) allows measurement of how ABIs change overall behaviour and risk. Studies also differed according to timeframe; Aalto et al (2000; 2001) followed up at 36 months, others at 12 or 3. Given that the effect of ABIs on longer term habits is deemed a greater challenge than immediate behaviour changes (Kaner et al, 2007); this may explain the limited impact found in the Aalto studies. In summary, the included studies are somewhat weak in terms of providing uniform outcomes for comparison, but strong in that this allows for multidimensional outcome assessments, going beyond previous evidence bases. The included studies display certain credentials, recognised by the CASP approaches, which reinforce the validity of their findings. An oft-cited weakness of quantitative research, particularly RCTs, is their tendency to be efficacy rather than effectiveness trials, so removed from real-world practice (Victoria et al, 2004). An intervention proven effective in ideal conditions may be considerably less straightforward in routine practice. A majority of the included studies were efficacy trials, with admittedly smaller sample sizes, but also
578
Considerably higher drop-out for stepped care (two thirds did not attend step 2 onwards). Outcomes measured as number of drinks consumed fell across both intervention and control. However reductions were more marked for at-risk patterns of drinking behaviour (specifically drink driving) for intervention group.
Considerable barriers to effective implementation of ABIs. Population-wide education campaigns favoured over one-toone interventions.
Qualitative data and outcome measures—including statements on by whom, whereabouts and to what effect ABIs could or should be delivered.
Denmark Participants all had experience of delivering ABIs as part of a previous study
Both stepped care and control groups reduced drinking levels at 12 months. The difference in amount of reduction between intervention and control was only 0.025, providing no statistically significant difference.
Doctors did not feel they could recommend ABIs as a strategy for reducing risky alcohol consumption. Community-wide preventative methods favoured over ABIs.
greater generalisability than often provided by effectiveness studies. The two qualitative studies were especially valuable for assessing whether ideal-condition findings can be generalised into practice, through gathering experiences of recipients and deliverers of ABIs. The concept that clinician and patient engagement are equally crucial for intervention effectiveness (Hutchings et al, 2006) makes Beich et al’s (2002) study particularly relevant for this review, and again demonstrates the value of including qualitative evidence to enable judgement on this particular intervention.
Discussion As noted, the studies as a whole offered no evidence that brief intervention, or further personalised intervention, yielded more impressive results than alcohol screening and minimal advice only. Figure 2 demonstrates Kaner et al’s (2013) data. Given that ABIs have been part of alcohol strategy for some time (Babor and Higgins-Biddle, 2001) and supported by acclaimed evidence (Kaner et al, 2007), it is important to assess whether the problem is with the validity of studies included in this review, or whether a challenge to the previous evidence base is warranted. Regarding study validity, certain weaknesses have been conceded above. However,
© 2014 MA Healthcare Ltd
Watson et al (2013)
British Journal of Nursing, 2014, Vol 23, No 11
British Journal of Nursing.Downloaded from magonlinelibrary.com by 147.188.128.074 on March 22, 2015. For personal use only. No other uses without permission. . All rights reserved.
LITERATURE REVIEW
© 2014 MA Healthcare Ltd
Conclusion The implications for this review are great; the suggestion being that to screen and to discuss alcohol with a heavy drinker is equally as effective as an ABI, primarily because there is such considerable overlap between the two. It is possible to conclude, then, that the lesser impact of ABIs found in this review than in the Cochrane review (Kaner et al, 2007) could be a result of the advancement of screening and ‘non-intervention’ practices, so that in this review’s later collection of studies, control effect on drinking is considerably greater.Two of the most recent studies, Kaner et al (2013) and Butler et al (2013) support this assumption, having provided
British Journal of Nursing, 2014, Vol 23, No 11
Baseline
6-month follow up
12-month follow up
40 35 Proportion of patients
Sanchez-Craig (1992) argues that commonality of findings across a number of studies serves to reinforce the validity of the outcomes. Strengths of the evidence reviewed, in terms of common findings, combination of qualitative and quantitative evidence, and specific credibility approaches used (including cluster randomisation, and ITT analysis) combine to allow an acceptable level of confidence that lack of validity is not the explanation for the limited intervention effect found. Indeed, Watson et al (2013) state that the absence of benefit of intervention over basic care is so noticeable that issues of methods and validity cannot provide an explanation. Clarifying the nature of the intervention’s evident weakness facilitates more in-depth conclusions. Rather than failing to reduce at-risk drinking behaviour per se, (they succeeded in doing so in 6 out of 9 studies), ABIs only failed to reduce drinking more than control arms. Therefore, discussion must look at the content of the controls themselves. While screening practices varied across studies, they were commonly thorough and multi-layered. The majority including the minimum of an AUDIT score assessment (Table 3), which itself often includes an introductory statement as to the hazards of heavy drinking. When assessment of drinking is done properly, the distinction between screening and outright intervention ‘is inevitably arbitrary’ (SanchezCraig, 1992: 661). AUDIT is arguably in itself a first-stage intervention (WHO, 2006). Follow-up procedures could also be interventionist, when involving re-assessment of drinking patterns and done frequently over a lengthy period of time. Butler et al (2013), for instance, required all participants to report and re-evaluate their drinking behaviours at 0, 3 and 12 months regardless of their being randomised to control or intervention. Maisto et al (2001) did so even more frequently, at 1, 3, 6, 9 and 12 months. Moreover, the control group often included basic support, namely short advice on initial visit and a leaflet, so there is an absence of a strictly non-treatment comparison group. The two Aalto studies (2000; 2001) and Watson et al (2013) affirm this, having explicitly ruled out the prospect of no-advice control as unethical. The argument is that, given long-standing evidence on the harmful effects of heavy drinking, nurses would be demonstrating unethical practice if they were to abstain altogether from commenting on such behaviour. Where outcomes were noticeably better for intervention than control (i.e. Curry et al, 2003) controls were strictly no-advice, separating this study from the others (and raising questions about the different ethical assumptions shaping their design).
30 25 20 15 10 5 0
Patient information leaflet
Brief advice
Brief lifestyle counselling
Intervention Figure 2. Proportion of patients (by intervention type) scoring negative status on alcohol use disorders identification (AUDIT) test (Kaner et al, 2013)
Table 3. Example of an AUDIT screen Alcohol Use Disorders Identification Test (Audit) Please circle the answer that is correct for you. ■■ How
Never
often do you have a drink containing alcohol? Monthly 2 or 4 2 or 3 or less times a month times per week
4 or more times a week
■■ How
many drinks containing alcohol do you have on a typical day when you are drinking? 1 or 2 3 or 4 5 or 6 7 or 9 10 or more ■■ How
often do you have six or more drinks on one occasion? Never Less than Monthly 2 or 3 4 or more monthly times per week times a week ■■ How
often during the last year have you found that you were not able to stop drinking once you had started? Never Less than Monthly 2 or 3 4 or more monthly times per week times a week ■■ How
often during the last year have you failed to do what was normally expected from you because of drinking? Never Less than Monthly 2 or 3 4 or more monthly times per week times a week ■■ How
often during the last year have you needed a first drink in the morning to get yourself going after a heavy drinking session? Never Less than Monthly 2 or 3 4 or more monthly times per week times a week ■■ How
often during the last year have you had a feeling of guilt or remorse after drinking? Never Less than Monthly 2 or 3 4 or more monthly times per week times a week ■■ How
often during the last year have you been unable to remember what happened the night before because you had been drinking? Never Less than Monthly 2 or 3 4 or more monthly times per week times a week ■■ Have
No
you or someone else been injured as a result of drinking? Yes, but not in last year Yes, during the last year
■■ Has
a relative or friend, or a doctor or other health worker, been concerned about your drinking or suggested you cut down? No Yes, but not in last year Yes, during the last year Source: National Institute of Alcohol Abuse and Alcoholism, 2014
579
British Journal of Nursing.Downloaded from magonlinelibrary.com by 147.188.128.074 on March 22, 2015. For personal use only. No other uses without permission. . All rights reserved.
Conflict of interest: none. Aalto M, Saksanen R, Laine P et al (2000) Brief intervention for female heavy drinkers in routine general practice: a 3-year randomized, controlled study. Alcohol Clin Exp Res 24(11): 1680-6 Aalto M, Seppä K, Mattila P et al (2001) Brief intervention for male heavy drinkers in routine general practice: a three-year randomized controlled study. Alcohol Alcohol 36(3): 224-30 Appleby J (2012) Drinking nation: have we had enough? BMJ 344: e2634. doi: 10.1136/bmj.e2634 Babor TF, Higgins-Biddle JC (2001) Brief Intervention For Hazardous and Harmful Drinking. World Health Organization, Geneva. http://tinyurl. com/dynmmmn (accessed 30 May 2014) Ballesteros J, Duffy JC, Querejeta I, Ariño J, González-Pinto A (2004) Efficacy
KEY POINTS n Current
levels of alcohol consumption in the UK impact negatively on the health of the population and add a considerable burden to public finances
n It
is appropriate to locate nursing interventions relating to alcohol within primary care in order to enable a preventative approach, aiming to tackle hazardous drinking before it escalates into critical health and social problems
n Discussing
issues surrounding alcohol consumption with individuals drinking harmful or hazardous amounts does have a beneficial effect on reducing their intake. However, thorough alcohol screening and basic alcohol advice is equally as effective, and more feasible to deliver, than lengthier or personalised interventions
580
of brief interventions for hazardous drinkers in primary care: systematic review and meta-analyses. Alcohol Clin Exp Res 28(4): 608-18 Beich A, Gannik D, Malterud K (2002) Screening and brief intervention for excessive alcohol use: qualitative interview study of the experiences of general practitioners. BMJ 325(7369): 870 Butler CC, Simpson SA, Hood K et al (2013).Training practitioners to deliver opportunistic multiple behaviour change counselling in primary care: a cluster randomised trial. BMJ 346: f1191. doi: http://dx.doi.org/10.1136/ bmj.f1191 Critical Appraisal Skills Programme International (2013) 10 questions to help you make sense of qualitative research. http://tinyurl.com/jwzbudn (accessed 30 May 2014) Curry SJ, Ludman EJ, Grothaus LC, Donovan D, Kim E (2003) A randomized trial of a brief primary-care-based intervention for reducing at-risk drinking practices. Health Psychol 22(2): 156-65 Drummond C, Coulton S, James D et al (2009) Effectiveness and costeffectiveness of a stepped care intervention for alcohol use disorders in primary care: pilot study. Br J Psychiatry 195(5): 448-56 Edwards AG, Rollnick S (1997) Outcome studies of brief alcohol intervention in general practice: the problem of lost subjects. Addiction 92(12): 1699-704 Fernandez J (2007) Client engagement with a specialist nurse-led alcohol service in primary care. Primary Health Care 17(1): 22-24 Health and Social Care Information Centre (2013) Statistics on Alcohol: England, 2013. http://tinyurl.com/pkm3eeg (accessed 30 May 2014) Hollis S, Campbell F (1999) What is meant by intention to treat analysis? Survey of published randomised controlled trials. BMJ 319(7211): 670-4 Hutchings D, Cassidy P, Dallolio E, Pearson P, Heather N, Kaner E (2006) Implementing screening and brief alcohol interventions in primary care: views from both sides of the consultation. Primary Health Care Research and Development 7(3): 221-9. doi: http://dx.doi. org/10.1191/1463423606pc292oa Kaner EF, Beyer F, Dickinson HO et al (2007) Effectiveness of brief alcohol interventions in primary care populations. Cochrane Database Syst Rev (2): CD004148 Kaner E, Bland M, Cassidy P et al (2013) Effectiveness of screening and brief alcohol intervention in primary care (SIPS trial): pragmatic cluster randomised controlled trial. BMJ 346: e8501. doi: 10.1136/bmj.e8501 Katikireddi SV, Higgins M, Bond L, Bonell C, Macintyre S (2011) How evidence based is English public health policy? BMJ 343: d7310. doi: 10.1136/bmj.d7310. Lacey J (2011) Alcohol brief interventions: facilitating effective delivery. Community Practitioner 84(4): 36-8 Lock C (2004) Alcohol and brief intervention in primary health care: what do patients think? Primary Health Care Research and Development 5(2): 162-78. Lock CA, Kaner E, Heather N et al (2006) Effectiveness of nurse-led brief alcohol intervention: a cluster randomized controlled trial. J Adv Nurs 54(4): 426-39 Macintyre S (2011) Good intentions and received wisdom are not good enough: the need for controlled trials in public health. J Epidemiol Community Health 65(7): 564-7. doi: 10.1136/jech.2010.124198. Epub 2010 Maisto SA, Conigliaro J, McNeil M, Kraemer K, Conigliaro RL, Kelley ME (2001) Effects of two types of brief intervention and readiness to change on alcohol use in hazardous drinkers. J Stud Alcohol 62(5): 605-14 McQueen J, Howe TE, Allan L, Mains D, Hardy V (2011) Brief interventions for heavy alcohol users admitted to general hospital wards. Cochrane Database Syst Rev (8): CD005191 doi: 10.1002/14651858.CD005191.pub3 National Institute for Health and Care Excellence (2010) Alcohol-use disorders: preventing harmful drinking. NICE, London. http://tinyurl.com/ctolc2c (accessed 30 May 2014) National Institute on Alcohol Abuse and Alcoholism (2014) Screening Tests. http://tinyurl.com/6w4zplw (accessed 30 May 2014) NHS Inform (2013) Alcohol Misuse, defining a drink problem. http://tinyurl. com/k82jscp (accessed 30 May 2014) Public Health Resource Unit (2006) Critical appraisal skills programme (CASP) making Sense of Evidence. 10 Questions to Help You Make Sense of Randomised Controlled Trials. http://tinyurl.com/8ms2ck9 (accessed 30 May 2014) Sanchez-Craig M (1992) Toward a public health model to preventing alcohol problems. Addiction 89(6): 660–2 Stade BC, Bailey C, Dzendoletas D, Sgro M, Dowswell T, Bennett D (2009) Psychological and/or educational interventions for reducing alcohol consumption in pregnant women and women planning pregnancy. Cochrane Database Syst Rev (2): CD004228. doi: 10.1002/14651858.CD004228.pub2 Victora CG, Habicht JP, Bryce J (2004) Evidence-based public health: moving beyond randomized trials. Am J Public Health 94(3): 400-5 Watson JM, Crosby H, Dale VM et al (2013) AESOPS: a randomised controlled trial of the clinical effectiveness and cost-effectiveness of opportunistic screening and stepped care interventions for older hazardous alcohol users in primary care. Health Technol Assess 17(25): 1-158. doi: 10.3310/hta17250 Whitehead D (2011) Health promotion in nursing: a Derridean discourse analysis. Health Promot Int 26(1): 117-27. doi: 10.1093/heapro/daq073. Epub 2010 World Health Organization (2006) Screening and brief intervention for alcohol problems in primary health care. http://tinyurl.com/ma98wje (accessed 30 May 2014)
© 2014 MA Healthcare Ltd
control arms with considerable non-personalised input, detailed assessments and frequent re-assessments, followed by trial results which do not advocate ABIs over usual care. Consequently, the recommendations of this review are not that nurses should avoid the matter of harmful drinking, given the apparently weak evidence as to the relative impact of ABIs. Rather, the opposite, as screening for alcohol use and raising the issue of drinking within routine practice has proven to lower risky drinking behaviour to a similar degree as structured ABIs, and so this practice should continue. The acceptability of nursing interventions is of course critical to their success. Therefore, the difficulties implementing ABIs or personalised advice, demonstrated by Watson et al (2013) and by the two qualitative studies in the review (Beich et al, 2002 and Lock, 2004), are also noteworthy. Standard screening and simple advice practices should be recommended for their transferability and impact, over lengthier, personalised ABI. To strengthen the theory established within this paper, it would be helpful to see further studies of nursing interventions for harmful levels of drinking. Studies should begin by accepting that a strictly no-treatment control approach would be potentially unethical, given the evidence that discussion or screening itself amounts to an intervention, and an effective one if done thoroughly. Different screening methods should be tested, and with a focus on effectiveness over efficacy trials to reinforce the need for interventions to be transferable to practice. Best practices from previous studies, notably ITT analysis and power calculations, should be copied to ensure maximum possible validity. Also, given the argument for a role for qualitative evidence in public health strategy, it is vital to gather larger and higher-quality qualitative studies on alcohol interventions to supplement the BJN existing RCTs.
British Journal of Nursing, 2014, Vol 23, No 11
British Journal of Nursing.Downloaded from magonlinelibrary.com by 147.188.128.074 on March 22, 2015. For personal use only. No other uses without permission. . All rights reserved.
Harmful levels of alcohol consumption are a longstanding but continually growing health concern affecting individuals, and consequently populations. Through personalised advice, alcohol brief interventions (ABIs) target drinking habits, which are deemed damaging to health. They are not targeted at dependent drinkers for whom a more intensive and specialist service would be better suited. In response to recent shifts in public-health strategy away from ABIs, and the publication of new trials into their effectiveness, this article evaluates the value of ABIs through the review of eleven relevant studies. Findings suggest that hazardous drinkers did reduce their drinking patterns following ABI, but not more so than control participants who were given standard treatment. The evolution of standard treatment to include thorough screening, non-personalised advice and sometimes literature on alcohol intake delivers an intervention, that is effective in reducing drinking. Therefore, future research and practice should focus on screening methods and quality of standard care rather than on brief interventions. Key words: Alcohol brief interventions (ABIs) ■ Harmful and hazardous drinking ■ Alcohol strategy ■ Public health ■ Practice nursing
I
n 2011–2012, there were 1 220300 alcohol-related admissions to hospital. This is more than twice as many admissions as in 2002–2003 and providing estimated costs of alcohol-related harm to the NHS in England of £3.5 billion for the year (Health and Social Care Information Centre, (HSCIC) 2013). The prevalence of at-risk levels of drinking is not a new phenomenon but remains a critical problem, affecting individual, community and public health. This literature review will evaluate the effectiveness of general-practice-based alcohol brief interventions (ABIs) in reducing alcohol consumption and alcohol-related problems among harmful and hazardous drinkers. According to Lacey (2011), harmful levels of alcohol consumption lead to collateral damage, including relationship problems, financial distress and alcohol-related offending. These are in addition to direct health problems such as liver disease, mental health problems and injury, among others (Babor and HigginsBiddle, 2001). Despite the adverse effects, approximately 39%
Emma Clossick is Staff Nurse, St George’s Hospital NHS Trust, London and Sue Woodward is Lecturer, Florence Nightingale School of Nursing and Midwifery, King’s College London Accepted for publication: May 2014
574
of men and 27% of women in the UK are reported to drink above the recommended maximum consumption at least once in a week (HSCIC, 2013). Thus, encouraging healthier drinking behaviour must be a pressing concern for health professionals wishing to improve health and quality-of-life outcomes for individuals and populations. ABIs are short (5–15 minute) sessions of information and advice on reducing drinking, often accompanied by selfhelp literature. ABIs offer drinkers personalised feedback and structured advice about how to reduce alcohol-related risk (Hutchings et al, 2006). Though the content of ABIs may vary according to the patient and professional involved, NICE guidelines suggest they should accommodate the six key aspects shown in Box 1. ABIs are delivered specifically to harmful or hazardous rather than dependent drinkers (Box 2), reflecting a principle of health policy that alcohol strategy should be tailored to target drinkers according to type (Babor and Higgins-Biddle, 2001). In this review, ABIs are explored as interventions within primary care settings, delivered opportunistically and by a member of the multidisciplinary team, commonly the practice nurse, during routine patient visits to practice. Placing Box 1. FRAMES acronym for structuring an ABI Content of an Alcohol Brief Intervention ■ Feedback: feedback on the client’s risk for alcohol problems ■ Responsibility: highlight that the individual should take responsibility for change ■ Advice: explicitly advise reduction ■ Menu: outline options for change ■ Empathy: offer a warm, reflective and understanding approach ■ Self-efficacy: encourage optimism about behaviour change Adapted from NICE, 2010 Box 2. Definitions of types of drinking behaviour Hazardous drinking - a pattern of alcohol consumption over the recommended weekly limit (21 units for men, 14 for women), or which put an individual at risk of alcohol-related health problems. ■ Harmful drinking - a pattern of alcohol consumption that is already causing damage to health. ■ Dependent drinking – symptoms of physical and/or psychological addiction and actual or perceived difficulty maintaining normal function without drinking. Usually including withdrawal symptoms if intake is stopped. ■
Adapted from NHS Inform, 2013
© 2014 MA Healthcare Ltd
Abstract
British Journal of Nursing, 2014, Vol 23, No 11
British Journal of Nursing.Downloaded from magonlinelibrary.com by 147.188.128.074 on March 22, 2015. For personal use only. No other uses without permission. . All rights reserved.
LITERATURE REVIEW the intervention within routine primary care potentially yields many advantages; including that intervention before referral for secondary care enables a preventative approach (Babor and Higgins-Biddle, 2001). ABIs are also relevant to contemporary nursing roles in particular, given that health promotion responsibilities are increasingly found within a nursing remit (Whitehead, 2011). Nevertheless, the focus of the alcohol strategy in the UK has shifted in recent years, away from the input of professionals and towards the legal limits, through advertising, cost and availability. Indeed, guidelines from the National Institute for Health and Care Excellence (NICE) (2010) highlight pricing and marketing of alcohol as likely to be more important than the interventions of health professionals. There is therefore a need to appraise existing evidence on ABIs with a view to making the case for their continued inclusion within public-health strategy.
© 2014 MA Healthcare Ltd
Evidence base ABIs are already frequently practised within primary care (World Health Organization (WHO), 2006) and there is an existing research base relating to their effectiveness. A recent Cochrane review (McQueen et al, 2011) found benefits in reducing alcohol consumption but looked only at interventions delivered to inpatients. Another Cochrane review cautiously but positively assessed advice-based alcohol interventions targeted at individuals during pregnancy (Stade et al, 2009). More relevantly to this article, an earlier Cochrane review (Kaner et al, 2007), specifically into ABIs delivered to a general primary care population, found an average reduction of six drinks per week for recipients of ABIs, who usually tend to be white males (Kaner et al, 2013). However, as with all public-health interventions, there is a need to keep the evidence base current. A number of years has passed since the Cochrane review was produced in 2007 and a several more recent notable randomised controlled trials (RCTs) have emerged (Drummond et al, 2009; Butler et al, 2013; Kaner et al, 2013; Watson et al, 2013). Thus, the Cochrane database cannot currently provide up-to-date judgement on whether ABIs should remain as recommended care pathways for general practices. Moreover, the Cochrane review (Kaner et al, 2007) and other previous reviews of ABIs (e.g. Ballasteros et al, 2004) have drawn solely on RCTs to demonstrate the utility of interventions. RCTs are often described as the ‘gold standard’ in research design (Katikireddietal, 2001; Macintyre, 2011) because of their potential to minimise the impact of confounding factors. Yet there are hesitations surrounding how fully they can provide answers to questions surrounding public-health interventions. For instance,Victoria et al (2004) argue that RCT evidence must be supplemented with qualitative information to adequately assess whether an intervention can impact on the complex and long-term behavioural processes that are specific to public health. Qualitative studies allow the gathering of attitudes and opinions surrounding interventions, which can complement the information provided by statistics/ RCTs to allow a fuller assessment of the viability of a public-health intervention. Therefore, in using both qualitative and quantitative evidence, this review will benefit from a more varied evidence base than has been available previously.
British Journal of Nursing, 2014, Vol 23, No 11
Table 1. Database search results—with limits applied of English language only, published 2000-2014 Search terms
BNI
CINAHL
Embase
Medline
Total combined
‘Brief intervention’ and ‘alcohol’
66
16
1038
621
1741
‘Brief intervention’ and 17 ‘alcohol’ and ‘primary care’
3
241
152
413
‘Brief intervention’ and ‘alcohol’ and ‘general practice’
1
43
22
72
Results from database searching = 72
6
Results from reference list searching = 7
Results from contacting leading academics = 1
Excluded papers = 68 ■■ No primary research ■■ Duplicates ■■ Wrong sample ■■ Pilot study only Papers for further evaluation = 12 Excluded = 1 Wrong intervention
Included = 11
Figure 1. A summary of the search and papers included
Method Four databases (MEDLINE, Embase, British Nursing Index (BNI) and Cumulative Index to Nursing and Allied Health Literature (CINAHL)) were searched for relevant studies. The keywords ‘brief intervention’ and ‘alcohol’ were entered into database search tools, and the search limited to English language studies published between January 2000 and the January 2014. It was deemed that alcohol brief interventions do not generally or regularly go by any other description or facet. To narrow the search to the setting in question, and provide a suitably sized set of papers for screening, ‘general practice’ was introduced as an additional search term rather than ‘primary care’ (Table 1). Figure 1 details the process of finding the final eleven included papers. Inclusions needed to collect and present primary data on ABIs as opposed to offering summaries or reviews of pre-existing evidence. Those looking at interventions within specialist-alcohol treatment, or aimed at populations with mental-health or dependency diagnoses were excluded on the basis that such individuals are best suited to specialist not routine care-based interventions (Babor and Higgins-Biddle, 2001). Fernandez (2007) was included for further evaluation but then excluded on these grounds, as intervention was delivered by alcohol specialist services rather than general primary care. Studies involving
575
British Journal of Nursing.Downloaded from magonlinelibrary.com by 147.188.128.074 on March 22, 2015. For personal use only. No other uses without permission. . All rights reserved.
Results Eleven studies were found to meet the criteria described. Nine were RCTs. Two of these (Aalto et al, 2000; 2001) used an identical research design but had different samples, according to gender and inclusion eligibility. Two studies were qualitative focus-group studies. Table 2 summarises the sample, methodology, findings and strengths and limitations of each. The Critical Appraisal Skills Programme (CASP) tools for quantitative (Public Health Resource Unit (PHRU), 2006) and qualitative studies (CASP International, 2013) were used to assess the validity and relevance of the included studies. Each of the quantitative studies found that patients receiving ABIs reduced alcohol consumption or experienced fewer alcohol-related problems after intervention. Neither of the qualitative studies found any evidence of this. Furthermore, although the RCTs did see reductions in drinking following intervention, only one of the seven RCTs (Curry et al, 2003) found ABI to be more effective than the control of nontreatment or very basic advice. A dominant theme that emerges from the studies is the problem of loss to follow-up. For instance, in Kaner et al’s (2013) study, only 57% of those randomised to extended intervention cooperated with the trial for a long enough period to receive it. The concern is that those likely to drop out between baseline and follow-up are potentially unrepresentative of the overall sample. Edwards and Rollnick (1997) suggest drop-out is most common among those younger than the average and, crucially, less likely to respond to the ABI. Furthermore, in all studies except Butler et al
576
(2013), drop-out tended to be higher within the intervention cohorts than control. The fact that a frequent consequence of the intervention was non-compliance presents as evidence against the effectiveness of ABIs. Some of the stronger studies (Lock et al, 2006; Butler et al, 2013) took care to address the issues of loss to follow-up and its potential effect on outcomes, specifically through using an intention-to-treat (ITT) analysis. Within this approach, every individual who begins treatment is considered part of the trial, whether they go on to finish or not (Hollis and Campbell, 1999), and their outcomes are analysed according to their original grouping, as recommended by the CASP framework (PHRU, 2006). This avoids drop-out breaking the initial randomisation in a study, and the resulting evidence provides comment on how effectively the intervention was delivered as well as its benefits post delivery. Curry et al (2003) make an exception to this approach, using multiple imputation, whereby statistics software assigns multiple plausible outcomes for lost participants, and the presumed results are obtained by combining them. This addressed the problem of differential loss to follow-up. However, it arguably overestimated treatment effect relative to ITT analysis, by allowing some lost subjects to be assigned positive response, despite their ultimate non-engagement. A number of studies failed to provide sufficiently large or representative samples. The CASP framework highlights sample size power as a key factor in minimising the play of chance within studies. Yet, only a small number of studies used power calculations to legitimise the size of their sample (as did the Kaner et al, 2013 study which required 435 patients and had an initial sample of 756), and within those that did, the size required was more often not met. (Lock et al, 2006; Drummond et al, 2009) The qualitative studies were unsurprisingly especially small; Beich et al (2002) had 24 participants and Lock (2004) had 31. As well as size, the quality of recruitment methods is worthy of particular reflection. Notable bias in recruitment procedures was seen in Drummond et al (2009) and in Watson et al (2013), where only males and older participants respectively were recruited owing to their perceived greater response to ABIs. So, it is important to acknowledge that sample quality and follow-up were methodological issues which affected the validity of the 11 included studies although certain compensatory strategies, such as ITT analysis, were often employed. Nevertheless, the commonality of findings across the RCTs in particular (8 out of 9 suggesting ABI was not statistically or qualitatively better than control) does suggest that sufficient rigour was maintained to allow a common theme to emerge. As discussed, inclusion criteria within the search strategy allowed for considerable variation in outcome measures in order to reflect the broad nature of public-health outcomes. The corresponding difficulty is that direct comparisons between study results are not straight-forward. One common measure is units or standards drinks per week, enabling certain like-for-like comparisons. For instance, Lock et al (2006), found a drop of 7 units per week among the intervention arm at 6 weeks, whereas Aalto et al (2001) did not find any reduction in number of drinks over 12 months. However, as argued by Sanchez-Craig (1992), such a reliance
© 2014 MA Healthcare Ltd
an additional health factor, such as pregnancy or injury, were excluded on the grounds of dissimilar sample and reduced transferability of findings. However, studies in which the intervention was conducted by a non-nursing professional, i.e. a doctor or researcher, were not excluded, so long as the intervention did not require skills or specialist training beyond those routinely used by a practice nurse. The search criteria left considerable room for variation between included studies. As mentioned, qualitative and quantitative studies were permitted. Moreover, the wide spectrum of consequences associated with a public-health problem such as alcohol consumption poses difficulty in finding unified outcome measures. For instance, it is known that the physical, psychological, social and financial costs of harmful drinking lead to serious and diverse knock-on effects from impaired liver function to emotional harm, relationship breakdown and social isolation. Patterns of drinking, plus secondary consequences related to drinking, are intended targets of ABIs as well as the number of units consumed. So, objective measurements such as blood tests and more subjective measures such as self-reported wellbeing are justifiably used within relevant literature. Studies with quite different outcome measures, from gamma-glutamyl transferase levels (Aalto et al, 2000) to episodes of drink-driving (Curry et al, 2003) were therefore included in this review. The value of assessing the complete, multi-dimensional impact of a public-health intervention is felt to outweigh the benefits of identical outcome measures.
British Journal of Nursing, 2014, Vol 23, No 11
British Journal of Nursing.Downloaded from magonlinelibrary.com by 147.188.128.074 on March 22, 2015. For personal use only. No other uses without permission. . All rights reserved.
LITERATURE REVIEW Table 2. Summary of literature search studies and findings Author and year
Sample size, country and inclusion criteria
Methodology
Results
Lock et al (2006)
127 adults
Intervention group received a 5–10 minute brief intervention. Control received ‘standard advice’ and a leaflet.
No statistically significant evidence that brief intervention provides superior outcomes to control. Reduced consumption for 51% of ABI participants and 59% of control
North east of England Randomised controlled trial
AUDIT score >8 or >7 for women
Drummond et al (2009)
112 male adults
Randomised controlled trial
AUDIT score >= 8
Butler et al (2013)
1827 adults
Randomised controlled trial
AUDIT-C score >= 4 (men) or 3 (women)
Maisto et al (2001)
301 adults USA
Kaner et al (2013)
Intervention group received stepped care (extended ABI) and control group received minimal intervention (standard ABI)
South Wales
South Wales
Randomised controlled trial
Follow-up at 6 and 12 months. Outcome measures—AUDIT scores, drink units per week and Drinking Problems Index scores.
Follow-up at 6 months post intervention. Primary outcome measures total number of drinks consumed in previous 180 days, number of days abstinent and average drinks/day. Intervention group received behaviour-change counselling from trained GP or practice nurse, control group received screening and consultations from GP/ nurse not trained in behaviour change.
Intervention groups received either brief advice or motivational enhancement (extended intervention), control group received standard care (generalised advice). Follow-up at 1, 3, 6, 9 and 12 months. Outcome measures as number of drinks consumed, and broader drinking behaviour measures.
756 adults
Intervention group received either brief advice + leaflet (basic ABI) or basic life counselling + brief advice + leaflet (extended ABI). Control group received screening and leaflet only.
AUDIT >= 8 Follow up at 6 and 12 months. Outcome measures—AUDIT Score, alcohol problems questionnaire, service use questionnaire, patient satisfaction questionnaire and use of ‘Drink Line’ helpline.
Randomised controlled trial Aalto et al (2000)
118 female adults Finland
Randomised controlled trial Aalto et al (2001)
© 2014 MA Healthcare Ltd
Randomised controlled trial
Those assigned to intervention received either 7 (intervention group A) or 3 (intervention group B) sessions. Control group received advice to reduce drinking upon baseline measurement only.
CAGE test giving >= Two affirmative answers or self-reported alcohol consumption >= 190 g absolute ethanol per week
Follow-up at 3 years post intervention. Main outcome measures self-reported weekly alcohol consumption, bloodtest markers including gamma-glutamyl transferase (GGT) and mean corpuscular volume (MCV).
296 male adults
Methodology matched Aalto et al (2000)
Finland CAGE test giving >= Three affirmative answers or self-reported alcohol consumption >= 280 g absolute ethanol per week
British Journal of Nursing, 2014, Vol 23, No 11
Greater reported intention to change at 3 months from intervention group. Yet no statistically significant difference in reported behaviour at 3 or 12 months between intervention and control.
In-person follow up immediately after consultation plus questionnaire follow up at 3 and 12 months.
AUDIT score >= 8 or quantity- frequency score of >= 16 (male)/ 12 (female) standard drinks per week
Practices from across England and Wales
Basic ABI led to significantly reduced alcohol consumption, as did extended intervention. The slight difference in effect (stepped intervention produced greater results than standard ABI) was not statistically significant.
Alcohol consumption was reduced in both interventions, as well as with the standard care cohort. Found that very minimal intervention is sufficient.
All groups reduced alcohol consumption and smaller reduction in control group was not statistically significant. High drop-out from extended intervention arm implies greater difficulty of implementing counselling relative to brief intervention or screening only.
Depending on outcome variable reduction was between 27% and 75% of total participants. No statistically different outcome between groups A, B or Control regardless of outcome measure chosen.
Greater reduction in intervention groups A and B than in control group but did not reach level of clinical significance (group B = 5–13% greater reduction) Statistically significant reduction in MCV was found across all groups.
577
British Journal of Nursing.Downloaded from magonlinelibrary.com by 147.188.128.074 on March 22, 2015. For personal use only. No other uses without permission. . All rights reserved.
529 adults over 55 years old England and Scotland AUDIT score >= 8
Intervention (‘stepped care’) group received 20-minute behavioural change counselling plus motivational enhancement therapy and additional referrals if indicated. Control group received 5-minute brief intervention. Primary outcome as average drinks per day, followed up at 12 months.
Randomised controlled trial Curry et al (2003)
307 adults USA
Randomised controlled trial Lock (2004)
Four eligibility criteria relating to quantity of drinks consumed and three indicators of at-risk drinking behaviour (relating to chronic drinking, binge drinking and drink driving) 31 adults
Intervention group received advice from physician during initial visit, self-help manual, written feedback and up to three follow-up counselling calls. Control group received no intervention. Follow-up at 3 months and 12 months. Outcome measures quantity of drinks consumed and three indicators of at-risk drinking (matching screening indicators).
Focus group based, using a funnel approach to discussion surrounding attitudes to ABIs
North east of England Combination of random and purposive sampling All drinking statuses included
Beich et al (2002)
24 general practitioners
Qualitative study using focus group and individual interviewing Specific focus on the feasibility of implementing ABIs, as opposed to their theoretical effectiveness.
on quantitative outcomes risks ignoring changes in pattern of consumption. Therefore, inclusion of aspects such as drinkdriving (Curry et al, 2003) or intention to change habits (Butler et al, 2013) allows measurement of how ABIs change overall behaviour and risk. Studies also differed according to timeframe; Aalto et al (2000; 2001) followed up at 36 months, others at 12 or 3. Given that the effect of ABIs on longer term habits is deemed a greater challenge than immediate behaviour changes (Kaner et al, 2007); this may explain the limited impact found in the Aalto studies. In summary, the included studies are somewhat weak in terms of providing uniform outcomes for comparison, but strong in that this allows for multidimensional outcome assessments, going beyond previous evidence bases. The included studies display certain credentials, recognised by the CASP approaches, which reinforce the validity of their findings. An oft-cited weakness of quantitative research, particularly RCTs, is their tendency to be efficacy rather than effectiveness trials, so removed from real-world practice (Victoria et al, 2004). An intervention proven effective in ideal conditions may be considerably less straightforward in routine practice. A majority of the included studies were efficacy trials, with admittedly smaller sample sizes, but also
578
Considerably higher drop-out for stepped care (two thirds did not attend step 2 onwards). Outcomes measured as number of drinks consumed fell across both intervention and control. However reductions were more marked for at-risk patterns of drinking behaviour (specifically drink driving) for intervention group.
Considerable barriers to effective implementation of ABIs. Population-wide education campaigns favoured over one-toone interventions.
Qualitative data and outcome measures—including statements on by whom, whereabouts and to what effect ABIs could or should be delivered.
Denmark Participants all had experience of delivering ABIs as part of a previous study
Both stepped care and control groups reduced drinking levels at 12 months. The difference in amount of reduction between intervention and control was only 0.025, providing no statistically significant difference.
Doctors did not feel they could recommend ABIs as a strategy for reducing risky alcohol consumption. Community-wide preventative methods favoured over ABIs.
greater generalisability than often provided by effectiveness studies. The two qualitative studies were especially valuable for assessing whether ideal-condition findings can be generalised into practice, through gathering experiences of recipients and deliverers of ABIs. The concept that clinician and patient engagement are equally crucial for intervention effectiveness (Hutchings et al, 2006) makes Beich et al’s (2002) study particularly relevant for this review, and again demonstrates the value of including qualitative evidence to enable judgement on this particular intervention.
Discussion As noted, the studies as a whole offered no evidence that brief intervention, or further personalised intervention, yielded more impressive results than alcohol screening and minimal advice only. Figure 2 demonstrates Kaner et al’s (2013) data. Given that ABIs have been part of alcohol strategy for some time (Babor and Higgins-Biddle, 2001) and supported by acclaimed evidence (Kaner et al, 2007), it is important to assess whether the problem is with the validity of studies included in this review, or whether a challenge to the previous evidence base is warranted. Regarding study validity, certain weaknesses have been conceded above. However,
© 2014 MA Healthcare Ltd
Watson et al (2013)
British Journal of Nursing, 2014, Vol 23, No 11
British Journal of Nursing.Downloaded from magonlinelibrary.com by 147.188.128.074 on March 22, 2015. For personal use only. No other uses without permission. . All rights reserved.
LITERATURE REVIEW
© 2014 MA Healthcare Ltd
Conclusion The implications for this review are great; the suggestion being that to screen and to discuss alcohol with a heavy drinker is equally as effective as an ABI, primarily because there is such considerable overlap between the two. It is possible to conclude, then, that the lesser impact of ABIs found in this review than in the Cochrane review (Kaner et al, 2007) could be a result of the advancement of screening and ‘non-intervention’ practices, so that in this review’s later collection of studies, control effect on drinking is considerably greater.Two of the most recent studies, Kaner et al (2013) and Butler et al (2013) support this assumption, having provided
British Journal of Nursing, 2014, Vol 23, No 11
Baseline
6-month follow up
12-month follow up
40 35 Proportion of patients
Sanchez-Craig (1992) argues that commonality of findings across a number of studies serves to reinforce the validity of the outcomes. Strengths of the evidence reviewed, in terms of common findings, combination of qualitative and quantitative evidence, and specific credibility approaches used (including cluster randomisation, and ITT analysis) combine to allow an acceptable level of confidence that lack of validity is not the explanation for the limited intervention effect found. Indeed, Watson et al (2013) state that the absence of benefit of intervention over basic care is so noticeable that issues of methods and validity cannot provide an explanation. Clarifying the nature of the intervention’s evident weakness facilitates more in-depth conclusions. Rather than failing to reduce at-risk drinking behaviour per se, (they succeeded in doing so in 6 out of 9 studies), ABIs only failed to reduce drinking more than control arms. Therefore, discussion must look at the content of the controls themselves. While screening practices varied across studies, they were commonly thorough and multi-layered. The majority including the minimum of an AUDIT score assessment (Table 3), which itself often includes an introductory statement as to the hazards of heavy drinking. When assessment of drinking is done properly, the distinction between screening and outright intervention ‘is inevitably arbitrary’ (SanchezCraig, 1992: 661). AUDIT is arguably in itself a first-stage intervention (WHO, 2006). Follow-up procedures could also be interventionist, when involving re-assessment of drinking patterns and done frequently over a lengthy period of time. Butler et al (2013), for instance, required all participants to report and re-evaluate their drinking behaviours at 0, 3 and 12 months regardless of their being randomised to control or intervention. Maisto et al (2001) did so even more frequently, at 1, 3, 6, 9 and 12 months. Moreover, the control group often included basic support, namely short advice on initial visit and a leaflet, so there is an absence of a strictly non-treatment comparison group. The two Aalto studies (2000; 2001) and Watson et al (2013) affirm this, having explicitly ruled out the prospect of no-advice control as unethical. The argument is that, given long-standing evidence on the harmful effects of heavy drinking, nurses would be demonstrating unethical practice if they were to abstain altogether from commenting on such behaviour. Where outcomes were noticeably better for intervention than control (i.e. Curry et al, 2003) controls were strictly no-advice, separating this study from the others (and raising questions about the different ethical assumptions shaping their design).
30 25 20 15 10 5 0
Patient information leaflet
Brief advice
Brief lifestyle counselling
Intervention Figure 2. Proportion of patients (by intervention type) scoring negative status on alcohol use disorders identification (AUDIT) test (Kaner et al, 2013)
Table 3. Example of an AUDIT screen Alcohol Use Disorders Identification Test (Audit) Please circle the answer that is correct for you. ■■ How
Never
often do you have a drink containing alcohol? Monthly 2 or 4 2 or 3 or less times a month times per week
4 or more times a week
■■ How
many drinks containing alcohol do you have on a typical day when you are drinking? 1 or 2 3 or 4 5 or 6 7 or 9 10 or more ■■ How
often do you have six or more drinks on one occasion? Never Less than Monthly 2 or 3 4 or more monthly times per week times a week ■■ How
often during the last year have you found that you were not able to stop drinking once you had started? Never Less than Monthly 2 or 3 4 or more monthly times per week times a week ■■ How
often during the last year have you failed to do what was normally expected from you because of drinking? Never Less than Monthly 2 or 3 4 or more monthly times per week times a week ■■ How
often during the last year have you needed a first drink in the morning to get yourself going after a heavy drinking session? Never Less than Monthly 2 or 3 4 or more monthly times per week times a week ■■ How
often during the last year have you had a feeling of guilt or remorse after drinking? Never Less than Monthly 2 or 3 4 or more monthly times per week times a week ■■ How
often during the last year have you been unable to remember what happened the night before because you had been drinking? Never Less than Monthly 2 or 3 4 or more monthly times per week times a week ■■ Have
No
you or someone else been injured as a result of drinking? Yes, but not in last year Yes, during the last year
■■ Has
a relative or friend, or a doctor or other health worker, been concerned about your drinking or suggested you cut down? No Yes, but not in last year Yes, during the last year Source: National Institute of Alcohol Abuse and Alcoholism, 2014
579
British Journal of Nursing.Downloaded from magonlinelibrary.com by 147.188.128.074 on March 22, 2015. For personal use only. No other uses without permission. . All rights reserved.
Conflict of interest: none. Aalto M, Saksanen R, Laine P et al (2000) Brief intervention for female heavy drinkers in routine general practice: a 3-year randomized, controlled study. Alcohol Clin Exp Res 24(11): 1680-6 Aalto M, Seppä K, Mattila P et al (2001) Brief intervention for male heavy drinkers in routine general practice: a three-year randomized controlled study. Alcohol Alcohol 36(3): 224-30 Appleby J (2012) Drinking nation: have we had enough? BMJ 344: e2634. doi: 10.1136/bmj.e2634 Babor TF, Higgins-Biddle JC (2001) Brief Intervention For Hazardous and Harmful Drinking. World Health Organization, Geneva. http://tinyurl. com/dynmmmn (accessed 30 May 2014) Ballesteros J, Duffy JC, Querejeta I, Ariño J, González-Pinto A (2004) Efficacy
KEY POINTS n Current
levels of alcohol consumption in the UK impact negatively on the health of the population and add a considerable burden to public finances
n It
is appropriate to locate nursing interventions relating to alcohol within primary care in order to enable a preventative approach, aiming to tackle hazardous drinking before it escalates into critical health and social problems
n Discussing
issues surrounding alcohol consumption with individuals drinking harmful or hazardous amounts does have a beneficial effect on reducing their intake. However, thorough alcohol screening and basic alcohol advice is equally as effective, and more feasible to deliver, than lengthier or personalised interventions
580
of brief interventions for hazardous drinkers in primary care: systematic review and meta-analyses. Alcohol Clin Exp Res 28(4): 608-18 Beich A, Gannik D, Malterud K (2002) Screening and brief intervention for excessive alcohol use: qualitative interview study of the experiences of general practitioners. BMJ 325(7369): 870 Butler CC, Simpson SA, Hood K et al (2013).Training practitioners to deliver opportunistic multiple behaviour change counselling in primary care: a cluster randomised trial. BMJ 346: f1191. doi: http://dx.doi.org/10.1136/ bmj.f1191 Critical Appraisal Skills Programme International (2013) 10 questions to help you make sense of qualitative research. http://tinyurl.com/jwzbudn (accessed 30 May 2014) Curry SJ, Ludman EJ, Grothaus LC, Donovan D, Kim E (2003) A randomized trial of a brief primary-care-based intervention for reducing at-risk drinking practices. Health Psychol 22(2): 156-65 Drummond C, Coulton S, James D et al (2009) Effectiveness and costeffectiveness of a stepped care intervention for alcohol use disorders in primary care: pilot study. Br J Psychiatry 195(5): 448-56 Edwards AG, Rollnick S (1997) Outcome studies of brief alcohol intervention in general practice: the problem of lost subjects. Addiction 92(12): 1699-704 Fernandez J (2007) Client engagement with a specialist nurse-led alcohol service in primary care. Primary Health Care 17(1): 22-24 Health and Social Care Information Centre (2013) Statistics on Alcohol: England, 2013. http://tinyurl.com/pkm3eeg (accessed 30 May 2014) Hollis S, Campbell F (1999) What is meant by intention to treat analysis? Survey of published randomised controlled trials. BMJ 319(7211): 670-4 Hutchings D, Cassidy P, Dallolio E, Pearson P, Heather N, Kaner E (2006) Implementing screening and brief alcohol interventions in primary care: views from both sides of the consultation. Primary Health Care Research and Development 7(3): 221-9. doi: http://dx.doi. org/10.1191/1463423606pc292oa Kaner EF, Beyer F, Dickinson HO et al (2007) Effectiveness of brief alcohol interventions in primary care populations. Cochrane Database Syst Rev (2): CD004148 Kaner E, Bland M, Cassidy P et al (2013) Effectiveness of screening and brief alcohol intervention in primary care (SIPS trial): pragmatic cluster randomised controlled trial. BMJ 346: e8501. doi: 10.1136/bmj.e8501 Katikireddi SV, Higgins M, Bond L, Bonell C, Macintyre S (2011) How evidence based is English public health policy? BMJ 343: d7310. doi: 10.1136/bmj.d7310. Lacey J (2011) Alcohol brief interventions: facilitating effective delivery. Community Practitioner 84(4): 36-8 Lock C (2004) Alcohol and brief intervention in primary health care: what do patients think? Primary Health Care Research and Development 5(2): 162-78. Lock CA, Kaner E, Heather N et al (2006) Effectiveness of nurse-led brief alcohol intervention: a cluster randomized controlled trial. J Adv Nurs 54(4): 426-39 Macintyre S (2011) Good intentions and received wisdom are not good enough: the need for controlled trials in public health. J Epidemiol Community Health 65(7): 564-7. doi: 10.1136/jech.2010.124198. Epub 2010 Maisto SA, Conigliaro J, McNeil M, Kraemer K, Conigliaro RL, Kelley ME (2001) Effects of two types of brief intervention and readiness to change on alcohol use in hazardous drinkers. J Stud Alcohol 62(5): 605-14 McQueen J, Howe TE, Allan L, Mains D, Hardy V (2011) Brief interventions for heavy alcohol users admitted to general hospital wards. Cochrane Database Syst Rev (8): CD005191 doi: 10.1002/14651858.CD005191.pub3 National Institute for Health and Care Excellence (2010) Alcohol-use disorders: preventing harmful drinking. NICE, London. http://tinyurl.com/ctolc2c (accessed 30 May 2014) National Institute on Alcohol Abuse and Alcoholism (2014) Screening Tests. http://tinyurl.com/6w4zplw (accessed 30 May 2014) NHS Inform (2013) Alcohol Misuse, defining a drink problem. http://tinyurl. com/k82jscp (accessed 30 May 2014) Public Health Resource Unit (2006) Critical appraisal skills programme (CASP) making Sense of Evidence. 10 Questions to Help You Make Sense of Randomised Controlled Trials. http://tinyurl.com/8ms2ck9 (accessed 30 May 2014) Sanchez-Craig M (1992) Toward a public health model to preventing alcohol problems. Addiction 89(6): 660–2 Stade BC, Bailey C, Dzendoletas D, Sgro M, Dowswell T, Bennett D (2009) Psychological and/or educational interventions for reducing alcohol consumption in pregnant women and women planning pregnancy. Cochrane Database Syst Rev (2): CD004228. doi: 10.1002/14651858.CD004228.pub2 Victora CG, Habicht JP, Bryce J (2004) Evidence-based public health: moving beyond randomized trials. Am J Public Health 94(3): 400-5 Watson JM, Crosby H, Dale VM et al (2013) AESOPS: a randomised controlled trial of the clinical effectiveness and cost-effectiveness of opportunistic screening and stepped care interventions for older hazardous alcohol users in primary care. Health Technol Assess 17(25): 1-158. doi: 10.3310/hta17250 Whitehead D (2011) Health promotion in nursing: a Derridean discourse analysis. Health Promot Int 26(1): 117-27. doi: 10.1093/heapro/daq073. Epub 2010 World Health Organization (2006) Screening and brief intervention for alcohol problems in primary health care. http://tinyurl.com/ma98wje (accessed 30 May 2014)
© 2014 MA Healthcare Ltd
control arms with considerable non-personalised input, detailed assessments and frequent re-assessments, followed by trial results which do not advocate ABIs over usual care. Consequently, the recommendations of this review are not that nurses should avoid the matter of harmful drinking, given the apparently weak evidence as to the relative impact of ABIs. Rather, the opposite, as screening for alcohol use and raising the issue of drinking within routine practice has proven to lower risky drinking behaviour to a similar degree as structured ABIs, and so this practice should continue. The acceptability of nursing interventions is of course critical to their success. Therefore, the difficulties implementing ABIs or personalised advice, demonstrated by Watson et al (2013) and by the two qualitative studies in the review (Beich et al, 2002 and Lock, 2004), are also noteworthy. Standard screening and simple advice practices should be recommended for their transferability and impact, over lengthier, personalised ABI. To strengthen the theory established within this paper, it would be helpful to see further studies of nursing interventions for harmful levels of drinking. Studies should begin by accepting that a strictly no-treatment control approach would be potentially unethical, given the evidence that discussion or screening itself amounts to an intervention, and an effective one if done thoroughly. Different screening methods should be tested, and with a focus on effectiveness over efficacy trials to reinforce the need for interventions to be transferable to practice. Best practices from previous studies, notably ITT analysis and power calculations, should be copied to ensure maximum possible validity. Also, given the argument for a role for qualitative evidence in public health strategy, it is vital to gather larger and higher-quality qualitative studies on alcohol interventions to supplement the BJN existing RCTs.
British Journal of Nursing, 2014, Vol 23, No 11
British Journal of Nursing.Downloaded from magonlinelibrary.com by 147.188.128.074 on March 22, 2015. For personal use only. No other uses without permission. . All rights reserved.
