Appendix C: The evidence

This appendix lists the evidence statements from four reviews (two effectiveness reviews and two cost-effectiveness reviews) and the economic modelling report provided by the public health collaborating centre (see appendix A). It links them to the relevant recommendations in section 4. (See appendix B for the key to quality assessments.)

The evidence statements are presented here without references – these can be found in the full reviews (see appendix E for details). It also sets out a brief summary of findings from the economic analysis.

The two effectiveness reviews, two cost-effectiveness reviews and economic modelling report are:

  • Effectiveness reviews:

    • Review 1: 'Interventions on control of alcohol price, promotion and availability for prevention of alcohol-use disorders in adults and young people'

    • Review 2: 'Screening and brief interventions for prevention and early identification of alcohol-use disorders in adults and young people'.

  • Economic analysis:

    • Review 3: 'Prevention and early identification of alcohol-use disorders in adults and young people. Macro-level interventions for alcohol-use disorders: cost-effectiveness review'

    • Review 4: 'Prevention and early identification of alcohol-use disorders in adults and young people. Screening and brief interventions: Cost-effectiveness review'

    • Economic modelling report: 'Modelling to assess the effectiveness and cost effectiveness of public health-related strategies and interventions to reduce alcohol attributable harm in England using the Sheffield alcohol policy model version 2.0'.

Evidence statements numbered 1.1 to 3.8 are from review 1. Evidence statements numbered 5.1 to 7.7 are from review 2. Evidence statements numbered e1.1 to e2.3 are from review 3. Evidence statements numbered e5.1 to e6.2 are from review 4. Modelling statements numbered M1 to M50 are from the economic modelling report.

Where a recommendation is not directly taken from the evidence statements, but is inferred from the evidence, this is indicated by IDE (inference derived from the evidence).

Recommendation 1: evidence statements 1.1, 1.2, 1.3, 1.4, 2.27, 2.30, e1.1; modelling statements M12, M21, M22, M23, M24, M26, M27, M29, M34, M35, M36, M37

Recommendation 2: evidence statements 2.19, 2.20, 2.21, 2.22, 2.24, 2.25, e.2.3; modelling statements M51, M55, IDE

Recommendation 3: evidence statements 3.1, 3.2, 3.3, 3.4, 3.6, 3.7, 3.8

Recommendation 4: evidence statements 2.4, 2.5, 2.8, 2.9, 2.19, 2.20, 2.21, 2.22, 2.24, 2.25

Recommendation 5: evidence statements 7.1, 7.2, 7.6

Recommendation 6: IDE

Recommendation 7: evidence statements 5.7, 5.9; IDE

Recommendation 8: evidence statements 5.1, 5.2, 5.5, 5.6, 5.7, 5.9, 5.10, 5.11, 7.3, 7.4, 7.5, 7.7, e5.1; modelling statements M2, M3

Recommendation 9: evidence statements 6.1, 6.2, 6.3, 6.4, 6.10, 7.3, e6.1, e6.2; modelling statement M6; IDE

Recommendation 10: evidence statement 6.11; modelling statement M6

Recommendation 11: IDE

Evidence statements

Please note that the wording of some evidence statements has been altered slightly from those in the review team's report to make them more consistent with each other and NICE's standard house style.

Evidence statement 1.1

A comprehensive systematic review was identified that demonstrated a clear association between price/tax increases and reductions in consumer demand for alcohol (++).These conclusions were based on two rigorous meta-analyses of price elasticities. Further evidence was supportive of a negative relationship between the price of alcohol and alcohol consumption among young people (one UK and one USA [not graded]). A positive relationship between alcohol affordability and alcohol consumption operating across the European Union was identified (one EU).

Evidence statement 1.2

A systematic review reported that there is some evidence that young people, binge drinkers and harmful drinkers tend to show a preference for cheaper drinks (++).

Evidence statement 1.3

A limited evidence base was identified that indicated that minimum pricing may be effective in reducing alcohol consumption (one [++] and one UK [not graded]). Consulted members of the community were supportive of such measures (one [++]).

Evidence statement 1.4

An evidence base comprising a large number of primary studies was identified that demonstrated a relationship between price/tax increases and reductions in harms (one [++] systematic review). Additional evidence indicates that decreases in the price of alcohol contribute towards increases in alcohol-related deaths, particularly in deaths attributable to chronic causes such as alcoholic liver disease (one [++] Finland). Population groups specifically affected included the older population, the unemployed and individuals with lower levels of education, social class and income (one [++] Finland). However, the same authors observed no increase in interpersonal violence rates following the decrease in alcohol prices (one [++] Finland). A time series analysis demonstrated that increases in tax were associated with decreases in alcohol-related disease mortality (one [++] USA).

Evidence statement 2.4

Evidence was identified demonstrating that serving staff in alcohol outlets were disapproving of under-age sales (one [+] USA) and generally positive of implementing under-age checks, including electronic age-verification devices (one [++] USA).

Evidence statement 2.5

The commitment of managers and licensees towards their legal responsibilities relating to under-age sales was variable (one [+] UK and one [+] USA).

Evidence statement 2.8

The effectiveness of enforcement checks in reducing alcohol sales to under-age young people was variable (one [+] systematic review).Compliance checks conducted by local police were not effective in reducing arrests in those aged under 18 years or reducing under-age sales (one [+] and one [++]) in the UK. Other studies showed favourable outcomes of compliance checks by local authorities in reducing under-age alcohol sales (two [+] USA, one [++] USA and one USA [not graded]). Checks enforced with a 30-day licence suspension or a fine were effective in reducing sales (one [+] USA). However, the deterrent effect of enforcement was found to decay over time (one [+] USA and one USA [not graded]). Additional UK-specific evidence demonstrated that enforcement of laws relating to under-age sales supported by a local multi-agency community alcohol partnership, helped reduce possession of alcohol and antisocial behaviour and improved the relationship between enforcers and retailers.

Evidence statement 2.9

A study based in Fife, Scotland indicated that on- and off-licensees perceived the most effective approach to preventing under-age sales to be test purchasing carried out in conjunction with a new, nationally-accepted proof-of-age card.

Evidence statement 2.19

Other UK-specific studies of the effects of changes in licensing hours presented mixed findings, with some studies reporting no apparent effects on alcohol-related outcomes (two [++] UK). However, following the extension of licensing hours, one (+) UK study reported an increase in admissions for self-poisoning by overdose in which alcohol was also involved. Another UK study found increases in the occurrence of slight accidents in the workplace.

Evidence statement 2.20

Extensions in trading hours in Australia were typically associated with increased violence (one [++]), motor vehicle crash rates (one [++]) and an increase in the apprehension of impaired male drivers aged 18 to 25 years (one [++]). Local community restrictions on alcohol availability were found to have modestly favourable outcomes, including reductions in alcohol consumption and violence. However, in one evaluation of the restriction of take-away trading hours and volumes for alcohol sales in Australia, many customers shifted their purchases to cheap cask port, providing an illustration of the ways in which consumers may respond to limitations in alcohol availability.

An increase in alcohol-related road traffic accidents followed the removal of the ban on Sunday sales of packaged alcohol in New Mexico (one USA [not graded]).

The introduction of unrestricted serving hours in Reykjavik, Iceland resulted in increased police work episodes, more emergency ward admissions for weekend nights, increased suspected drink-driving incidents, and more people circulating in the city centre at 6am (one [+]).

The Saturday opening of alcohol retail outlets in Sweden also led to an increase in sales (two ++) but no apparent change in alcohol-related harms (one [++]).

A range of evidence from Scandinavia, based on largely small-scale, local natural experiments, showed the variable impact of changes in alcohol licensing, with decreased alcohol consumption typically observed as a result of restrictions. However, a USA-based study suggested that restrictions on Sunday alcohol sales had no apparent impact on consumption, whilst earlier closing hours in bars appeared to result in increased alcohol sales.

Evidence statement 2.21

A clear positive relationship between increased outlet density and alcohol consumption among adults was demonstrated in a range of association studies (three USA [not graded], one [++] USA, two [++] Canada and one Canada [not graded]). However, one USA study (not graded) found no significant association between alcohol outlet density and heavy drinking.

Evidence statement 2.22

A positive relationship between alcohol outlet density and alcohol consumption was also observed in studies focusing on young people (one USA, one Australia, two Switzerland and two New Zealand [not graded]).

Evidence statement 2.24

A number of natural experiments demonstrated the effects of changes in alcohol outlet density on alcohol consumption and alcohol-related outcomes. Increases in alcohol outlet density tended to be associated with increases in alcohol consumption and alcohol-related morbidity and mortality in Scandinavia. A literature review found that the privatisation of alcohol retail monopolies in the USA, Canada and Scandinavia (not graded) was linked with higher outlet densities, longer hours or more days of sale and changes in price and promotion, typically resulting in increased alcohol consumption (international). A positive association between alcohol outlet density and gonorrhoea (one USA [not graded]) was also observed following the civil unrest in Los Angeles.

Evidence statement 2.25

An evidence base, within one literature review, was described demonstrating positive relationships between outlet density and a range of outcomes including rates of violence, drink-driving, pedestrian injury, and child maltreatment.

Evidence statement 2.27

Evidence was identified that pre-drinking [drinking before going out] is a prevalent activity, both in the UK (one [++] UK and one UK [not graded]) and within one international literature review.

Evidence statement 2.30

Evidence was identified that demonstrated that pre-drinking is associated with heavy alcohol consumption (one [++] UK and one international [not graded])and increased risk of alcohol-related harm (one [++] UK).

Evidence statement 3.1

One systematic review (++) demonstrated a small but consistent relationship between advertising and alcohol consumption at a population level.

Evidence statement 3.2

A systematic review of longitudinal studies found that exposure to alcohol advertising and promotion was associated with the onset of adolescent alcohol consumption and with increased consumption among adolescents who were already drinking at baseline assessment (++).Another systematic reviewpresented evidence of a small but consistent relationship between advertising and alcohol consumption among young people at an individual level (++).Another review concluded that the evidence base suggested the existence of an association between exposure to alcohol advertising and promotion and alcohol consumption among young people (++). Further literature reviews were also indicative of alcohol advertising having an impact on young people. There was evidence of awareness, familiarity and appreciation of alcohol advertisements among this age group.

Evidence statement 3.3

One systematic review presented evidence of a moderate but consistent association between point of purchase promotions and effects on alcohol consumption among under-age drinkers, binge drinkers and regular drinkers (++).

Evidence statement 3.4

A systematic review reported that outdoor and print advertising media may increase the probability of onset of adolescent alcohol consumption and also influence quantity and frequency of alcohol consumption among young people (++). Another review included one USA-based study that reported that outdoor advertising media did not have any effect on alcohol behaviour, but was a predictor of intention to use alcohol among adolescents (++).

Evidence statement 3.6

One systematic review reported that evidence from longitudinal studies consistently demonstrated that exposure to television and other broadcast media was linked with the onset of and levels of alcohol consumption (++). Further evidence was included in a review thatindicated that exposure to alcohol portrayals via television (including advertisements aired during sports programmes) and other broadcast media may be linked with alcohol use among adolescents (++).

Evidence statement 3.7

The content of alcohol advertising was reported to be attractive to young people, conveying desirable lifestyles and images of alcohol consumption. Younger age groups and girls aged15 to 17 years were reported to be potentially experiencing the greatest impact of alcohol advertising (++).A further UK-specific report showed that, despite changes to the Advertising Code, while advertising recall fell (potentially due to reduced television advertising expenditure over the study period), there was an increased perception among young people that television alcohol advertisements were appealing and would encourage people to drink. However, there was a decrease in the proportion of young people who considered alcohol commercials to be aimed at them. A literature review stated that there was no scientific evidence available to describe the effectiveness of self-regulation in alcohol advertising.

Evidence statement 3.8

Inconclusive evidence was identified, within one systematic review (++) and one literature review (not graded), of the impact of advertising bans on alcohol consumption .

Evidence statement 5.1

The Alcohol-use disorders identification test (AUDIT) is effective in the identification of hazardous and harmful drinking in adults in primary care (three [++] systematic reviews, one [++] Finland, one [++] UK and one literature review [not graded]). The use of lower thresholds in conjunction with alcohol screening questionnaires was recommended for women (one [++] Finland, one [++] Belgium, one [++] systematic review and one literature review [not graded]). Optimal screening thresholds for the detection of hazardous or harmful drinking using AUDIT appeared to be greater than or equal to seven or eight among men (two [++] systematic reviews) and greater than or equal to six to eight among women (one [++] systematic review, one [++] Finland and one literature review [not graded]).Optimal screening thresholds for identifying binge drinking using AUDIT were greater than or equal to seven or eight for adult males (no data available for females) (one [++] Finland).Primary studies included in a systematic review (++) recommended higher AUDIT thresholds for males (five to eight) than females (two to six).

Evidence statement 5.2

The evidence for the effectiveness of shorter versions of AUDIT in adults in primary care was variable. Some authors of cross-sectional diagnostic evaluations observed comparable performance between the full AUDIT and shorter versions (two [++] Finland, one [++] Belgium and one [++] USA). Other findings drawn from primary care were more cautious of the utility of the shorter forms of this questionnaire (one [++] systematic review).The optimal screening threshold for the detection of hazardous drinking using AUDIT-C was greater than or equal to three among men and women (one [++] systematic review and one[++] USA).However, thresholds of greater than or equal to five for the detection of heavy drinking among femalesand greater than or equal to six for identifying bingeing moderate and heavy drinking men were also recommended (one [++] Finland). Primary studies included in a systematic review recommended higher AUDIT-C thresholds for males (three to six) than females (two to five) (one [++]). FAST was described, within a literature review (not graded), as being effective in the detection of alcohol problems at a cut-off point of greater than or equal to one in males and females in a primary care setting in the UK.

Evidence statement 5.5

Only a limited amount of evidence could be identified relating to the performance of alcohol screening questionnaires in hospital settings. The 'Five-shot questionnaire' was shown to detect alcohol misuse in adult male inpatients at a cut-off of greater than or equal to 2.5 (one [++] Belgium). AUDIT was effective in screening UK male and female adult general medical admissions for hazardous and harmful alcohol consumption (one [+] UK). AUDIT was also reported to perform effectively among general hospital inpatients (one [++] systematic review).

Evidence statement 5.6

Evidence was identified for the use of alcohol screening questionnaires among adults in emergency care settings. One study found that the CAGE questionnaire was effective in screening for a lifetime diagnosis of alcohol dependencein trauma centre patients ([++] USA). AUDIT-C was shown to effectively identify hazardous drinking among male and female adult traffic casualties in an emergency department (one [+] Spain). One literature review indicated that FAST displayed good screening properties in the identification of alcohol problems among males and females presenting to an A&E setting in the UK. The 'Paddington alcohol test' has been shown to be rapid, feasible to use, be UK-specific and to have reasonably good screening properties for the detection of alcohol misuse when implemented in response to clinical 'trigger' conditions in A&E care. These are listed as follows: fall; collapse; head injury; assault; accident; unwell; non-specific gastrointestinal conditions; psychiatric; cardiac; repeat attender (three [++] UK).

Evidence statement 5.7

AUDIT was shown to perform more effectively in the identification of alcohol abuse or dependence (when used at a cut-off of greater than or equal to 10) than CAGE, CRAFFT (car, relax, alone, forget, friends, trouble) or RAPS-QF (rapid alcohol problems screen) questionnaires among young people (median age of 19 years) (one [++] USA). AUDIT was also demonstrated to have higher sensitivity (when used at an optimal cut-off of greater than or equal to three) than CAGE, CRAFFT or POSIT (problem oriented screening instrument for teenagers) in the detection of problem use (that is, hazardous or harmful consumption not reaching the diagnostic threshold for an alcohol-related disorder, abuse and dependence) in a sample aged between 14 and 18 years (one [++] USA). The identified evidence for the effectiveness of SASSI (substance abuse subtle screening inventory) in screening for alcohol misuse was limited and inconclusive (two [++] USA and one [+] USA). AUDIT was found to perform reasonably well in elderly populations (one [++] systematic review), while AUDIT-5 was described as showing potential as an appropriate tool for use among older people (one [+] systematic review).

Evidence statement 5.9

The screening properties of questionnaires were influenced by the ethnicity of recipients and authors suggested that the use of appropriate cut-off scores should be considered (one [++ systematic review, one [++] USA and one literature review [not graded]).

Evidence statement 5.10

Laboratory markers are of limited value in the detection of alcohol misuse when compared with alcohol screening questionnaires (two [++] UK, one [++] Belgium and one [+] Germany). However, the use of blood-alcohol concentration testing may complement the use of later questionnaire screening in the identification of alcohol misuse among patients treated in the emergency department resuscitation room (one [++] UK).

Evidence statement 5.11

A number of clinical indicators were described, within a cross-sectional study, a literature review and a case study, as being associated with excessive alcohol consumption (one [++] Spain, one literature reivew and one UK [not graded]). Awareness of such indicators may be useful in alerting health professionals to alcohol-related physical problems.

Evidence statement 6.1

Twenty seven systematic reviews provided a considerable body of evidence supportive of the effectiveness of brief interventions for alcohol misuse. Brief interventions were found to reduce alcohol consumption, alcohol-related mortality, morbidity, injuries, social consequences and the consequent use of healthcare resources and laboratory indicators of alcohol misuse.

Evidence statement 6.2

Six systematic reviews (all [++]) demonstrated that interventions delivered in primary care are effective in reducing alcohol-related negative outcomes. Three systematic reviews specifically focusing on the use of brief interventions in emergency care(one [+] and two [++]) found limited evidence of effectiveness. A further review (++) presented inconclusive evidence of the effectiveness of brief interventions in inpatient and outpatient settings. A systematic review of brief interventions for alcohol misuse in the workplace presented limited and inconclusive findings for the effectiveness of interventions in this setting (++).

Evidence statement 6.3

Brief interventions are effective in reducing alcohol consumption in both men and women (seven [++]).

Evidence statement 6.4

Most of the primary evidence was drawn from populations with an age range of 12 to 70 years. Therefore, brief interventions for adults have been shown to be effective among adult populations.

Evidence statement 6.10

Extensive heterogeneity was evident in the characteristics of evaluated brief interventions. However, limited evidence would suggest that even very brief interventions may be effective in reducing alcohol-related negative outcomes, (one [++] systematic review) with inconclusive evidence for an additional positive impact resulting from increased dose (three [++] systematic reviews). Evidence from an additional review (++) suggests that brief interventions are effectivebut the impact of including motivational interviewing principles was unclear.

Evidence statement 6.11

Extended brief interventions were demonstrated to be effective in the reduction of alcohol consumption (evaluated interventions consisted of two to seven sessions with a duration of initial and booster sessions of 15 to 50 minutes (one [++] systematic review) or 10 to 15 minutes in one session with a number of specific booster sessions of 10 to 15 minutes duration (one [++] systematic review).

Evidence statement 7.1

Organisational factors such as adequate support and resources can influence the acceptability and implementation of screening and brief intervention for alcohol misuse.

Implementation of screening and brief interventions is influenced by factors other than effectiveness. Positive support from the government, management and involvement of non-clinical members of staff are more likely to result in successful implementation.

There is also evidence from a range of studies in primary care settings that adequate practitioner training and support in alcohol misuse screening and use of brief intervention materials facilitates – or would facilitate – effective implementation rates and appropriate detection of 'at risk' drinkers. Evidence suggests that the extent of training and support available to practitioners is variable.

One RCT ([++] USA) showed more successful implementation of screening and brief intervention where there was prior experience of this type of work, management stability and positive support in terms of coordination of programmes. Financial incentives and successful management of staff changes, as well as assistance from receptionists, were also important. However, barriers to success included competing priorities and lack of time. The importance of financial and other incentives for GPs along with readily available materials and training was also highlighted in one survey in New Zealand (+).

Evidence from RCTs (one [++] USA, one [+] USA and one [+] UK) suggests that the extent to which brief interventions are implemented, though not necessarily the appropriateness of implementation, is increased with use of a training and support intervention for GPs and nurses. One cross-sectional study ([++] Germany) provides evidence that GPs holding a qualification in addiction medicine are more likely to detect problem drinkers. However, a cross-national survey (++) found that training did not improve baseline role insecurity for GPs.

One cross-sectional study([+] Finland) and one qualitative study ([++] Finland) found that practitioner training rates and ratings of their own familiarity with screening tools and knowledge of brief intervention content was low. The importance of training to practitioners in this survey was evident, as were practitioner views that they lacked training to carry out counselling ([++] UK). The latter point was also evidenced in one cross-sectional study ([++] UK). A Delphi survey ([++] UK) provides evidence in the form of expert opinion that practitioner training should help raise awareness of risk factors and typical presentations of individuals with potential drinking problems.

Evidence from qualitative studies show that some nurses in the UK (one [++]) see training as an incentive to carrying out alcohol-related work.

A sample of GPs in Finland perceived that they lacked training in identifying the early stages of alcohol misuse; and GPs in a Danish focus group study (+) felt they lacked training in counselling skills.

In a probationary setting, forensic medical examiners in a UK qualitative study set in custody suites (-) felt they lacked the required training to carry out assessments of drinking behaviour.

Evidence statement 7.2

Extending the current practitioner workload is a potential barrier to implementing screening and brief interventions on a large scale, particularly if all young people and adults are screened as routine practice. The extra time that implementation demands can be a barrier to acceptability and therefore a willingness to deliver such a programme.

Implementation of routine screening and brief intervention programmes requires team-working between physicians, nurses and non-clinical personnel, with consideration required regarding the extent of involvement and specific roles of team members.

Evidence from one systematic review ([++] Denmark) challenges the model of universal screening. The study concluded that implementation of universal screening does not benefit sufficient numbers of individuals to warrant the extra workload required. Nurses in one qualitative study ([++] UK) felt 'overloaded' with preventative work generally, with resources such as space, staff and sufficient time in short supply. In another qualitative study ([+] Denmark), the additional workload of screening and brief interventions was found to be creating stress among practitioners in primary care. In terms of time available, a Canadian qualitative study (++) found that time was constrained in terms of assessing each patient.

A qualitative study of Finnish GPs(++) showed that they felt they lacked time to carry out a drinking assessment in the context of other consultation demands and weak evidence. One (-) study in Sweden found that nurses regarded time constraints as a barrier to engaging in alcohol prevention. There is mixed evidence from one RCT ([++] USA) for the utilisation of non-clinical staff in implementation in order to delegate work and thus to decrease the workload of clinicians. Another RCT found that receptionists did not have a particularly positive attitude to being involved in this type of work without adequate reimbursement ([++] UK), or to changing their perceived role ([++] USA).

In an emergency care setting, one cross-sectional study ([-] USA) provides weak evidence (from a survey of physicians) that, despite support for brief interventions in theory, lack of time is a barrier to implementation. A further UK-based study set in an emergency department also reported that lack of time was viewed as a limiting factor in delivering screening (++).

In a briefly reported UK qualitative study set in custody suites (-), forensic medical examiners felt they lacked the required time to carry out assessments of drinking behaviour.

Evidence statement 7.3

There is evidence that implementation of screening and brief interventions would be facilitated by use of environments where alcohol can be discussed in a non-threatening way. Integrating screening and advice into general lifestyle discussions might increase the acceptability of screening and brief intervention for users. In a range of studies, providers and experts emphasise the importance of appropriate contexts for discussion of alcohol use with users in order to increase acceptability.

Clinical consultations for non-alcohol-related medical problems can be an inappropriate time to discuss alcohol use, given that users are focused on the condition for which they are seeking advice. Instead, sessions such as new patient registrations and well-person clinics, where health promotion is often discussed, provide a less threatening opportunity to discuss drinking, as part of a general discussion on lifestyle issues such as diet, exercise and smoking.

Evidence was found from a cross-sectional study ([+] Sweden) that primary care users attending for scheduled appointments are more likely to be asked about their drinking behaviour. This suggests that practitioners deem certain contexts as more appropriate or more convenient in some way for carrying out screening and a brief intervention. A Delphi survey ([++] UK) also provides expert-view evidence that clinics and new registration sessions are an appropriate context for assessing drinking behaviour (in terms of user acceptability). This study also suggests that interventions might be more acceptable to users if they are tailormade to the individual rather than global in design. There is further evidence from five UK qualitative studies (four [++] and one [+]) that practitioners and users regard clinics, registration sessions and routine consultations as opportunities for discussions in a less-threatening environment and context. That is, they provide an opportunity to discuss drinking in a context that is related to the purpose of the visit (such as lifestyle assessment or chronic condition monitoring).

Emergency care and probation settings are regarded as two contexts that provide a potential opportunity to carry out alcohol screening and give advice. However, there is scarce evidence available.

One survey of Scottish emergency care units (++) and one qualitative study ([-] UK) set in custody suites found that staff thought the location unsuitable for alcohol screening and intervention. However, two surveys from the US (both [+]) reported that both patients and surgeons found the emergency care setting acceptable and appropriate. One US evaluation (+) provided evidence that emergency care staff may not feel adequately supported either by management or financially, with training and workload as two particular concerns. One UK survey (+) provided mixed views, with some nurses preferring an holistic approach, and others prioritising care of injuries over health promotion. A further UK-based (++) study found that the majority of consulted professionals judged the emergency department to be an appropriate place to perform alcohol screening but that implementation rates were low, potentially due to clinical inertia.

The importance of having resources in place to rapidly refer positively screened patients from the emergency department for a brief intervention was emphasised, because the rate of attendance for brief interventions dropped off markedly 2 days following referral (one [++] UK).

Implementation of alcohol screening and brief interventions in emergency care settings is not as consistent as in primary care. The setting differs from primary care in terms of patient population and types of presenting cases, and as such, account needs to be taken of barriers and facilitators to implementation that are specific to the emergency care context, where attendance is brief and often traumatic, patients are more likely to be injured, traumatised, or intoxicated, and staff may feel less prepared to give advice.

Evidence statement 7.4

There is evidence that service users have preferences regarding the status of the person dealing with their alcohol issues. Although experts consider alcohol and counselling specialists to be better qualified to carry out interventions, service users might feel stigmatised or rejected should their needs be referred on to such practitioners.

Evidence from one RCT ([+] USA) carried out in a general medicine setting showed that service users are no more likely to attend for counselling with an alcohol specialist than with a physician or nurse. In addition, qualitative evidence from the UK (one [++]) focusing on user views shows that counselling with alcohol specialists can sometimes be perceived as stigmatising. These views contrast with expert views (one [++] UK) that alcohol workers and counsellors might be best placed to deliver a brief intervention. There are mixed views from three UK studies (all [++]) in that professionals and some users perceive the nurse as having more time for discussing drinking with users, whereas other users report that they are more likely to discuss alcohol-related issues with their GP.

Evidence statement 7.5

There is some evidence that service users are generally positive about screening and intervention. There is also evidence for general under-activity in discussing drinking with service users.

Negative service user behaviour, such as aggression at being asked about their drinking, while rare, may serve as deterrents to practitioners approaching the topic of drinking with users. Actual drunkenness at consultations limits the likelihood that users will appreciate or remember the advice given. Practitioners may benefit from training in dealing with such situations, and in approaching the topic with individuals that they perceive as 'low risk' in appropriate contexts.

Two studies (one [+] USA and one [++] UK) provide evidence that the majority of service users are positive about screening, and another ([+] Finland) that they are positive about discussing drinking. However, two qualitative studies (one [++] UK and one [+] Denmark) found that some professionals had encountered negative reactions from users in terms of embarrassment and unease, and that this led some to change their GP practice.

Evidence from two UK cross-sectional studies (both ++) shows under-activity in terms of practitioner management of hazardous drinking, with a majority of GPs in the first study only intervening in between one and six cases of hazardous drinking per year. Even in cases of heavy drinking, service users are not being asked about their consumption ([+] Finland). Another cross-sectional study ([+] Sweden) found that advice on drinking behaviour is provided less often than for other lifestyle behaviours, such as exercise, diet, and smoking, and less often than service users expect. One cross-sectional study ([++] Finland) found that the time being spent on asking users about their drinking was typically less than 4 minutes, and another recent cross-sectional study ([+] Germany) found that detection rates of problem drinkers are low, at one in three. Possible reasons are found in a Finnish qualitative study (++) of GPs, who reported their reluctance to ask users about their drinking unless they saw clear signs of risky drinking behaviour.

Evidence statement 7.6

Evidence was found that provider attitudes, knowledge, skills and behaviour can influence the implementation of screening and brief interventions for alcohol misuse.

There is evidence from primary care practitioner views of a shortfall in perceived knowledge in terms of detecting 'at-risk' individuals. There is also evidence of confusion regarding current guidelines around drinking behaviour, and the known benefits of drinking in moderation. This can affect practitioner confidence in, and motivation towards, implementing screening and brief intervention programmes effectively. In addition, the practitioner's own drinking behaviour and the user-practitioner relationship may affect the way that alcohol-related interventions are implemented.

One UK qualitative study (++) provides evidence that GPs found difficulty in identifying early-stage heavy drinkers. The study also reports difficulty working with multiple definitions of problematic drinking. One qualitative study ([+] Finland) found that GPs and nurses saw the lack of clear guidance as a barrier to carrying out brief interventions. Utilising the skills of receptionists can be useful, but there is evidence from one RCT ([++] UK) that receptionist attitudes toward the work may be less positive than that of clinicians, and that this might have an impact upon implementation.

There is weak evidence ([-] UK) that forensic medical examiners perceive that they lack the knowledge to carry out an assessment in custody suites in the UK.

Two UK qualitative studies (one [++] and one [+]) found that nurses saw alcohol as a difficult and emotive topic to broach with users. In addition, nurses reported confusion for themselves and service users around the issue of standard drink units, and the potential benefits of alcohol that create ambiguity in discussing drinking from a health promotion perspective. Other studies (one [+] UK and one [+] Finland) found that GPs relationship with alcohol could affect their behaviour in terms of addressing service user drinking, with feelings of guilt and hypocrisy potential barriers to open discussion, or facilitators to empathy. There is qualitative evidence from three studies focusing on user views (two [++] UK and one [+] USA) that discussing drinking is facilitated by a good relationship with the health professional. In addition, there is evidence (one [+] Denmark) that practitioners are concerned not to offend users by discussing alcohol for fear of disturbing the therapeutic relationship.

Evidence statement 7.7

Evidence was identified that shows disparities in the way screening and brief interventions for alcohol misuse are implemented in realtion to certain groups within the population. While certain groups, such as males and high earners, are more 'at-risk' than others from alcohol misuse, individuals from groups that are 'low-risk' – such as females, younger and older people – may be missed. Conversely, over-targeting can also occur due to misconceptions of the populations most at-risk of alcohol misuse.

One systematic review (+) provides inconclusive evidence that socioeconomic status affects the uptake of brief interventions. However, one cross-sectional study ([++] UK) found that unemployed individuals were more likely to receive a brief intervention than those in employment. In terms of ethnicity, there is evidence from one cross-sectional study ([+] USA) that minorty ethnic groups, in this instance black and Hispanic, and particularly Hispanic people, were more likely to be approached by practitioners regarding their alcohol consumption.

Four cross-sectional studies (one [++] UK, one [+] Sweden, one [++] Germany and one [+] Finland) provide evidence that primary care users most likely to be given advice on drinking are males. Another cross-sectional study ([+] Finland) suggests that males, as well as heavy drinkers, are also more likely to adhere to the advice provided in a brief intervention. One qualitative study ([+] Denmark) found that GPs were reluctant to address drinking with young people as they felt that they would be likely to grow out of the habit of hazardous drinking.

Evidence statement e1.1

There is limited evidence of the cost effectiveness of price controls in a UK setting. One systematic review (+) suggests that the available evidence is limited to two studies, one which takes an international perspective, and one set in Estonia. The review reports that the evidence is suggestive that in areas with a high prevalence (greater than 5%) of hazardous drinkers, as is the case in the UK, taxation will be more cost effective than other alcohol misuse macro interventions, but that the evidence base for this is not strong.

Evidence statement e2.3

There is limited evidence of the cost effectiveness of opening hours interventions in a UK setting. One study of moderate quality that takes an international perspective (+) provides evidence that reducing licensing hours provides relatively small quality of life benefits compared to other alcohol misuse interventions.

Evidence statement e5.1

One study shows that the AUDIT test is a more cost effective screening tool than measures of y-glutamyltransferase, aspartate aminotransferase, per cent carbohydrate deficient transferrin, and ethrocyte mean cell volume. This is because AUDIT is both cheaper and more effective than these other tests ([+] UK). The evidence does not allow a ranking of the cost effectiveness of these other screening methods.

Evidence statement e6.1

Cost effectiveness evidence for screening and brief interventions in the emergency care setting is scarce. The available evidence does not allow firm conclusions regarding the long-term cost effectiveness of these interventions in a UK setting. However, the evidence does suggest that brief interventions in the emergency care setting may be cost effective in the UK. One study suggests that screening plus a brief intervention may produce long-term cost savings ([+] USA), but the applicability of this evidence to the UK is uncertain. One UK study suggests that a brief intervention administered by alcohol health workers in a hospital setting will reduce consumption in the short term without significantly increasing costs, but long-term evidence is lacking (++).

Evidence statement e6.2

Cost effectiveness evidence for screening and brief interventions in the hospital setting is scarce. The available evidence does not allow conclusions regarding the cost effectiveness of these interventions in a UK setting to be made. A UK study presents evidence for screening by doctors and nurses in a general hospital setting (+), but this does not allow a conclusion to be reached regarding the most cost-effective screening method. One study suggests that screening plus a brief intervention may produce long-term cost savings ([-] Australia), but the reliability of this evidence is low due to the uncertainty in resource use estimates.

Modelling statement M2

A policy of screening and brief intervention at next GP registration is a more phased approach over time than screening at next GP consultation. The former approach would screen an estimated 39% of the population, with 36% of hazardous and harmful drinkers receiving a brief intervention over the modelled 10-year screening programme. A policy of screening and brief intervention at next GP consultation is a very large-scale implementation, with an estimated 96% of the population screened after 10 years (of whom the majority would be screened in the first year of implementation), and 79% of hazardous and harmful drinkers receiving a brief intervention.

Modelling statement M3

Screening and brief intervention in an A&E setting is estimated to screen 78% of the population within 10 years, but because the estimated uptake of brief interventions is just 30%, only 18% of hazardous and harmful drinkers are estimated to receive the brief intervention.

Modelling statement M6

Sensitivity analysis shows that even fairly long brief interventions (for example, 25 minutes) would appear cost effective versus a 'do nothing' policy. There is currently no conclusive evidence of the differential effectiveness of delivery by different types of staff. On this basis, decision makers might consider the less costly staffing options that were modelled for screening and the brief intervention to be attractive. Evidence around the differential effectiveness of interventions of different duration is also inconclusive. Sensitivity analyses show that shorter duration interventions remain cost effective when using the best available evidence on the relationship between duration and effectiveness.

Modelling statement M12

Increasing levels of minimum pricing show very steep increases in effectiveness. Overall changes in consumption for 20p, 25p, 30p, 35p, 40p, 45p, 50p, 60p, 70p are: --0.0%, -0.1%, -0.4%, -1.1%, -2.4%, -4.3%, -6.7%,
-11.9% and -17.7%. Higher minimum prices reduce switching effects. Note that estimates for lower minimum prices are subject to less modelling uncertainty than those for higher minimum prices. This is because the consideration of supply-side responses and, in particular, a possible restructuring of the market following large mandated price increases in sections of the market, was outside the scope of the model. As an example, a minimum price of 40p per unit has the following estimated effects:

% change in consumption

Deaths p.a. (full effect )

Hospital admissions p.a.

Crimes pa

Work absences (days p.a.)

Un-employment (persons p.a.)

-2.4%

-1,190

-39,000

-10,000

-134,000

-11,500

Modelling statement M21

As prices increase, alcohol-attributable hospital admissions and deaths are estimated to reduce. Prevented deaths occur disproportionately in harmful drinkers. On balance, the health-harm reductions mostly relate to chronic diseases rather than acute conditions such as injuries. This is because much of the alcohol-attributable health harm occurs in middle or older age groups at significant risk of developing and potentially dying from chronic disease.

Modelling statement M22

For chronic diseases, the time for a change in consumption to achieve the full effect in changing the prevalence of disease is important in the modelling. The reductions in health-harms, for chronic disease, observed 1-year following implementation are estimated to be around one tenth of the level that will accrue when the full effect of consumption changes occurs.

Modelling statement M23

Crime harms are estimated to reduce as prices are increased. The crime reductions observed for policies take place across the spectrum of violent crime, criminal damage and theft, robbery and other crimes. A minimum price of 40p is estimated to reduce total crimes by 9000 per annum.

Modelling statement M24

The evidence base for under-age purchasing is limited (because the youngest ages for which purchasing data exists in the 'Expenditure and food survey' are 16 and 17, and there are concerns on reliability even for this). Given this caveat, crime harms are estimated to reduce particularly for young people aged 11 to 18 years because they are disproportionately involved in alcohol-related crime and are affected significantly by targeting price rises at low-priced products.

Modelling statement M26

Unemployment harm estimates [that is, estimated unemployment due to alcohol consumption], reduce proportionately more than health or crime harms. Generally, all policy options that target harmful and hazardous drinkers are effective in reducing alcohol-related harm in the workplace. The size of the effect is dependent on the extent of price increases.

Modelling statement M27

Unemployment due to alcohol problems among harmful drinkers is estimated to reduce as prices increase: for example, a 40p minimum price is estimated to result in 11,500 avoided unemployment cases, while a 50p minimum price is estimated to result in 25,900 avoided unemployment cases. Absence reductions are particularly focused on hazardous and harmful drinkers: for example, for 40p, the 134,000 estimated reduction in days absence is made up of 38,000 days for hazardous and 78,000 days for harmful drinkers.

Modelling statement M29

The societal value of harm reduction for many of the potential policies can be substantial. When accumulated over the 10 year time horizon of the model, many policies have estimated reductions in harm valued over £500m. For example, a 40p minimum price is valued at £4bn over the 10-year period. The financial value of harm reductions becomes larger as prices are increased.

Modelling statement M34

Moderate drinkers are affected in only very small ways by the policy options examined, both in terms of their consumption of alcohol and their spending.

Modelling statement M35

In terms of the differential effectiveness for priority groups, harmful drinkers are expected to reduce their absolute consumption the most, but in the more effective policy options, they also spend significantly more on their purchases.

Modelling statement M36

Policies which target low-priced alcohol affect harmful drinkers disproportionately. This is because moderate drinkers tend to drink a smaller proportion of the very low priced products available.

Modelling statement M37

There are significant effects on harmful drinkers, but important health gains also occur in hazardous and moderate drinkers. Even though moderate drinkers are at a lower risk of health-related harms, small changes in the consumption of the large number of moderate drinkers feed through in the model to small changes in risk and appreciable changes in population health.

Modelling statement M51

Though smaller than price effects, outlet density reductions have been proven to reduce both consumption and harm. As an example, the 10% reduction in outlet density has the following estimated effects:

% change in consumption

Deaths per annum (full effect)

Hospital admissions per annum

Crimes per annum

Work absences (days per annum)

Un-employment (persons per annum)

- 2.3%

-710

-25,000

-61,000

-284,000

-8100

Modelling statement M55

Modelling a 10% change in licensing hours produces changes in alcohol consumption based on three studies of -1.2% (Canadian), +0.2% (US), and -3.5% (Swedish). As an example, the 10% reduction in licensing hours has the following estimated effects:

% change in consumption

Deaths per annum (full effect)

Hospital admissions per annum

Crimes per annum

Work absences (daysper annum)

Un-employment (persons per annum)

- 1.2%

-420

-14,000

-27,000

-138,000

-3400

Economic analysis

The cost-effectiveness reviews and economic modelling showed that increasing the price of alcohol is likely to be a cost effective way of reducing consumption and alcohol-related harm. This could involve a general price increase, imposing a minimum price per unit or placing restrictions on discounting.

There was limited evidence on the effectiveness of reducing the availability of alcohol and restricting or banning advertising. Exploratory analyses suggested that policies to address these issues would probably have a smaller positive effect than that expected by a price increase.

The cost effectiveness reviews and economic modelling suggested that screening plus a brief intervention at the next GP consultation, the next registration with a new GP, or the next A&E visit would be cost effective when compared against 'doing nothing'.

Fieldwork findings

Fieldwork aimed to test the relevance, usefulness and feasibility of putting the recommendations into practice. The PDG considered the findings when developing the final recommendations. For details, go to the fieldwork section in appendix B and 'Alcohol-use disorders: preventing the development of hazardous or harmful drinking'.

Fieldwork participants who work within the alcohol field were positive about the recommendations and their potential to help prevent alcohol-use disorders. However, they felt that a number of areas should be given further consideration as follows.

  • A treatment pathway should be provided which not only illustrates the stages of care that the recommendations cover, but also outlines the roles and responsibilities of different professional groups.

  • Good communication is needed between NICE and organisations in non-healthcare settings to ensure alcohol is tackled as part of partnership working.

  • NICE should work closely with the National Treatment Agency (NTA) to ensure commissioners' concerns about the relative lack of investment in alcohol services (compared with drug services) is considered.

  • The term 'motivational counselling' should be reconsidered or clearly differentiated from other motivational approaches.

  • The presentation of the guidance will contribute to its impact and likely adoption. A standard approach should be used whereby each recommendation is preceded by a short statement of the evidence and a discussion of the likely outcomes of implementing the proposed actions.

  • The contribution that community and voluntary groups make to reducing alcohol-related harm should be acknowledged and organisations working in these sectors should be mentioned throughout the guidance.

  • National Institute for Health and Care Excellence (NICE)