The alcohol improvement programme: evaluation of an initiative to address alcohol-related health harm in England.

AIMS: The evaluation aimed to assess the impact of The Alcohol Improvement Programme (AIP). This was a UK Department of Health initiative (April 2008-March 2011) aiming to contribute to the reduction of alcohol-related harm as measured by a reduction in the rate of increase in alcohol-related hospital admissions (ARHAs).
METHODS: The evaluation (March 2010-September 2011) used a mix of qualitative and quantitative methods to assess the impact of the AIP on ARHAs, to describe and assess the process of implementation, and to identify elements of the programme which might serve as a 'legacy' for the future.
RESULTS: There was no evidence that the AIP had an impact on reducing the rise in the rate of ARHAs. The AIP was successfully delivered, increased the priority given to alcohol-related harm on local policy agendas and strengthened the infrastructure for the delivery of interventions.
CONCLUSION: Although there was no measurable short-term impact on the rise in the rate of ARHAs, the AIP helped to set up a strategic response and a delivery infrastructure as a first, necessary step in working towards that goal. There are a number of valuable elements in the AIP which should be retained and repackaged to fit into new policy contexts.

BACKGROUND

Alcohol policy initiatives 2004–2008

The Alcohol Harm Reduction Strategy for England (The Prime Minister's Strategy Unit, 2004), published in 2004 after considerable delay, was greeted by a storm of criticism from health professionals and alcohol specialists, particularly regarding the adequacy of its evidence base, the mechanisms for implementation and the political processes underlying its formulation (e.g. Room, 2004). Nevertheless, the Strategy was followed by a number of policy actions, which, it could be argued, served to raise awareness of alcohol-related harms, to augment the evidence base for action, to engage health professionals to a greater extent and to begin the process of revitalizing the structures for service delivery.

Safe Sensible Social (Department of Health, 2007) proposed structures and initiatives for reducing the social and economic costs of alcohol misuse. Local Area Agreement partnerships were identified as best placed to plan a comprehensive approach, bringing together the various interests (e.g. crime, health, education) involved in tackling alcohol-related harms. In 2007, Public Service Agreement 25 (PSA 25) Indicator 2 was put in place to measure the number of alcohol-related hospital admissions (ARHAs), and Vital Signs Indicator 26 (and National Indicator 39), introduced in 2008, measured variation in the rate of ARHAs. Statistics on alcohol and detailed regional and local profiles of alcohol-related health harm, developed by the North West Public Health Observatory (NWPHO) (Jones ), provided a national picture, updated regularly, of the extent to which alcohol was associated with acute and chronic conditions. An Effectiveness Review (Raistrick ), and a commissioning guide, Signs for Improvement (Department of Health, 2008), drew on available research, expert opinion and case studies of good practice to provide evidence for service development. In particular, an argument was made for the improvement of specialist treatment, expansion of the number of alcohol liaison nurses/alcohol health workers (AHWs), expanded identification and brief advice (IBA), social marketing and advocacy activities. These initiatives, among others, were consolidated into the Department of Health's Alcohol Improvement Programme (AIP), which ran for 3 years from April 2008.

The AIP could be seen as a concerted effort to address alcohol-related harm and to refocus the balance of action towards health after decades when alcohol had a low policy profile on health service agendas, when it was subject to relative resource deprivation in comparison with drugs and when it tended to be viewed through the lens of crime and disorder. This paper presents a brief overview of the AIP and some of the main evaluation findings. It argues that a long-term view needs to be taken in assessing the impact of action to change alcohol consumption and related harms.

The AIP 2008–2011

The main formal aim of the AIP was to contribute to the reduction of alcohol-related harm as measured by a reduction in the rate of increase in ARHAs. ARHA was chosen by the DH as the main outcome indicator because it focused on a standard measure which had been used since 2003, and could, therefore, provide information on trends. To achieve the aim, the AIP objectives were to:

support capacity and capability-building in local areas to ensure sustainable improvement in interventions to reduce alcohol-related harm;

collate and disseminate evidence, data, tools and guidance to support the NHS and local partners and make efficiency savings available from improved alcohol services; and

produce guidance on the key enablers and activities for change.

The AIP was a complex multi-component programme consisting of seven coordinated elements:

The Alcohol Policy Team at the Department of Health led the initiative and provided guidance.

The NWPHO produced statistics on alcohol-related health harms at national, regional, local and PCT levels.

The Alcohol Learning Centre (ALC), an online site, provided up-to-date information on alcohol, training packages, forums for communication and discussion between groups working on alcohol.

Regional Offices and Regional Alcohol Managers (RAMs) were employed specifically for the programme; their role included fostering implementation of ‘high-impact changes’ (HICs) in their regions.

The Alcohol National Support Team (ANST) provided PCTs with an intensive 4-day ‘diagnostic’ visit aiming to leave the PCT with a recommended plan of action for implementing the HICs.

Twenty ‘early implementer’ PCTs (EIs) were given additional funds and support to encourage them to go ‘a little further faster’ in implementing the HICs.

Seven HICs were chosen as the most effective, evidence-based actions likely to contribute to reducing alcohol-related harm. The first three HICs were enabling actions intended to facilitate intervention. The remaining four HICs were interventions that could be commissioned and implemented at local level.

High-impact changes

1. Work in partnership

2. Develop activities to control the impact of alcohol misuse in the community

3. Influence change through advocacy

4. Improve the effectiveness and capacity of specialist treatment

5. Appoint an AHW

6. IBA—provide more help to encourage people to drink less

7. Amplify national and social marketing priorities.

The programme had a budget of £22,352,336 over a 3-year period from 2008 to 2011.

EVALUATION METHODS

Ideally, the evaluation would have started at the same time as the AIP and continued to collect data for a period after the end of the programme. However, the AIP was not a ‘planned’ programme. As discussed above, it evolved over time from activities following the 2004 and 2007 policy documents prior to being ‘branded’ as a programme and it continued to evolve during its 3-year lifespan from 2008 to 2011. The evaluation was commissioned to begin in March 2010 (after the start of the programme) and ended 6 months after the end of the programme. Thus, both the nature of the programme and the period of the evaluation influenced the evaluation approach and what could be achieved. The evaluation had to take account of the changes occurring before and during the evaluation, of variable regional and local contexts which influenced the implementation of the HICs and of the different starting points across the country. In some places, action on alcohol had a long history and well-established structures for service delivery; in other areas, there was a need for awareness-raising, and new mechanisms for developing appropriate responses had to be put in place. Local cultures and histories of implementation resulted, therefore, in wide variations in the implementation of specific elements of the AIP which could not be investigated in detail within the resources and timescale of the evaluation. Given these constraints, a qualitative approach, in combination with a quantitative element, was judged to be best suited to evaluating both process and impact, to reflecting the diversity of perspectives and experiences of those involved in the AIP and to examining some of the main assumptions of the programme.

The primary evaluation aim was to assess the extent to which the AIP had impacted on alcohol-related harms and contributed to efforts to reduce the rate of increase in ARHAs (the main outcome). In addition, the evaluation aimed to describe and assess the process of implementation, to gain some understanding of the elements of the programme which had worked well and which had potential for mainstreaming into existing service structures or adapting within new policy and delivery contexts. Thus, both process and impact evaluation were included.

A combination of existing statistical information (in particular, statistics generated by the NWPHO), documentary sources and new data was used:

Six in-depth, taped interviews were conducted with national level policy-makers: the DH Alcohol Policy Team, the ANST and the ALC. Thematic content analysis provided contextual understanding of the development of the AIP and identified themes for further exploration.

Twenty-five in-depth, taped interviews were held with regional level informants: all RAMs; selected members of Government Regional Offices; other key informants, recommended by the RAMs. Interviews provided information about the implementation of the AIP in the regions and their histories of alcohol action and afforded further contacts for follow-up. Analysis of these data provided descriptive accounts and identified issues for further examination. Using the RAMs as ‘gatekeepers’ was the quickest and most efficient way to get relevant contacts who knew about the AIP and this may have resulted in more positive assessment of AIP elements. However, we were aware of this issue, and attempts to offset bias were made by snowballing out from initial contacts and by including open discussion questions in the telephone survey which reached less involved stakeholders.

Structured telephone interviews (with open comments) were conducted with 16 EI and 28 non-EI PCT alcohol leads. Sampling used a list of PCTs which ranked PCTs by levels of ARHAs and deprivation. All 20 EIs were selected (but 5 did not respond). From each region, a quota of five other PCTs with the highest levels of ARHAs was selected. Survey data (analysed using SPSS) generated descriptive statistics. Open-ended questions generated illustrative themes. Survey responses guided case study development.

Four case studies (two EI and two non-EI) were conducted. They did not aspire to be ‘typical’; they aimed to obtain understanding of the complexity and dynamics of implementation and, importantly, the influence of local context. Main criteria for inclusion were the PCT had engaged with the survey; had received an ANST visit; and the alcohol lead agreed to a second interview and to facilitate access to other key informants. Interviews were conducted with 26 individuals from a variety of professional backgrounds. Content analysis produced a narrative account of AIP implementation in each area.

Examination of ARHA trend data. Data collated by NWPHO were analysed to address the impact of the AIP on reducing the rate of increase of ARHAs. A linear regression model estimated whether the AIP had an impact on the rate of change over time using these covariates: region, year: pre-intervention (2002–2008) and post-intervention (2008–2010). The interaction between year and whether pre- or post-AIP implementation was the variable of interest reported. A second model controlled for whether the PCT was an EI or not. A descriptive approach was taken to examine the year-on-year mean percentage change in ARHAs before and after the introduction of the AIP. PCT-level data were analysed nationally, regionally and by those EI and non-EI PCTs included in the survey. Gender differences and whether admissions were wholly or partially attributable to alcohol were examined.

Visits by the evaluation team were made to all regions; RAM meetings at the DH were attended; some regional events were attended; email contact with RAMs was ongoing.

A modified version of framework analysis was used to inform data collection and analysis (Ritchie and Spencer, 2002). This allows for the inclusion of themes drawn from the literature/other research as well as themes emerging from the data and for analyses and interpretation of data within a thematic framework. Data collection and analysis were iterative, with insights from each phase of data collection feeding into the development of subsequent data collection. Analyses of the qualitative data (interviews) used standard procedures for developing categories and themes (Ritchie and Spencer, 2002); the four case studies were compiled by looking across individual interviews in the same PCT to create a new set of categories and themes which were drawn from all case study interviews. Case study themes were developed through team discussion (largely between two researchers, P.T. and J.T.). This process incorporated a limited triangulation with the dual purpose of ‘checking’ data reliability but also revealing contradictions and new perspectives within the data.

Ethical approval was obtained from researchers' universities (LSHTM, Middlesex, York), and established ethical guidelines were followed in conducting and reporting the evaluation.

The present paper provides an overview of the main results from three key evaluation questions:

Did the AIP impact on ARHAs?

Did the additional support to selected early implementer PCTs (EIs) result in quicker, more effective delivery of HICs?

To what extent were the separate programme components successful and how did they contribute to the AIP overall?

RESULTS

Did the AIP impact on ARHAs?

Data collated by the NWPHO (2002/2003 to 2009/2010—the only data available at the time) were examined to address the question of the impact of the AIP on reducing the rate of increase of ARHAs.

A linear regression model was used to investigate the relationship between intervention and rate of admission over time. We found that the rate of admission had increased year on year and this was reflected in the model as the year covariate was highly significant (P < 0.001); for each increase in year, the rate of admission increased by 180 (95% CI 115–246). The interaction between year and intervention was not significant in predicting rate of admission, indicating that there was no evidence that the AIP had an impact on changing the rate of admission (see Table 1 for details).

Table 1.

Parameter estimates for linear regression model

Parameter	B	Standard error	Significance	95% CI
Intercept	405	253	0.11	−92 to 901
East Midlandsa	74	44	0.09	−12 to 160
East of Englanda	−219	40	P < 0.001	−296 to −141
Londona	−127	33	P < 0.001	−192 to −63
North Easta	457	40	P < 0.001	378 to 536
North Westa	380	35	P < 0.001	312 to 448
South Easta	−309	37	P < 0.001	−382 to −236
South Westa	−80	39	0.04	−156 to −3
West Midlandsa	−23	37	0.53	−96 to 49
Pre-interventionb	427	253	0.09	−68 to 923
Year	180	34	P < 0.001	115 to 246
Pre-interventionb * Year	−56	34	0.10	−123 to 10

aCompared with Yorkshire and the Humber.

bCompared with post-intervention.

R2 = 0.63.

Figure 1 shows that levels of ARHAs varied greatly across the regions. In the North West and North East, they were well above the others, with East of England and South East England being slightly lower. However, all regions showed a similar, rather regular, increasing trend over time, with no obvious changes associated with the introduction of the AIP.

Fig. 1.

Regional rate of ARHAs per 100,000 of the population.

Examining national data, as expected, male rates of ARHAs were significantly higher (P < 0.001) than female rates, and partially attributable admissions were significantly higher than wholly attributable (P < 0.001). The attributable fraction may be defined as the proportion of disease risk in a population that would not have occurred if exposure to a risk factor or set of factors had not occurred. The alcohol-attributable fraction (AAF) is therefore calculated as a positive function of the prevalence of drinking (the exposure) and the relative risk function of each alcohol-related condition (the disease risk) to enable the estimation of the proportion of cases of a disease or type of injury that may be attributed to the consumption of alcohol (Jones ). From December 2008, AAFs are calculated for 45 conditions, of which 13 are wholly attributable to alcohol consumption and 32 are partially attributable to alcohol consumption. Regional data indicated that wholly attributable admissions had remained stable over the years, with increases in admissions showing for partially attributable admissions.

Frustration with the indicator itself was commonly voiced by study respondents. For some, NI39 and VSC26 were too narrowly cast and failed to embrace the broader range of alcohol-related harm. A main source of frustration stemmed from the way the indicator was measured—and in particular the partly attributable hospital admissions such as hypertensive diseases and cardiac arrhythmias. A number of respondents referred to ‘wild’ fluctuations in ARHAs at a PCT level. The reasons for such fluctuations were often in doubt but one frequently proffered explanation was that other health initiatives had impacted on an alcohol-related disease—hypertensive diseases in particular. In some areas, dramatic local fluctuations were put down to changes in the way in which conditions were being coded in particular hospitals. These problems led to some cynicism about the potential for the AIP to have a measurable impact on ARHAs. At the same time, the value of having an indicator was widely acknowledged. Indicators were seen as ‘valuable as they gave us a focus on which to pin actions’ (RAM interview). It was felt that without a target for alcohol, ‘it would never get the attention it deserves and never get any investment … if you're not failing against your target then there is no reason why PCTs should be paying any attention to it’ (PCT respondent).

Did the additional support to selected EIs result in quicker, more effective delivery of HICs?

The EIs were 20 PCTs chosen from a group identified as ‘spearheads’ (areas of highest health inequality, which were being given wider health policy priority at this time); these were largely areas where rates of ARHAs were also high. They were provided with additional support to encourage speedier, more effective implementation of the HICs. The linear regression model indicated that there was no evidence that the rate at which ARHAs were increasing was different in EI compared with non-EI sites or that the intervention had a differing effect on these sites.

However, findings from the case studies and responses from the PCT interviews indicated that EI status was seen as important in helping to initiate and strengthen strategic development and build infrastructures for service delivery. The additional funds had been used to support the development of new initiatives, to consolidate partnerships and to expand existing work which otherwise would not have been achieved.

What the Early Implementer funding gave us was much more of a corporate and partnership buy-in as a shared agenda because the EI funding was not just invested in health related initiatives. It was invested in initiatives, some were innovative, some were based on the best evidence we could get and some weren't because we were trying them out. (Director, Public Health, EI case study area)

To what extent were the separate components successful and how did they contribute to the AIP overall?

The advantage of a multi-component programme, such as the AIP, is that initiatives are intended to be complementary and the impact of the programme is derived as much from the synergistic effects of its various components as from the outcomes of the separate elements (Holder, 1998). Evaluations of multi-component programmes have generally concluded that it is not possible to quantify specifically the contribution of any one component to the outcomes of the programme as a whole (US Department of Health and Human Services, 2000). This applies also to the evaluation of the AIP. We were able to investigate the extent to which each component was seen to have met its objectives and the extent to which it was deemed to be ‘successful’; but we could not apportion any weight to individual components regarding their contribution to the AIP overall.

Each component of the AIP attracted both positive and negative comments. For example, despite suggestions on how the ALC might be improved and some criticism of the tools on the website, its popularity as a one-stop shop was reflected in many of the responses:

It was good: for example, I wanted to know what screening tools were out there and it told me, I wanted to know about SIPS and that told me all about SIPS, you didn't have to go searching all over the place, you have got a lot of information in there. Also there was an IBA training course which I dipped into. (PCT Alcohol Lead)

It was notable that even where respondents were negative about a particular aspect of a component, this was generally couched within a more positive assessment of the AIP as a whole. Taken together, the components aimed to provide the field with best available evidence, to improve the infrastructure for raising awareness and delivering initiatives to reduce alcohol-related harms, and to improve workforce capacity and skills. Table 2 provides an overview of AIP achievements across these three domains.

Table 2.

Measures of progress and achievements

Measure	Achievement
Best available evidence
Compilation of existing evidence	These were all achieved. SIPS research still to report but interim presentations have been given. Signs for Improvement provided evidence and guidance in the context of world class commissioning
Data on alcohol consumption/trends, etc./national and local profiles (NWPHO)	Evidence made widely available through ALC and through events run by national policy team and the RAMs
SIPS research	SIPS research looked at the implementation of IBA in various health and criminal justice contexts
Dissemination of evidence
Improvement of infrastructure
Establishment of RAOs/RAMs	These were established but at different points in time and they were more or less well embedded in regional structures—in any case, the regional level was abolished when the new Coalition government came into power
Identification of EIs and related activities	These were identified and were focus of activity but with funding cuts the number was reduced in year 3 and activities in each were somewhat reduced; variable results
Development of infrastructure for service delivery	New networks and partnerships set up, e.g. network of commissioners, network of AHWs; aspects of the HICs—increased number/range of delivery locations of IBA/AHWs
Number of PCTs prioritizing VSC26	Two-thirds (100/153)
Number of visits made by ANST	34 (20 EIs and 14 others)
Directly Enhanced Service (DES)	A DES for alcohol was established in 2008/2009 to run for 2 years and was extended into 2010/2011. The DES provides an incentive to GPs to provide IBA to their newly registered patients. This amounts to 8% of adults, or about 3.3 million adults who will annually be the target of this activity
Development and use of ALC	In 2011: 10,981 visits; 4488 registered users; 2394 e-news subscribers
	RAMs thought it was very important that the ALC should remain after the end of the AIP
Improving workforce capacity and skills
Implementation of HICs	Capacity improved through support for new IBA and AHW and by learning generated in course of HIC implementation
Capturing and sharing learning	ALC did this as well as networks developed within regions by RAMs and others
Development of e-learning modules	For example, three IBA e-learning modules: primary care, community pharmacy, hospital settings, on ALC
IBA Train the trainer events	National ‘Train the trainer’ events provided by the Improvement Team, training over 100 health and criminal justice practitioners to deliver IBA
Establishment of post-graduate certificate in the management of alcohol in primary care	Since the launch in September 2009: over 28 training events in the UK; over 658 health professionals trained
	Certificate from the Royal College of General Practitioners
Undergraduate medical training	Developed consensus guidance on the teaching of substance misuse in the undergraduate medical curriculum (2005–2008). The project to embed the guidance in all English medical schools commenced in 2009
Development of tools by Improvement Support Programme	For example, ready reckoner; Alcohol Systems Model; GP templates linked to DES scheme; Rush model spreadsheet, etc. All available on ALC

Some elements of components were emphasized in particular as key levers for change:

dedicated issue-specific champions at regional and local levels (such as RAMs and local professionals nominated as ‘appointed’ champions as part of their role);

a mechanism (the RAMs and the ALC) to secure good communication links between local and national levels and across local areas/PCTs;

a mechanism (ALC) to facilitate information transfer, and promote workforce skills development;

availability of best evidence to inform policy and practice developments;

availability of tools to assist the development of an economic rationale for action on alcohol specific to individual PCTs/local conditions;

support to legitimize and promote innovative, outcomes-focused interventions;

an agreed target (ARHAs) to stimulate and sustain action.

The HICs were adopted and developed variably as suited the specific local context. Working in partnership was accepted without question as a prerequisite for action and this was promoted in all areas. Other HICs developed most extensively were introduction and expansion of IBA and appointment of AHWs; these services were developed in different ways across local areas. As mentioned earlier, the ethos underpinning the AIP was to allow local areas flexibility in implementing change; as a result, we were not able to investigate the consistency or precise details of service delivery across provider services within the resources and time span of the evaluation. A number of respondents were sceptical of social marketing, which they felt was less well evidenced than other HICs. Part of the problem may have arisen from lack of understanding of social marketing as an attempt to prompt behavioural and environmental change in the long term rather than convey traditional health awareness messages. According to other informants, social marketing in some areas did result in successful ‘branding’ of the local programme in that it was picked up and disseminated by local media, and provided a recognizable banner for coordinating partnerships and local action. The branding of the AIP initiative as a national policy effort did have some impact in some regions; it aimed mainly to impact on middle-level decision-makers and encourage them to view their activities in a coordinated way and to raise alcohol on local agendas—we have concluded that in this the AIP was often successful despite the expressed scepticism.

Overall, the majority of respondents were positive about the AIP, which was seen to have contributed to raising awareness of alcohol-related harm, stimulating a more strategic and coordinated response to alcohol-related harm at local level, supporting capacity building and skills development in the workforce and promoting a culture of ‘spend to save’ and outcomes-focused approaches to service delivery.

[T]he art of this really, as a national project, has been in making sure that people feel that they are leading it locally and I think it has been quite cleverly managed the whole thing and the years to come will tell, won't they, but I do think it will have a big impact. (Senior Manager, acute NHS Trust).

DISCUSSION

There were many reasons why the AIP did not result in a measurable decrease in the rate of increase of ARHAs. A £22 million programme such as the AIP represents a modest intervention when set against national trends in drinking and related ill-health, and trend data would need to be collected for a longer time following the end of the programme. Furthermore, along with improved methods of identification and recording (AAFs), AIP activity may have led, over the short term, to an increase in the numbers identified, referred to hospital care and recorded as alcohol-related health problems. This could be seen as a positive outcome of the AIP. In both EI and non-EI PCTs, local factors were extremely influential in driving responses to the AIP initiatives. It was not surprising, therefore, that the outcome measure, a reduction in the increasing rate of AHRAs, did not show a measurable reduction nationally and did not differentiate between PCTs. Even if there were to be a reduction in the rate of increase of ARHAs in the longer term, it would be difficult to attribute causality definitively to the AIP. The programme augmented, complemented and consolidated other national and local action, some of which had been in place prior to 2008 and some initiatives (e.g. alcohol liaison workers) encouraged by AIP action have continued to expand (Patton, 2012).

Key factors which influenced the impact of the AIP were similar to those identified in other multi-component programmes (Thom and Bayley, 2007):

the nature of the local context, especially the extent to which there was a local tradition of concern and action on alcohol-related harms;

the degree of strategic buy-in at national, regional and local levels;

the existence of well-focused aims and objectives and agreed goals and targets;

the existence of champions and local ‘activists’;

the extent of good partnership working and sharing of information;

the availability of appropriate, adequate resources.

The evaluation supports findings from a study of local strategic partnerships by Geddes and the National Evaluation Team (2006). They reported that more successful partnerships flourish within local contexts characterized by a virtuous circle—one with a history of strategic partnership working, with trust and good working relations, where there is a stable local political environment, where potential partners are willing to engage and where there is effective local leadership, the presence of local champions and good management. Where these features are lacking, partnerships find themselves within a vicious circle where it is more difficult to achieve effective working. As noted before, the AIP was implemented in local contexts with very variable starting points regarding action on alcohol-related harm. In some areas, the AIP was able to build on and augment existing activities; in other areas, the first task was to begin to raise awareness of the issues and to put in place appropriate structures to deliver initiatives and lay the foundations for building a virtuous circle. Sustainability of effort is, therefore, a major consideration when looking at the potential legacy from the AIP and was an important aspiration for the programme.

It was hoped that AIP support would drive change by providing a springboard for new activities, by enabling new working patterns to emerge and by initiating service development. While the AIP cannot be seen in isolation from prior trends and action to address alcohol-related harm, the evaluation findings indicate that it did function as an important change agent. In particular, it fostered the expansion and mainstreaming of IBA, AHWs and collaborative networks for information sharing and partnership working around commissioning and service delivery. The development of a ‘brand’ in some regions was also a way of ensuring that future action could be pulled together under a recognizable ‘umbrella’. To some extent, it succeeded in implanting an outcomes and cost driven approach to commissioning and service development which could be used as a rationale for gaining PCT cooperation and securing funds for alcohol services.

There are a number of elements in the AIP which, evaluation findings indicated, should be retained and repackaged to fit into the new policy context following the 2010 election of the Coalition government. It is recognized that there will be considerable challenges to sustaining or adapting valued aspects of the AIP in the current climate which has been characterized as a ‘cycle of continuous change’ (Hunter and Perkins, 2012, p. 51). Nevertheless, accepting the need for adaptation to new administrative and clinical structures, five core aspects of the AIP legacy are suggested for future policy consideration:

Sustain vertical and horizontal communication by continuing to provide a ‘one stop shop’ vehicle for communication, information sharing on good practice and evidence-based approaches, workforce development and training; and ensure a conduit for communication and liaison between central government and local authorities, public health and service providers.

Retain targets (such as achieving a reduction in ARHAs) as an incentive to action and attempt to improve the appropriateness, type and relevance of outcomes to measure and reflect the range of harms and intervention approaches in the alcohol field.

Continue to support collaborative structures, networks and partnerships and ensure that these operate effectively within changing policy contexts. Adapt them to the new health, local government and policing structures developing at local level but retain and foster a ‘virtuous circle’ approach.

Retain a strong policy focus on alcohol by identifying single-issue alcohol champions in local areas; ensuring awareness of alcohol issues and a position on policy agendas; targeting commissioners and developing the business case for alcohol.

Provide support for service development and innovation to encourage change and development within local provider settings and to provide resources to encourage innovative, outcomes-focused initiatives.

Finally, although some aspects of the AIP were potentially directed towards developing hospital services—through the appointment of AHWs, for example—the AIP was intended to stimulate awareness and action at primary care and community level and engagement with hospital consultants/services was not a core part of the RAM role or of the programme objectives. With longer term objectives in mind, it is suggested that future policy action and research should consider the impact of primary care and other community level activity on the delivery of hospital services and on rates of ARHAs.

CONCLUSION

Although there was no measurable short-term impact on the rise in the rate of ARHAs, evaluation findings indicate that the AIP helped set up a strategic response and a delivery infrastructure as a first, necessary step in working towards that goal. The AIP had always been a time-limited programme, but the intention had been to leave behind a legacy, in terms of improved evidence, infrastructure and workforce capacity, and the evaluation found that there were a number of valuable elements of the AIP which respondents believed should be retained and repackaged to fit into new policy contexts.

Funding

This is an independent report commissioned and funded by the Policy Research Programme in the Department of Health (PRP reference number: 0030006). The views expressed are not necessarily those of the Department. Funding to pay the Open Access publication charges for this article was provided by Middlesex University using the study grant from the Department of Health Policy Research Programme, PRP Reference Number: 0030006.

Conflict of interest statement: In the last three years, B.T. has attended an International Centre for Alcohol Policy (ICAP) research advisory group meeting (expenses and honorarium) and written a chapter for in ICAP book (honorarium). She is a member of ICAP research advisory group since May 2012 (no remuneration). Other authors have no interests to declare.