Aims: To describe the development and evaluation of an innovative web-based service for hazardous or harmful drinkers. Methods: Patients were identified in General Practice and referred to the new service. Mixed methods were used for evaluation. Results: The service was feasible and acceptable to patients, primary care professionals and commissioners. Users appeared to reduce their alcohol consumption. Conclusion: This model may be of interest to other primary care commissioners looking to increase access to alcohol treatments at low cost.
Excess alcohol consumption is a major public health problem, causing about as much of the global burden of disease as smoking (Rehm et al., 2009). Nearly 1:4 of the adult population, or ∼7 million people in England alone, drink more than recommended limits, and are hence at increased risk of harm from alcohol consumption (Drummond et al., 2005). Effective treatments exist (Raistrick et al., 2006) but there have been real problems with implementation (Nilsen, 2010)—in 2004 it was estimated that only 1:18 people who could benefit from alcohol treatment received it (Drummond et al., 2005).
In 2006, the Department of Health for England and Wales mandated Primary Care Trusts (PCTs) to improve service provision with a four-tiered framework (Department of Health, 2006). ‘Tier 1’ is delivered in primary care by generic health care professionals and is equivalent to screening and brief intervention (SBI) plus onward referral where needed. ‘Tier 2’ requires staff with some training in alcohol misuse and includes extended brief interventions to help alcohol misusers reduce alcohol-related harm plus assessment and referral of those with more serious alcohol-related problems to Tier 3 or Tier 4 services. These require highly trained specialized staff and focus on patients with complex needs (Department of Health, 2005, 2006).
For most areas in the UK, this required a large increase in service provision at all tiers. Barriers to increasing delivery of SBI in primary care (Tier 1) include reluctance of health professionals to bring up the subject of alcohol for fear of offending patients, a widespread nihilism about the effectiveness of alcohol treatments and reluctance to identify a problem if it is difficult to provide treatment for it (Beich et al., 2002; Rapley et al., 2006; Nilsen, 2010). Barriers to increasing provision of Tier 2, 3 and 4 interventions include cost and a lack of personnel trained in providing effective therapeutic interventions (Drummond et al., 2005).
Web-based interventions for alcohol misuse may help overcome some of these barriers. Computer-based interventions can reduce alcohol consumption among adult problem drinkers (Khadjesari et al., 2011). They are likely to appeal to users because they are convenient (can be accessed at any time of day), anonymous (important in stigmatized conditions like alcohol misuse) and available at time of need (important in conditions with high rates of relapse like alcohol misuse).
We report here on the experience of one PCT that introduced an innovative web-based service to deliver Tier 2 interventions as this model could increase levels of service provision nationally. The service could be delivered by staff without specific training in alcohol or other substance misuse. The aim of this paper is to describe the service along with the results of the initial evaluation of the pilot with a view to helping commissioners and policy-makers who may wish to introduce a similar service. Specific objectives are to:
describe the innovative service and its implementation;
determine the feasibility and acceptability of the new service;
describe its effects on users; and
describe the costs associated with implementing the service.
This was a single-group experiment with no comparator using mixed methods. Ethical approval was obtained from a local NHS National Research Ethics Committee (Ref 09/H0722/60).
The study was undertaken in one PCT on the outskirts of London which covered a relatively affluent area (Anon, 2007). There were fewer binge drinkers (11.3%) in the area compared with the national average (20.1%). Despite this, 1863 hospital admissions due to alcohol-related harm were recorded in 2008/2009 (Public Health Observatories, England, http://www.apho.org.uk/). Prior to the service described here, the only statutory alcohol service was for heavily dependent drinkers requiring specialist input. The main voluntary service was Alcoholics Anonymous, also aimed at dependent drinkers. There were 28 General Practices with a combined registered population of ∼180,000 (Hildebrand, 2007).
The study was carried out in adults aged 18 or over identified in General Practice as drinking hazardously or harmfully who accepted referral for help to reduce their alcohol consumption after a brief intervention. General Practitioners (GPs) were asked to continue to direct patients requiring specialist input (e.g. patients with severe dependency, co-existing drug dependency or severe mental health problems) to the pre-existing specialist alcohol service.
The intervention/service had three components:
assessment of alcohol consumption, dependence and alcohol-related harm;
a web-based self-help programme known as Down Your Drink (DYD);
support in using the self-help programme.
An Alcohol Project Co-ordinator (APC) was employed to set up the service. The initial post-holder had experience of working in the field of alcohol, but the service was designed to be delivered by staff without specialist training.
Assessment of alcohol consumption and alcohol-related harm can trigger change even without further intervention (Kypri et al., 2007; McCambridge and Day, 2008). An assessment was undertaken by the APC using a pre-determined script and validated assessment instruments completed online by the patient on a project-specific laptop. Measures included past-week alcohol consumption (TOT-AL) (Khadjesari et al., 2009), the Alcohol Use Disorders Identification Test (AUDIT) (Babor et al., 2001), the Leeds Dependence Questionnaire (LDQ) to measure dependence (Raistrick et al., 1994), the Clinical Outcomes in Routine Evaluation (CORE-10) as a measure of mental health (Connell and Barkham, 2007) and the EQ-5D to measure health-related quality of life (Rabin and de Charro, 2001). Demographic data, including age, gender, ethnicity, internet access and internet experience were recorded.
The online questionnaires were automatically scored with the results provided to the APC for feeding back to the patient according to pre-specified algorithms. Where patients’ scores suggested high levels of dependency (LDQ scores 21 or more), or high levels of psychological distress with risk of suicide or self-harm, the results were sent to the GP with recommendations for additional support.
Web-based self-help programme
The programme used in this service was DYD. It consisted of three phases each with a strong theoretical underpinning and delivered treatments known to be effective face-to-face (Raistrick et al., 2006). The first phase (‘It's Up To You’) was based on motivational interviewing and aimed to encourage the user to reach a considered decision about changing drinking behaviours. The second phase (‘Making the Change’) used computerized cognitive behavioural therapy (CCBT) and behavioural self-control techniques to help users cut down. The third phase (‘Keeping on Track’) focused on relapse prevention (Linke et al., 2008).
Support to use the web-based programme
In common with many web-based interventions (Eysenbach, 2005), considerable attrition from DYD has been observed (Linke et al., 2007; Wallace et al., 2011). ‘Facilitated access’ or support in use of the programme can reduce attrition from CCBT (Robinson et al., 2010; Titov et al., 2010; Johnston et al., 2011). In the UK, facilitated access to CCBT is part of the Improving Access to Psychological Therapies (IAPT) programme, where facilitation is offered by ‘graduate mental health workers’—staff with a degree, usually in psychology, but no specific training in mental health and no previous clinical experience (Department of Health, 2008). Experience from IAPT informed development of this new alcohol service.
The APC gave the patient personalized log-in details for the web-based programme (user name and password), introduced the patient to the main sections of the website and provided as much support as was needed to enable the patient to navigate around the site on their own. The APC advised patients to start with some of the exercises in Phase 1 of the website. Patients were given a booklet containing the DYD url, their personal log-in details, reminders of the contents of the three phases, information about free/low-cost internet access locally and the APCs name and contact details. Patients were encouraged to ring the APC if they experienced any difficulties using the website. Finally, the APC arranged to telephone the patient three times at fortnightly intervals. The purpose of these phone calls was to provide support in use of the website, not to provide counselling about their alcohol use.
The service was designed to fit as closely as possible with existing practice to promote implementation, integration and sustainability. It was launched in September 2009 with a presentation to local GPs at a regular lunchtime meeting for continuing professional development, followed by a session offering training in SBI (mostly to practice nurses), visits by the APC to individual practices in the PCT and various publicity initiatives, including articles in the PCT GP newsletter in December 2009, a patient magazine in February 2010, the PCT website from March 2010 and a local authority community newsletter in April 2010.
Hazardous or harmful drinking was defined as scoring 5 or more on the three-item AUDIT-C, a screening test for alcohol misuse (Dawson et al., 2005; Alcohol Learning Centre, 2012), and/or drinking more than government guidelines (more than 14 units a week or 2–3 units a day for women or more than 21 units a week or 3–4 units a day for men). Once a health professional (GP or practice nurse) identified a patient as drinking hazardously or harmfully, they were asked to discuss this with the patient, offer advice about cutting down and explore whether the patient would like any further help. Patients who accepted were referred to the service either electronically or by faxed hard copy. On receipt of the referral, the APC contacted the patient to offer an appointment for the assessment. Appointments were arranged between 9 a.m. and 7 p.m. and held in a safe environment, such as the PCT, local authority or referring general practice. GPs were informed of the result of the referral.
Mixed methods were used for the evaluation. Quantitative data were collected on: numbers of patients referred; numbers of patients who attended the assessment; clinical and demographic characteristics of patients attending and costs associated with the service. As staff time was the major cost, the APC kept a diary of time taken to respond to each referral, make an appointment, undertake the assessment, complete facilitation phone calls and obtain follow-up data. The initial plan had been to collect outcome data at 3 months, but the APC left the post, preventing follow-up on all but the earliest patients seen. To reduce costs, the PCT decided to incorporate the work of responding to referrals, undertaking assessments, introducing patients to the web-based intervention and offering follow-up phone calls into the job description of several generic health promotion staff. These staff were already heavily committed and the PCT decided not to continue collecting follow-up data.
Qualitative data were collected using semi-structured interviews with patients and health professionals. The semi-structured interviews were undertaken by an academic GP trainee not previously involved in the service (SC). The topic guide for the interviews with health professionals included participants’ experience of the new service, their views about the advantages and disadvantages of the service, whether and why they had (not) referred patients and suggestions for improvement. Interviews with patients explored their experience of referral, their views about the advantages and disadvantages of the service and suggestions for improvement. Interviews were recorded and transcribed verbatim, with additional field notes kept by the researcher to capture non-verbal and contextual data. Quotes are identified by the disciplinary background of the participant, with GP indicating a General Practitioner, PN indicating a Practice Nurse and P indicating a patient.
In the first 12 months after the service launched, staff from 18 out of the 28 practices (64%) attended face-to-face briefing about the service. This included nineteen staff (mostly GPs) from 9 practices who attended the lunchtime meeting, 14 staff from 11 practices who attended SBI training and 6 practices who agreed to a visit from the APC.
Thirty-one patients were referred within the first 12 months. Of these, 10 could not be contacted or did not respond to the offer of an appointment and a further 2 declined the offer of an appointment. Of the 19 patients who attended the assessment, the mean age was 42 years (range 20–68 years), 10 (53%) were female, 9 (47%) were married or in a relationship, 9 (47%) had children, 12 (63%) were currently working, 6 (32%) had a health condition or disability which prevented the patient from engaging in employment or study and 15 (79%) had home internet access. They were drinking heavily, with a mean past week alcohol consumption of 75 Unit/week (SD 36; range 10–117) where 1 Unit = 8 g of ethanol and mean AUDIT scores of 23 (SD 7.2) [median AUDIT score 23, interquartile range (IQR) 15–28]. They showed low dependence with a mean LDQ score of 10 (SD 6.9). The CORE-10 scores were in the clinical range (mean = 15, SD 9.1) and the mean EQ-5D score was 0.806 (SD 0.176).
Use of the intervention
The average duration of the initial appointment was 43 min (SD = 11 min). Six patients (32%) logged on at least once after the initial appointment; among these users, the mean number of log-ins was 8 (SD 7.3, range: 2–22). The mean number of pages visited per session was 13 (SD 26, range: 1–134).
By the end of the first 12 months, the APC had succeeded in contacting 13 (68%) patients for at least one of the planned facilitative phone calls. Achieving contact was time-consuming and required multiple attempts. Forty phone calls were made at the 2 week point, resulting in 10 contacts; 39 calls made at 4 weeks, resulting in 10 contacts and 43 attempts made at 6 weeks resulting in 9 contacts.
Impact of the service
Owing to the APC leaving 12 months after the service launched, follow-up data were available only for a few patients (n = 7; 37%). The mean reduction in past week alcohol consumption among these seven individuals was 35 units (P = 0.018). Mean AUDIT scores also reduced significantly from 23 to 19 (P = 0.018). LDQ scores reduced from 10 to 8 (P = 0.027). There was no significant change in CORE-10 or EQ-5D scores.
A log of the time taken by the alcohol project worker suggested that the total time required per patient referred, seen and followed-up was 2.5 h. This total was made up of 15 min for responding to each referral and setting up an appointment, 1 h for the baseline assessment appointment and 15 min for each of three follow-up phone calls (allowing up to three attempts to connect with each patient at each call), making a total of 45 min and a further 30 min for the 3-month follow-up (including sending e-mails, reminders, making phone contact and data entry).
Additional costs included website hosting and maintenance @ £120 per month, £3200 for developing the online baseline data assessment instruments and database and £3600 for developing the automated e-mails and follow-up questionnaires.
Ten interviews were conducted: eight with primary care health professionals (five GPs, two PNs and one Health Care Assistant) and two with patients. Saturation, i.e. where no new themes were emerging from subsequent interviews, was achieved for health professionals but not for patients. However, no other patients were prepared to be interviewed.
Patients and health care professionals found the service highly acceptable. Both groups welcomed the addition of a service aimed at non-dependent drinkers. As this practice nurse said:
But no, I think it's brilliant. I think it's something we've welcomed and we've been waiting for, for a long time, something like this that we can work with.
Or as a patient put it more vividly:
They [specialist service] basically just said, well you're not drinking enough. And I thought, that's wrong, if you're asking for help.
Health professionals were pleased that the service included support in using the website as they questioned how motivated patients would be to use it without additional support. Patients and health professionals reported that the service worked smoothly, with one patient very positive about the follow-up phone calls. Patients like the anonymity and convenience of the web-based service, which they could work through in their own time. Both health professionals and patients liked it being slightly distanced from the practice.
The interview data suggested several reasons for the low level of referrals in the first year of the service. Health professionals from all backgrounds (GPs, practice nurses and health care assistants) had considerable difficulty discussing alcohol during consultations unless the patient introduced the subject first. Health professionals were concerned that they would offend the patient by mentioning alcohol consumption, and many had doubts about whether drinking more than recommended guidelines was problematic, reflecting the prevalence of excessive alcohol consumption. As this GP put it:
Our nurse today said ‘I'm having a real problem with this’. She said, she said that ‘people who are regular social drinkers are all scoring enough for me to offer referral and I feel really awkward offering it to them because I don't think there's an issue’.
Health professionals reported difficulty in remembering the new service existed or in remembering how to access it. Some health professionals also wanted to know whether the service worked and what patients thought of it before making any referrals.
Two additional challenges arose in the first year: equity of access and technical issues. Patients who could not read English or who were physically unable to use a computer could not use the service. It was hard to find venues with high-speed internet access. Slow internet access led to the computers freezing and offered an unprofessional image.
To our knowledge, this is the first example of a statutory alcohol service using a web-based intervention. The service was shown to be feasible and acceptable to stakeholders. Low referral rates may have been due to reluctance on the part of health professionals to discuss alcohol with their patients and difficulty in remembering the existence of the new service. Patients who used the service appeared to experience benefit, but this may not have been due to the service. At the end of the 12-month evaluation period, the APC left. Despite this, the PCT were able to continue the service by using generic health promotion staff to respond to referrals using the scripts and algorithms already developed. The PCT subsequently decided to expand the service by accepting referrals from Accident and Emergency staff and nurses on the medical wards at the local hospital, probation services, housing services and social services.
There are many limitations to this study. It was set up as a development project with little funding for evaluation. The PCT was going through substantial change and reorganization which impacted on PCT staff's ability to commit the necessary time and resources to developing the new service. The APC left after 12 months and was not replaced, preventing follow-up data collection on all but seven patients. The number of patients referred was low, and it was not possible to interview more than two patients. Hence, the outcome data reported cannot be viewed as robust.
Despite all these limitations, this report should be of interest to health care commissioners and policy-makers. The challenges faced by this PCT of a large amount of unmet need for alcohol services for non-dependent drinkers, a government policy to increase the availability of alcohol services, in an economic environment of shrinking resource, and a policy context of substantial upheaval were common to many PCTs and will be common to many clinical commissioning groups. Services based on access to internet self-help programmes with a small amount of face-to-face input from relatively unskilled staff are likely to be easier and cheaper to implement than services that require substantial input from highly skilled staff (Smit et al., 2011). The reluctance we identified among primary care staff to offer alcohol-related services to ‘social’ or non-dependent drinkers is also likely to be a widespread phenomenon (Rapley et al., 2006; Wilson et al., 2011).
—Funding for the development and initial quantitative evaluation of the service was provided by Kingston Primary Care Trust. Sophie Conroy was an academic GP ST4 funded by the London Deanery at the time of the interviews.
We are grateful to Iona Lidington, Associate Director of Public Health at Kingston Primary Care Trust, for her support and encouragement throughout the project and to Paul Wallace for his expert advice.