Scholin L., O’Donell A., Fitzgerald N.
Scottish Health Action on Alcohol Problems, 2017
Coinciding with a ‘refresh’ of Scotland’s alcohol strategy, evidence that financial incentives in primary care can affect delivery of brief alcohol interventions.
Summary Local contractual arrangements for brief alcohol interventions are known as ‘local enhanced service contracts’ in Scotland. They typically include fees for the delivery of screening and/or brief interventions, but differ in terms of how payments are structured and the level of remuneration.
As the Scottish Government prepares a ‘refresh’ of its national alcohol strategy, this study provides a timely consideration of the effectiveness of incentivising delivery of brief alcohol interventions in primary care.
Stakeholders say that incentives may help to facilitate the monitoring of delivery in ways that could not otherwise have been achieved, and also formalise or enhance the quality and skill of conversations between patients and GPs about alcohol.
However, local data is limited, meaning local health boards are designing and subsequently modifying their models without evaluating their effectiveness.
Little is known about the availability, depth or quality of data held by local health boards about the delivery of brief alcohol interventions in primary care, and the impact of different models and rates of payment on the implementation of screening and brief interventions. As the Scottish Government prepares a ‘refresh’ of the national alcohol strategy, it is timely to consider the importance and optimum use of local contracts and incentives for the delivery of brief alcohol interventions in primary care.
The featured study aimed to provide an overview of evidence for using financial incentives by conducting:
• A rapid literature review of the design and impact of financial incentives in the delivery of screening and/or brief alcohol interventions in primary care.
• An analysis of financial incentive models in three local health boards, and an assessment of the availability and utility of brief alcohol intervention data.
• Interviews with five key local and national stakeholders on the design and impact of financial incentive models.
The findings are reported under these three strands.
Studies have recorded incentives for screening patients for risky drinking ranged from about £5.11 to £7.67 per patient. Only one study reported specifically on payment per intervention, which ranged from about £11.56 to £21.40.
There was limited evidence on the rate or impact of reimbursement schemes on delivery of screening and brief interventions. Only four studies were identified, of which just two (1 2) reported on payment schemes implemented in routine practice, as opposed to within a research context.
There was some evidence that financial incentives encourage primary care providers to screen patients for risky drinking, and to deliver brief interventions to those at risk. Two studies (1 2) reported a significant rise in screening rates as a result of the introduction of payment schemes. However, there was limited data to assess their impact on the delivery of subsequent advice/interventions for alcohol, or on the effectiveness of such schemes on other important outcomes such as patients’ alcohol consumption.
One study provided a hypothetical assessment of the potential for financial incentives to reduce overall consumption amongst patients when introduced within the Australian primary care context. However, these payments were calculated to be substantially less effective than alternative strategies such as the introduction of a computerised reminder system, and something called ‘academic detailing’ involving educational outreach with physicians to encourage delivery of screening/brief interventions.
There was only one example of a study that tested the impact of a combination of strategies on reported screening and brief alcohol intervention delivery rates. Whilst training/support and financial reimbursement were both effective when introduced in isolation, the greatest impacts were realised via the implementation of multi-component strategies – specifically, a combination of training, support, and financial reimbursement.
The Scottish regions of Lanarkshire, Lothian, and Tayside were selected as case studies based on the level of detail that could be obtained. The data below refers to the years 2012/13 to 2014/15 (unless otherwise stated).
Case study 1: Lanarkshire
Across the three-year period, the number of brief alcohol interventions reported in primary care varied from about 5,000 to 5,700. This represented a delivery rate of approximately nine brief interventions per 1,000 population in 2014/15.
All brief interventions delivered were reported to a central data analyst using a standard form and entered into a database. The form included questions from the Fast Alcohol Screening Test (known as ‘FAST’). A brief intervention did not count towards the allocated target unless delivered by a practitioner who was recorded as having received brief alcohol intervention training (typically a one-day course).
The ‘conversion rate’ from screening to brief interventions – that is, the ratio of interventions delivered to the number of screenings – was 41%, the equivalent of one brief intervention per 2.45 people screened using FAST. In community mental health teams, 16% of screenings were converted into brief interventions, and in antenatal settings the rate was 0.15% (just two brief interventions delivered from 1,333 screenings).
Between 2008 and 2011, GPs were paid a fee for each patient they screened, with no fee in place for the delivery of a brief intervention following screening. This was felt to have resulted (though no figures were available) in a high level of reported screening but a low level of subsequent brief intervention delivery – known as a ‘low conversion rate’.
From 2011 onwards, funding was linked to a new performance framework under which the funding was used to pay for additional community nursing capacity. This was intended to free up community nurses across the area to provide brief interventions, but meant that the payment was for additional nursing capacity only, and there were no payments per individual for screening or intervention delivered.
Case Study 2: Lothian
The number of brief interventions delivered in primary care in Lothian fell between 2012 and 2015, but remained high compared with other areas, and across the three years exceeded the target set for Lothian by a considerable margin (134% of the target in 2014/15). The decrease in brief interventions delivered in primary care over this time coincided with a similar decrease in emergency and antenatal services.
From 2008/09 (up to 2015/2016), a local enhanced service contract set the payment at £30 per brief intervention reported, with no more than one brief intervention per patient reimbursable in one year. In the year 2008/2009 only, a payment of £20 was available for providing a follow-up consultation within 12 months of the delivery of the initial brief intervention, but this was subsequently dropped.
In 2015/2016, the payment system changed from remunerating practices for interventions already delivered, to paying practices up-front based on an allocated number of brief interventions according to the number of patients in the practice. Under this system, money was reclaimed from GP practices which did not deliver their allocated interventions and redistributed to those who had reached their allocation. Available funds for contracts in the year 2016/17 were reduced from £200,000 to £100,000 due to cutbacks. The basic rate of £30 per intervention was still paid, but for a smaller allocated number of brief interventions per practice. The impact of this was not known.
Under the first contract in 2008/09, an engagement fee of £600 was included for infrastructure and also to support the requirement that within the first 12 months of the contract, at least one clinician from each contracted practice had to attend training. A national two-day training course delivered by NHS Lothian was the recommended course. A half-day practice-based course was also offered. Later contracts included information about available training and stated that when an intervention is delivered it is assumed that the practitioner has “a basic professional level of health behaviour change expertise and communications skills (covered by the training programme)”. Unlike in Lanarkshire, no mechanism was in place to require that brief interventions contributing to the reported figures are delivered only by trained practitioners.
Case Study 3: Tayside
Tayside delivered 3,929 brief interventions in primary care in 2014/2015, a rate of approximately 9.5 per 1,000 of the population. Delivery fell in 2014/15 compared with earlier years.
A local enhanced service contract was in place in Tayside from 2008. Early on, a fixed fee of £20 per intervention was paid, without a separate fee for screening. From 2008, a payment of £2.20 was made for each patient screened and £9 per brief intervention delivered. From 2010, £10 was paid per brief intervention, unless practices exceeded their targets, in which case brief interventions over and above the targets would be remunerated at £15 each. Two additional lump sums were included in the contracts: £250 per practice for “audit and concluding work”; and £500 for “engagement and preparation”, and to cover participation in training. Prior to service commencement, at least one named individual from each practice team was required to have completed an approved NHS Tayside short training session on brief interventions, and to have cascaded training to team members within the practice.
All participating practices were required to: develop and maintain a register of patients who were screened using FAST, and identified as hazardous or harmful drinkers; and record all related information in a consistent manner using ‘codes’ specified in the contracts (eg, the code “9k11 Alcohol consumption counselling” was used as a proxy for a brief intervention).
From April 2014, practices were required to record reasons for not delivering brief interventions to patients. They were also given the “opportunity to contribute towards [a] more detailed audit of clinical outcomes at a regional level, to ensure maximum learning from the outcomes associated with this [local enhanced service] at a regional level”. Information from each screening and brief intervention delivered was collected for an audit in 2014. This included gender, date of birth, postcode, and FAST screening score. The audit report revealed the conversion rate for a six-month period in 2014, during which in Tayside as a whole, one brief intervention was delivered for every 12 screenings. This figure masked wide variation across the region, with a 1:7 ratio in Dundee, 1:26 ratio in Perth and Kinross, and 1:29 ratio in Angus.
Practices were asked to follow clinical guidelines from the Scottish Intercollegiate Guidelines Network, and screen patients opportunistically using FAST, but with an explicit focus on patient groups with occupational, social, psychiatric, and physical problems. From 1st April 2009, new patient registration and pregnancy checks were the focus for opportunistic screening.
Practices were advised to deliver brief interventions (usually 3–10 minutes long) within the consultation, to provide written and verbal information about safe levels and patterns of drinking, to provide patients with information on additional support services, and to refer to specialist services “where appropriate”.
Benefits of financial incentives
There was no overall consensus on whether financial incentives for GPs were worthwhile: some argued that they had negative consequences; while others felt that they were necessary and there was a risk that delivery would fall away if financial incentives were withdrawn.
Incentives specified in local contracts enabled professionals to be mandated to participate in training and data collection. This was the case in Lanarkshire, where conditions on funding required payments for community nursing, training, and data collection. Under the local enhanced service contracts in some areas, GP practices were also required to participate in audits, and provide data to the health board about delivery. It was suggested that this benefit may only be possible with a local enhanced contract, and would not be easy to mandate if delivery of brief alcohol interventions was included in a national GP contract.
Optimal models and levels of incentive
Participants discussed the different models of remuneration in place in primary care, including separate payments for individual components of brief interventions (screening, brief intervention, and follow-up). Amongst those who supported incentives, there was a consensus that it was better to have separate payments for screenings and brief interventions, than payments for delivery of the ‘whole package’.
There were divergent views on what the ‘right’ conversion rate would be. One participant suggested that a 1:3 or 1:4 ratio of screenings to brief interventions was indicative of GPs “perhaps subconsciously pre-selecting patients they might screen” who are more likely to screen positive, whereas 1:25 indicated that the alcohol screening had been “included in a suite of screening that they’re doing” or based on new patient registrations. It was not clear on what evidence the judgements were being made about conversion rates. However, participants reported that (in line with the national guidance) the local contracts were not intended to incentivise universal screening.
In addition to incentives, participants discussed the importance of other aspects of the implementation effort in securing the delivery of brief interventions, including:
• providing practice-based training;
• involving a GP champion in supporting practices and resolving difficulties;
• ensuring practices have the materials and resources needed for brief intervention delivery;
• the value of having a history of brief intervention work initiated prior to the national programme;
• the importance of local funding from the national brief intervention programme;
• setting delivery targets for each local area within a health board;
• commitment from management and other staff.
Unintended or negative aspects of incentives
Two participants felt that incentivising GPs to deliver brief interventions was not the best use of funding, and argued instead that other staff within primary care, such as nurses, were better placed to screen and deliver brief interventions. One of them, a GP, felt that GPs should focus their time on “chronically ill patients”, and on the “parts of the job that only [GPs] can do”. This participant said that GPs used their professional judgement to deliver less formalised brief interventions anyway as part of a normal consultation, and that the incentives unnecessarily formalised the way in which interventions were delivered. The other participant said that nurses sometimes delivered the interventions under the local enhanced service contract, and that the funding arrangement offered poorer value for money than investing in community nursing.
The main risk of incentives noted by participants was that people might be “just kind of ticking a box to get the money”, but this was not felt to have been happening on a large scale, and appeared to be more of a risk where payment was made for screening, and especially if only for screening.
One participant saw the lack of national co-ordination or evaluation of the local arrangements in primary care as a missed opportunity, and described in detail how it could have been done differently to learn for the future: “If I was starting again I would have … I would try and have either different models in different areas or a uniform model across the country and try and kind of compare and contrast and do it that way and then have a much more closer scrutiny of it over time to see how it’s worked and evaluate it rather than what happened, just leave everybody to their own devices and then try and scrap for information after that to see if it worked and not really fully understand what went on.”
The lack of local data from the featured study indicates that health boards had designed and subsequently modified their funding models without evaluating their effectiveness. Interviews with local and national stakeholders revealed that at least some decisions to change remuneration systems were based on a combination of local intelligence gathering and some shared learning, but assumptions about the likely impact of such changes on outcomes were not tested.
Benefits of incentives were perceived to include facilitating the monitoring of delivery in ways that could not otherwise have been achieved, and formalising or enhancing the quality and skill in conversations about drinking that may have already been taking place.
There was a general consensus that paying separately for both screening and intervention delivery was better than paying only for one or the other, but different views on the optimal ‘conversion rate’ of screenings to brief interventions. In practice, the so-called conversion rate varied by area and type of practitioner – for example, a 1:2.5 ratio of interventions to screenings delivered in Tayside, where delivery was largely by GPs, versus 1:12 in Lanarkshire, where delivery was led by community nurses. This could perhaps be indicative of differences in levels of drinking problems amongst the population accessing the service, the willingness of the population to disclose such problems to different practitioners, and the choice of patients being identified for screening.
UK studies published since the featured review have been supportive of incentives as an implementation strategy (1 2). Another recent study based in different locations across Europe (England, Catalonia, the Netherlands, Poland, and Sweden) also found that incentives and training had increased the reported delivery of screening by primary care doctors, though of course this could have been a reflection of improved recording and reporting of conversations that were already happening.
The commitment to brief interventions in Scotland thus far is rare in scale and scope globally, yet has not been matched by a significant contribution to the wider evidence base. In the authors’ own words:
“The failure to contribute to longstanding research questions about alcohol brief intervention implementation, quality and outcomes in primary care is a serious and abiding weakness of the otherwise widely-admired Scottish national alcohol programme.”
The refreshment of the national alcohol strategy presents an opportunity to address this ‘weakness’, and inform future alcohol policy in Scotland.
commentary The featured study, funded by Scottish Health Action on Alcohol Problems, found evidence that delivery of brief interventions in primary care was affected by financial incentives. In one case study area when screenings were incentivised, screenings were abundant but few patients received a brief intervention. When the system changed to target the intervention, an unusually high 41% of recorded screenings were followed by a brief intervention. Interviewees said money was not the only factor, but it was a key one. In a companion study, interviews with 13 GPs made clear that the biggest barrier to an effective brief intervention was the lack of time in a consultation lasting barely 10 minutes, and which was primarily for addressing the issues the patient brought to the consultation.
National policy in Scotland to date has prioritised screening and brief interventions in three key settings: primary care; antenatal care; and accident and emergency departments. Targets set for 2008/09–2010/11 to deliver 149,449 brief interventions were exceeded; over the three-year period, 174,205 brief alcohol interventions were recorded across the three priority settings. Most of the work has taken place in primary care, where a study found leverage and acceptance to be greatest. One such recognised lever is the one investigated by the featured study – specific funding through local enhanced contracts to incentivise primary care practices.
An Effectiveness Bank hot topic considers the potential for brief alcohol interventions to improve health population-wide, including the extent to which benefits in research can translate into routine practice.
Last revised 23 October 2017. First uploaded 10 September 2017
STUDY 2016 Improving the delivery of brief interventions for heavy drinking in primary health care: outcome results of the Optimizing Delivery of Health Care Intervention (ODHIN) five-country cluster randomized factorial trial