Stemming the referral tide
We look at the options for CCGs in tackling the bigger, more challenging commissioning areas. This month we look at controlling referrals
GPs made nearly 12 million referrals last year, according to the latest figures from the Department of Health – a 4.4% increase on the previous year. Every referral triggers thousands of pounds' worth of spending in secondary care, a total cost to the taxpayer of £15bn according to 2010 King's Fund figures. Referrals are a product of GPs' individual decisions and experience, making them a highly variable and unpredictable area of NHS expenditure. It's also unclear what is causing the increase in referrals.
A BMA analysis in 2009 identified dozens of possible factors, including:
- better data collection because of payment by results
- inaccurate hospital coding, either through error or gaming the system
- increased supply through Choose and Book
- better primary care diagnostics
- QOF incentives for referral
- fragmentation of care because of increased use of salaried GPs and locums
- improved patient awareness of treatment options.
Consultant-to-consultant referrals are another contributing factor. A 2009 report from consultancy CHKS identified a five-year trend towards a disproportionate increase in numbers of referrals between consultants.1
Referral management centres
A major report into referrals carried out by the King's Fund last year2 looked at a referral management centre that had been set up by a CCG and taken over by a PCT.
The report concluded the strengths of such centres are that they:
- filter out inappropriate referrals
- direct referrals to the most appropriate setting
- fast-track referrals when appropriate
- improve the quality of referral letters
- develop a body of expertise about local services
- provide evidence to support commissioning decisions.
However, the King's Fund report warned the centres could:
- increase costs, with centres costing up to £23 per referral
- demotivate local GPs
- misdirect referrals in the absence of full information
- create barriers to closer working between GPs and consultants
- lose or delay referrals in the absence of robust governance.
A study presented at the Society for Academy of Primary Care (SAPC) in July3 concluded referral management centres made no difference to the quality of GP referrals in terms of accuracy or appropriateness.
Clinical assessment and treatment centres
The King's Fund report also examined a clinical assessment and treatment service (CATS) run by a CCG.
It found such centres could:
- direct referrals to the most appropriate setting
- make services more accessible.
However, CATS could also:
- increase overall costs – if anything by more than referral management centres
- misdirect referrals
- delay access to a specialist.
Peer review and feedback
These methods were found by the King's Fund study to:
- increase the likelihood of GPs referring when necessary
- improve the quality of referral letters
- increase the likelihood of GPs directing referrals to the most appropriate setting.
A study published earlier this year at the SAPC conference looked at 10 practices in Wales4 that participated in incentivised peer review with consultants as well as GP peers.
Variation in individual GP referral rates changed from between 2.6 and 7.7 per 1,000 patients per quarter, to between 3.0 and 6.5 per 1,000 per quarter after implementation of the system.
The same study found that once the interventions stopped, GPs went back to their old ways with referrals increasing by 40% for practices that left the scheme.
NICE launched a database last year offering GPs advice on referring to secondary care. It highlights NICE guidance with the aim of improving clinical outcomes and patient experience, as well as saving money. But the King's Fund report said such initiatives did not change GP behaviour.
Financial incentives could change referral behaviour, but might reduce appropriate referrals as well as inappropriate ones and create a conflict of interest for the GP, the report warned.
The King's Fund found that although half of the 22 PCTs studied believed their referral management schemes had curtailed referrals, PCTs with active referral management were in fact no more likely to curtail demand than other PCTs.
A collaborative approach
As the baton for controlling GP referrals passes to CCGs, they must decide whether to continue what PCTs have started – with controversial referral management centres or incentive payments – or find new solutions. Dr Martyn Diaper, a GP in Winchester and member of the primary care team at the NHS Institute for Innovation and Improvement, believes the future solution lies in a collaborative approach rather than the ‘them and us' relationship that often pervaded with PCTs and GPs.
‘You need to be able to discuss openly why Dr A is referring four times more than other GPs. Have we got something to learn from Dr A or has Dr A got something to learn from the rest?' he asks. ‘The institute believes the way to tackle this is by creating a context where everyone can get on board around a common language.'
Chris Naylor, a senior researcher at the King's Fund and co-author of its report on referral management, agrees that CCGs will lend themselves to a collaborative rather than adversarial approach to referrals, and expects to see peer review and education form an increasing part of the process.
‘Whether referral management works post reform depends very much on how the systems are set up. If it is GP owned and led, we believe their chances of success are higher. If they're not and they just see their role as batting referrals back, they won't work,' he said.
1 Robinson P. Consultant to consultant referrals. CHKS. February 2009
2 Imison C and Naylor C. Referral management: lessons for success. The King's Fund, 2010
3 SAPC 2011 conference, poster 1.61
4 SAPC 2011 conference, abstract book
Case study: Peer-led referral management centre
Dr Peter Devlin, joint clinical director of Brighton and Hove Integrated Care Service
All referrals from all GPs in the 48 practices in Brighton and Hove come through a central point and are subject to peer review by GPs. We have a pool of 25 GP triagers, and also use extended-scope practitioner physiotherapists, GPSIs and a small number of consultants and specialist nurses.
There are two steps. First, all referrals are recorded, which builds up a highly detailed database about demand broken down into specialty and referring condition. Second, all referrals are peer-reviewed, which gives real-time feedback to referring doctors.
Ultimately the decision is down to the referring doctor – there will be a conversation, but we would never block a referral against the referrer's wishes.
One aspect to consider is how many community services are in place that you can stream appropriate patients to. But we have found that there are specific areas GPs find more difficult than others – for example, digestive diseases and ENT. It may be that in the future we move to a more targeted system.
In 2010/11, referrals nationally went up by 6%, but in Brighton and Hove we saw a drop of 2.2%. In April to June this year, we processed 16,486 referrals and in that time received only four complaints about clinical triage decisions and four about the administration process, only two of which were from GPs.
Our service does deliver a budget surplus overall, but more important than that it's the additional value given by the ability to give live feedback, offer learning and development to GPs and rapidly embed new clinical pathways.
Case study Referral balanced scorecards
Dr Sam Everington, chair of NHS Tower Hamlets CCG
In Tower Hamlets, we have had a data-sharing agreement with GPs in place for a long time. We have also split the 36 practices into eight confederations so they are very used to looking at data.
More recently we have been focusing on referral balanced scorecards, or dashboards, which are sent to practices on a monthly basis. We use as many data sources as possible, but it is largely based on SUS and Choose and Book data.
It is mainly done at the practice level, but some bits can go down to GP level. We're currently focusing on musculoskeletal, dermatology, urology and ENT, which are all areas where there are good alternative pathways in place.
The GP sees a dashboard for each specialty, which shows their position for GP-referred activity per 1,000 patients. Then we show them first outpatient referrals this year compared with last year so they can see if they have a similar level of activity. They can also see consultant-to-consultant activity as well, so they can be aware of what issues there might be, and finally follow-up activity.
They also get a line graph showing the monthly trend and how they are performing against the PCT average and within their confederated networks.
It is very much a developmental process, not a performance-management one. Practices and GPs can look at where they are against their colleagues.
It is about getting people to reflect on their referrals – improving quality, not just reducing referrals. I think this is what every CCG needs to do. All of us will be bottom of something so it is not about penalising. Its strength is in its simplicity.
Case study Referral facilitation software
Dr James Findlay, clinical governor at Nene Commissioning
After initially developing an Excel-based system for guiding GP decision-making on referrals, we realised we needed to be able to upload data more frequently and moved to a web-based programme.
Last year, we linked up with software company Plain Healthcare to improve functionality of the website – known as PathfinderRF – to be able to offer it to other commissioning groups.
The system enables the user to look up the type of referral and seek guidance on best practice. It gives them a form to fill in that can be copied into their clinical system and attached to Choose and Book or emailed to a specialist.
The advantage to the GP is if there is a specific reason they should not refer, it will all be listed. And it details the provider's criteria for a specific service – for instance, chest pain or specifications on who would not be suitable. It lists tips for diagnosis and medication and links to other useful websites. It is a one-stop shop that helps the referrer make the right decision while the patient is in front of them.
We really wanted to do that rather than have a referral centre where patients might be bounced back a few weeks later with no idea why.
For referral volume, the increase in 2008/9 was only 3%, compared with 13% in neighbouring PCTs. In an anonymous survey we found 50% of GPs rated its usefulness as five out of five, and 73% four out of five.
We will be running a forum to share ideas between CCGs using the system.
It is certainly very cost-effective, and it helps by telling you what to do, not what not to do.