Differences in diagnosis and treatment mean there are special challenges in developing payment by results in mental health. Now a group of trusts in the North are pioneering a 'whole life' approach which could hold the solution. Emma Dent reports

The history of payment by results style funding systems in mental health is neither straightforward nor successful - at least not yet.

When PbR was introduced in the acute sector, the Department of Health said repeating the exercise in mental health would take longer. To begin with it excluded the entire sector from the system.

A number of health services in other countries, such as New Zealand and Australia, have attempted introducing something similar to PbR, and all have failed. Until recently this has been taken as an indication that such an outcome would be inevitable in the UK too.

Everyone accepted that the sector would need an entirely different approach; managing someone's mental health is not the same as fixing a broken leg. What could the approach be?

In December 2004 the DoH gathered together a number of mental health trusts and told them that work was to begin on developing a case-mix, or healthcare resource group, approach to allocate resources based on diagnosis. The work was based loosely on what had been attempted earlier in New Zealand but intended to build a bespoke model.

'A number of trusts pulled out there and then,' recalls Tees, Esk and Wear Valleys trust finance director Colin Martin. 'In mental health you can rarely diagnose someone when they walk through the door and things can always change later.'

Mr Martin's former trust, Tees and North East Yorkshire Mental Health trust, was of among the 20 that did initially participate but numbers dwindled until only a handful were left by last year.

Participating trusts needed to do a data collection exercise to see if service users could be grouped together by diagnosis. It was quickly found that many organisations' information systems were simply not up to the job of carrying out the necessary data collection, reflecting the long-standing sore point of poor quality information systems in mental health. Data often had to be collected manually, which was time consuming, unwieldy and led to staff resentment.

New lease of life

The project lost impetus and its future looked uncertain until six months ago when the DoH Information Centre for health and social care brought new direction to the work. But this revival did not happen quickly enough to stop a group of trusts in the North banding together to pursue their own method due to concerns that the DoH's approach was unworkable.

Several of the trusts were working on their own care pathway systems, but decided to develop the work carried out in a project by South West Yorkshire Mental Health trust in Kirklees on what has been tagged an 'integrated packages approach to care'.

The Kirklees work has developed 13 clusters based on needs for adults and older people. They range from short-term, non-severe problems in cluster one through to assertive outreach-style patients, who have long-term problems and are difficult to engage, in cluster 13 (see below).

What marks the model apart from an HRG-style approach is that it aims to include a variety of interventions, from medications to therapy to reducing the social isolation of a patient or sorting out their problems with finding a place to live.

Supporters say that as mental health services have grown and adapted to take a 'whole life' approach, including a wide range of community-based services and social interventions, any funding system should reflect that.

'We are hoping the data that is produced through the pilot will support the integrated packages approach and demonstrate fully what it can do,' says South West Yorkshire Mental Health trust associate director of psychology Dr Roland Self. One of the leading clinicians behind the northern project, Dr Self describes the integrated packages scheme as a whole systems approach.

For him, the development of a funding system is almost a sideline to the work that is going on to improve care. The care packages approach work at the trust will go on regardless of payment by results in mental health succeeding or not.

Each of the six trusts in the northern project agreed to collect data in at least one inpatient unit and one community-based service. Over six months they have recorded 70,000 episodes on about 14,500 patients.

Carole Green, project director for payment by results in the National Institute for Mental Health in England's north east Yorkshire and Humber region, is leading the project. 'Some of the recorded patients started and ended having treatment within that period, some were receiving treatment before the pilot started and some started receiving treatment and still are,' she says. 'We have a wide range of data.'

Ms Green is now working all but single-handedly (she has the part-time support of a PhD student) to clean the data so that it can be sent to the Information Centre for it to compare it with its own work.

'We need to see what kind of variation there is and what the best care packages are,' she says, 'and that what happens in Bridlington could happen in Bournemouth and Birmingham.'

The work has uncovered patients who do not need specialist services and should actually be getting treated in primary care. That the clusters help commissioners and clinicians demonstrate this could explain why commissioners are showing an interest in the work and clinicians have bought into it. A tariff also has the potential to show providers where they have plurality of care provision and could be more cost effective.

The northern pilot has also confirmed huge variations in cost and spending around mental health but, promisingly, the Kirklees model suggested that 85 per cent of service users will be able to be accommodated in the 13 clusters, a conclusion Mr Martin concurs with.

He admits that there will clearly be some service users who will not fit in the clusters but adds that there are always those whose circumstances are unique to the point where their care pathway does not correspond to any other and has to be written from scratch for them.

'The 13 clusters may have to be expanded to 14 or 15, and they will need development. We can see that from the acute model,' he says.

Class action

The Information Centre has had 22 trusts - including the six also pursuing their own model - collecting data based on the health of the nation outcome scores (HONOS), a standard way of assessing the mental health of service users. The data has been examined for HONOS scores and diagnosis to estimate what kind of groups service users could be put in based on the amount of resource allocation they need.

'We want to see if we can develop a classification system but first we have to see what the data looks like,' says Michael Bewell, case-mix service project manager for mental health classification at the Information Centre.

It has collected around 14,000 patient records on care for adults and older people. Child and adolescent mental health services, forensic and secure services including prisons, and learning disability and substance misuse services, are excluded. In the latter category this is a marked difference from the northern trusts' model, which includes a cluster for drug and alcohol misuse.

'The data used was already in existence and has been taken from the minimum mental health data set [information that all trusts have to record on their activity]. We want a system that will really represent what the patient actually sees, what actually happens and is clinically appropriate,' says Mr Bewell.

South London and the Maudsley trust chief executive Stuart Bell has involved his trust in the work around PbR for some years. The DoH policy and advisory mental health payment by results expert reference panel chair he says his principle concern with any system is that it improves care.

He says it is vital that no funding scheme incentivises admitting patients.

'It is very important to find a way to get a sensible tariff-based system in place, irrespective of how it has been produced. We need a system that goes with the grain of mental health policy, that encourages support to be given to people when they need it, without necessarily having incentives to admit,' he says.

'Any scheme should encourage people to live as independently as possible, with admissions only taking place if they are absolutely needed.'

Mr Bell believes a tariff could have scope to incentivise certain treatments and improve access to poorly provided treatments such as talking therapies. However he cautions against payments being able to be made against every intervention possible: 'To attempt that would be a complete nightmare,' he says.

Data analysis of the work carried out by both the northern trusts and the Information Centre's case-mix team should be complete by the end of October, and will then be submitted to the DoH for a decision on what to do next. Consultations with stakeholders, likely to involve service users and carers, are expected to begin in November.

Some close to the two schemes predict it will come down to a battle between the two methodologies, while others believe it is possible that any future system may learn from both.

One senior manager close to the work says he fears that it may all come to down to the cost of work, based on diagnosis, with less emphasis on activities carried out. This could put at risk any potential to have an emphasis on need.

'The politics behind it all say that we must be seen to be doing something, but the danger is it could all be messed up completely and we could end up with something like the acute system when mental health needs something completely different,' he says. He adds that the amount of knowledge in the mental health sector that is based on diagnostic skills could mean some will be reluctant to work with a system that does not take diagnosis as its starting point.

Careful progress

Mr Martin and Mr Bell agree that a pilot or initial phase could at first involve elective-style services, such as eating disorders or talking therapies. An embryonic working system could emerge in 2007-08, perhaps with a shadow project to see how a more extensive system could work in practice.

But many are cautioning against an overly ambitious timetable. Ms Green suggests it will be early 2007 before there is an indication of what will happen next, followed by a year of testing.

Mr Bewell agrees that any system will have to evolve to be fully workable, adding that regardless of the ultimate outcome work done on data classification has been successful.

'In the earlier programme there was not enough explanation around what was being done with trusts. We have done that now. There are a lot of people excited about what is going on,' he says.

'We are seeing trusts realise what it could mean for them, in terms of knowing what work they do, and preparing for foundation trust status.' Any successful foundation applicant must be able to prove to regulator Monitor that there is a clear relationship between their activity and funding.

Getting staff on board has been vital for both projects, and data collection continues to be an issue for any system to succeed. Those involved in the projects say it could be a carrot approach to encourage trusts to improve their IT and data collection.

'The information needed can usually be taken from care pathway approaches, but it is about joining that up to get what is needed. Any system could act as an incentive to improve their data collection; this work has moved data collection up the agenda and staff have felt value in doing it,' says Mr Martin. 'Our confidence is growing and we and some other trusts intend to have in place a national IT programme solution that will boost confidence again.'

Terminology could also play an important role in how staff perceive any ultimate solution - the concept of PbR does not sit happily with mental health services.

'We are not about fixing knees and legs; calling it payment by results could mislead people,' says Mr Martin.

Any new system will have to be clearly explained to stakeholders to avoid unfavourable comparisons with the acute sector. Mr Martin suggests that terms have been used overseas, such as reimbursement, could be more acceptable.

Whatever the name, he insists it is time for mental health to have a tariff system, if only to counter what many perceive to be a diversion of funds to the acute sector as a result of PbR.

'The sector has felt the squeeze with PbR being introduced into the acute sector and with the current financial climate,' he says. 'Mental health needs to be out there.'

The northern model : who is involved

Formerly seven trusts, following a merger there now are six trusts working on a needs-led cluster model for a payment system.

They are:

  • South West Yorkshire Mental Health trust
  • Humber Mental Health Teaching trust
  • Doncaster and South Humber Healthcare trust
  • Tees, Esk and Wear Valleys trust
  • Northumberland, Tyne and Wear trust
  • Leeds Mental Health Teaching trust

Lucky 13 - the clusters

1 Acute non-psychotic (low severity)

Likely to be short-term, mild problems of mood, anxiety with no psychotic symptoms.

2 Acute non-psychotic (medium severity)

Low risk, characterised by moderate amounts of depression or anxiety.

3 Non-psychotic (high severity)

Severely depressed or anxious, will not present with hallucinations or delusions but may have some unreasonable beliefs and often a high risk of suicide.

4 Non-psychotic disorder of over-valued ideas

Moderate-to-severe disorders such as eating disorders and obsessive compulsive disorders, where extreme beliefs are strongly held. Likely to affect functioning in many ways.

5 Non-psychotic chaotic and challenging disorder

Wide range of symptoms, chaotic and challenging lifestyles, characterised by self-harm and over-dependent engagement with services.

6 Drug and alcohol

Misuse of drug and alcohol, may have some problems with anxiety or depression but will not have psychotic symptoms.

7 First episode psychosis

Presenting to services for the first or second time with mild to moderate psychotic symptoms, may also have depressed mood and/or anxiety or other behaviours.

8 Chronic severe mental illness (low symptoms)

History of psychotic symptoms currently controlled and causing minor problems if any. Likely to be vulnerable to abuse or exploitation and may be suffering problems linked to old age, long-term illness and prolonged medication use.

9 Chronic severe mental illness (high symptoms)

This group will have been known to the service for a long time and be in receipt of ongoing support from the service, with moderate psychotic symptoms and likely to have poor role functioning.

10 Severe psychotic episode

Likely to have been known to the service for a long time, experiencing severe symptoms that cause severe disruption to role functioning. May be at risk to others or themselves.

11 Severe depression

Acute episode of moderate to severe depressive symptoms, likely to present a risk of suicide and have disruption in many areas of their lives.

12 Dual diagnosis

Chronic, moderate to severe psychotic symptoms with unstable, chaotic lives and drug or alcohol misuse. May present a risk to others and poorly engage with services.

13 Assertive outreach

Severe psychotic symptoms with unstable, chaotic lifestyles, likely to be non-compliant, vulnerable and engage poorly with services. May present a risk due to violent behaviour.