Payment by results was introduced by the Department of Health in a publication Reforming Financial Flows (2002). The government plans to implement it throughout the health service, including in mental health services, although specialist mental health services are likely to remain as they are, at least for the time being. Payment by results already applies to some acute hospital trusts (general hospitals). In mental health, it will replace the existing ‘block contract payment’ arrangements for services (Reference FairbairnFairbairn, 2007), possibly in 2009, although there are some concerns that the implementation will be difficult. What is clear however, is that this new system of paying for healthcare services and procedures will have a major influence on how mental health services are commissioned and financed (Reference FairbairnFairbairn, 2007).

Payment by results – rudiments

Payment by results will see the introduction of nationally agreed tariffs for treatment and services based on average costs throughout the National Health Service (NHS; Box 1), but there will be some flexibility to take local cost variations into account. It is thought that by having this arrangement there will in theory be less need to negotiate on price, as this is fixed and predetermined, and more room for negotiation on quality and how best the services can meet the needs of the local population.

Tariffs will be calculated by comparing similar cases, probably based on level of need and resource use and loosely on diagnosis. Groupings of cases in the acute hospital sector are based on diagnosis (these clusters are known as healthcare resource groups). In mental health, however, classifying level of need and resource use is more complicated. There are very few textbook cases or examples of illnesses or treatments, and comorbidity and complications are frequent. Individuals are often treated both as an in-patient and in the community and also by different services and agencies during the same treatment episode. Traditionally, far less focus is applied to quantifying the cost of community care (Department of Health, 2007). There is also wide variation in clinical practice and available services between different regions of the UK. Any such coding would have to take these issues into account.

In mental health, health benefit groups will be used instead of healthcare resource groups. ‘Case-mix funding’ is a term sometimes used to describe the situation where a combination of factors influence cost (Information Centre for Health and Social Care, 2006), for example dual diagnosis or comorbidity. It is hoped that case-mix funding will apply to mental health services.

Activity-based funding implies that more ‘activity’ leads to greater funding. This is in contrast to the block contract agreements that already exist between primary care trusts commissioning services and local service providers. For example, a primary care trust commissioning medium secure psychiatric beds will usually buy a set number of beds per year, regardless of how many they will actually use. Thus, activity-based funding may seem fairer, at least on the surface. However, strict application of activity-based funding may be less appropriate for mental health services (Department of Health, 2008a ), particularly when one considers the wide variation in resource use by people with the same mental health diagnosis who may have different social circumstances and different needs not necessarily caused by their mental health problems.

Developing national tariffs for mental health problems

A key issue for mental health and payment by results is that in many cases diagnosis and severity of the illness do not predict resource use accurately. Other factors such as social support, accommodation, area of residence and marital status are also important (Reference Oyebode, Cumella and GardenOyebode et al, 1990). Treatment is often shared between different agencies, leading to complicated care pathways and it is also sometimes provided informally by non-professionals (Sainsbury Centre for Mental Health, 2004). Support from friends and family might, for instance, have a significant effect on a person's recovery, but this would not be measured.

Resource use can be related to behaviour and incidents which are sometimes independent of mental illness. Violence, unrelated to mental health problems and directed towards a staff member, carer or patient resulting in personal injury, time off work for the victim and increased levels of observation is an illustrative example. It is important that these other factors are taken into consideration. Tariffs or currencies based solely on diagnosis or severity of psychopathology and not other behaviours will be inaccurate if they do not reflect these other costs.

Developing tariffs or currencies for payment by results in mental health is clearly a challenging task. The Department of Health commissioned the Health and Social Care Information Centre (HSCIC) with the remit of developing the tariffs or currencies (Reference FairbairnFairbairn, 2007). The HSCIC then set up the Care Packages and Pathways Project which involves a number of mental health trusts in England (Department of Health, 2008b ) These organisations collected data on their patients using a needs assessment tool, the minimum data-set and a modified Health of the Nation Outcome Score (HoNOS) (Reference FairbairnFairbairn, 2007). The project is ongoing (Department of Health, 2008b ). However, there is limited evidence on how closely HoNOS ratings correlate with resource use and factors such as various therapies, medication and extra staffing for levels of observation. These issues are relevant when calculating resource use and tariffs for diagnoses.

The Care Packages and Pathways Project aimed to gather information on 10 000 in-patient treatment episodes and 200 000 non-in-patient contacts from various mental health trusts across England. The information was to be categorised in groups reflecting resource use, level of need and diagnosis. The programme eventually collected information on over 54 000 service users and 1.2 million service user contacts, although the vast majority of the data was of poor quality and only 22% could be used to develop the tariffs or currencies. Approximately 54% of service users were female and 46% male, with 81% White, 4% Asian and 2% Black (Information Centre for Health and Social Care, 2006). However, given that only 22% of the data could be used, it may not be representative and ultimately questions the tariff's or currency's utility.

The tariffs or currencies will be calculated by drawing data from a comprehensive clinical assessment of these cases and will include: psychiatric symptoms, ‘strong unreasonable beliefs’, self-harm, depression and suicide, substance misuse, aggressive behaviour and engagement. These results will be combined with a modified HoNOS. It is hoped then that the final results will form clusters which correspond to resource use. Each cluster will then be assigned a code corresponding to a cost (Information Centre for Health and Social Care, 2006).

Implications of payment by results for mental health trusts

The very concept of payment by results arouses degrees of pessimism and perhaps resentment among mental health services (Sainsbury Centre for Mental Health, 2006). In their survey of financial directors of English mental health trusts, the Sainsbury Centre for Mental Health found that of those that responded (42% of the total) 68% believed that the introduction of payment by results in acute hospital trusts had had a second order effect of diverting funding away from mental health services. This was despite the fact that in general, compared with acute hospital trusts, mental health trusts appeared to have managed their finances more efficiently in 2005/6 (Sainsbury Centre for Mental Health, 2006).

Despite presenting genuine concerns and risks, there is some evidence to suggest that payment by results could work (Sainsbury Centre for Mental Health, 2004). A broad, lateral but cautious approach to coding and developing tariffs, subject to independent audit and monitoring, might lead to a fairer and more accurate system for those commissioning services. Commissioners will be able to source high-quality services from a variety of providers. Efficiency, activity and potentially quality from providers should therefore also grow, at least in the short term. Service providers that flourish under payment by results might increase in size and power, some of them potentially becoming ‘too powerful’. They could spread and monopolise services in certain areas.

It is important that there will be a greater emphasis on measuring quality of service and activity so that patient care and standards of treatment are not jeopardised by a focus or desire to increase activity. Mental health trusts may become increasingly dependent on evidence-based and cost-effective protocols and guidelines.

A clear benefit to implementing payment by results to mental health services will be to make the daily workings of service providers more transparent to the public and patients (Reference FairbairnFairbairn, 2007). Unrecorded activity is not going to be paid for and record keeping, an area notoriously under par in mental health services, will have to improve.

Payment by results supports the decentralisation of NHS budgets (Reference FairbairnFairbairn, 2007). Given that most of the money is kept by primary care trusts to commission services whose budgets are set for 3 years at a time, the trusts will have more power in planning future services more appropriate to the needs of the population they serve.

I have mentioned some concerns about how representative the tariffs or currencies will be. Even if they are accurate, there may still be problems applying them. Some service users may become more ‘attractive’ to treat than others – for example, they could see themselves ‘up coded’ by trusts to a more expensive tariff so that service providers could charge more from those commissioning the service. This is known in the acute hospital sector as ‘diagnosis-related group (DRG) creep’ (Sainsbury Centre for Mental Health, 2004). Trusts could also potentially add additional and superfluous treatments or interventions leading to higher tariffs and inflated costs.

Conversely, trusts may find themselves under greater pressure from activity-based funding to get patients treated more quickly and discharged from in-patient settings or case-loads as early as possible. Inevitably, in some cases this might be premature leading to potentially greater suffering and risk to the patient and sometimes also to the community (Table 1).

Table 1. Potential advantages and disadvantages of payment by results and mental health services

Advantages	Disadvantages
Better value for money — service providers should be more active and efficient	How accurate and representative will the tariffs be? Tariffs are evolving from a limited evidence base
Improvements in quality — encourages competition between providers and greater emphasis on performance management and evidence-based medicine	Quality could worsen — early discharges and cheaper treatments to reduce unit costs and increase activity levels may compromise quality
De-centralisation — local primary care trusts will have more power in commissioning appropriate services for the local area	Abuse of the system — some patients may be more ‘attractive’ to treat than others as they could generate more money for their treatment
Transparency — daily working of services to be made more accessible to service users and the public	Resentment from staff — morale may worsen when services have to explain and measure levels of activity
Better record keeping — providers will have to improve in this area; unrecorded activity will not be paid for	Local monopolies — some providers may become so powerful that others are squeezed out, leading eventually to less choice for commissioners when sourcing services

The future

It seems that the Department of Health recognises that introducing payment by results to mental health trusts will be challenging. The development of tariffs for mental health problems is complicated, far more so than in the medical and surgical sectors and relies on high-quality and representative data. Concerns remain about how representative these data are and how closely they will correlate with resource use. There is a real risk that inaccurate tariffs will be generated.

Whether successful or not, payment by results will have a radical impact on financing and far-reaching consequences for mental health trusts. Some trusts will do well as others struggle and it is likely that the trusts that benefit and thrive with payment by results, will already have robust and organised information and cost management systems, and work with high-quality and efficient service providers. Across the board, we will see service providers forced to increase and closely monitor their levels of activity, or reduce their unit costs to survive.