The British Journal of Psychiatry
Care clusters and mental health Payment by Results
David G. Kingdon, Bohdan Solomka, Hamish McAllister-Williams, Douglas Turkington, Alain Gregoire, Hesham Elnazer, Mahesh Thagadur, Lars Hansen, Shanaya Rathod, Stefan Gleeson, Mo Zoha, Pratima Singh, Farooq Naeem

MacDonald & Elphick1 have lucidly described the proposed introduction of Payment by Results into mental health. They mention, however, that ‘concerns include the validity and reliability of the MHCT’ (mental health clustering tool), and there is also the major issue of how cost can be firmly linked to the quality of services delivered.

The Department of Health approach to reliability has been to rely on local initiatives and to commission the development of an algorithm based on the MHCT to ensure that clusters are reliably allocated. Exactly why the Department of Health believes that this is possible by either route is not clear. Local initiatives are the route to mayhem and none of the attempts at devising algorithms so far have been successful. The instrument on which the MHCT is based, Health of the Nation Outcome Scales (HoNOS), was not devised for this purpose. Additional items have been added but this was for clinical reasons. The HoNOS was devised as a clinical outcome measure, not for needs assessment and certainly not as a classification tool. Internationally recognised tools (e.g. Schedules for Clinical Assessment in Neuropsychiatry, Structured Clinical Interview for DSM Disorders) have been devised to classify conditions but these use a range of information (e.g. symptoms, beliefs and timescales), with specified criteria that have been and continue to be subject to international expert scrutiny.

A unit of costing must be directly related to quality and outcome measures or the UK will have the same problems as the USA have encountered in its payment systems. It is difficult to understand how clusters can be such units of cost unless there is a very substantial body of research investigating evidence for efficacy of interventions (e.g. cognitive therapy and medication) for individual clusters, and for the development and reliability testing of outcome measures – which would take years. Attempting to match cluster to pathway/intervention has to be done by using diagnosis as an intermediate step because that is where the evidence currently exists. The problem then is that each cluster relates to a number of guidelines and monitoring quality becomes complicated – as trusts, and previously the Department of Health, are finding in attempting to devise cluster pathways. General practice commissioners won’t have the time, resource or inclination to undertake such complex monitoring – so cost will rule.

Broad diagnoses, as used by the National Institute for Health and Clinical Excellence, have proved satisfactory for clinical purposes and have readily available, reliable and applicable outcome measures2 and, although diagnosis alone is not sufficient for costing, in combination with clinical pathways3 they can be costed and used for tariffs with a much better chance of reliability and homogeneity. The very limited data that have been produced so far are promising (available on request) but there needs to be more extensive examination of data. The Department of Health needs to take a lead because trusts are not going to re-analyse their data using diagnosis and allocation to pathways unless the Department of Health asks them to do so.

As MacDonald & Elphick1 describe, outcome measurement is needed in any system and clustering has been very effective at promoting use of HoNOS. However, combining diagnosis and pathways could provide a simpler, practical approach to gathering data for costing and tracking change than can making use of clusters. It would also lead to an improved quality of care by linking cost directly to the use of evidence-based guidelines and care pathways by empowered patients, carers, providers and commissioners.