In England, the focus of Department of Health attention has shifted from compliance with centrally endorsed processes towards measures of outcome. Reference Elphick1 The aspiration is to use the same data within units to improve the effectiveness of treatments and in commissioning to provide quality indicators. This should be especially welcome in mental health services, where the current set of management information is only loosely related to actual care processes.
But to provide the maximum benefit, for most people account must be taken of quantity as well as quality. With the advent of electronic care records it is technically feasible to collect and analyse both types of data, but will it be worth the time and effort? This review describes two key intertwined strands of development in the new approach: routine clinical outcomes measurement and the new system for funding, Payment by Results.
Routine clinical outcomes measurement in mental health services
In a perfect world, clinicians would routinely examine whether or not their patients improved after their interventions, tempering their findings with context data. Sparse interest in this until recently may have been the result of inadequate information systems, 2 predictions of insuperable obstacles for which little evidence has as yet emerged Reference Macdonald and Trauer3 and also perhaps the overweening influence of evidence-based medicine – why bother to check whether an intervention is actually working if trial data say it should?
Routine clinical outcomes measurement comprises measurement of changes in: health and social status; context data such as age, comorbidity and diagnosis; and interventions. Reference Macdonald4 Although the change in scores in health and social status over a given time period is usually called the ‘outcome’, all three dimensions are necessary for interpretation. These effectiveness data complement those produced by empirical studies, may reveal associations not apparent in smaller research studies, and, when fed back to clinicians and managers, can encourage reflective practice. When these data are complete enough, making changes to the organisation of services, and perhaps interventions themselves, may be justified on their basis. Reference Hill5
In England, the ‘family’ of Health of the Nation Outcome Scales (HoNOS) Reference Wing, Beevor, Curtis, Park, Hadden and Burns6 are the most popular outcome measures in adult secondary mental health services. Most have 12 scales, each scored 0–4, covering a range of symptoms, functioning and relationships. Health of the Nation Outcome Scales, and latterly HoNOS for older adults (HoNOS65+), have been nominally ‘mandatory’ in statistical returns for several years, but the Department of Health has never convincingly demanded their implementation. Reflecting the difficulty in choosing a universally appropriate instrument they finally published a compendium of optional measures in 2008. 7
Both HoNOS and HoNOS65+ were developed by psychiatrists and psychologists and reflect the interests of these professional groups. These may not accord with what patients would regard as important. The emergence of patient-reported outcome measures has not necessarily resolved this, at least in mental health, since despite being completed by the patient, most of these measures still reflect professional preoccupations.
Progress in routine clinical outcomes measurement development has now been galvanised by marriage with a more pressing, finance-driven policy: Payment by Results.
What is Payment by Results?
Payment by Results is the English version of a worldwide ‘case mix’ approach to health funding. 8,Reference Fetter, Shin, Freeman, Averill and Thompson9 Healthcare provision is remunerated by the payment of varying tariffs for each defined group of procedures or for episodes with specified diagnoses: the ‘costing currency’. Since payment depends on the recorded activity level, the more you do, the more you earn.
Before Payment by Results, per capita ‘block contracts’ left National Health Service providers to prioritise between patients with different types of problems. They tend to split their allocated budgets between teams along historical lines, leaving care staff with patient-by-patient rationing decisions. Commissioners have little information upon which to allocate resources or assess quality, so block contracts might well be dubbed ‘Payment by Intent’. In contrast, Payment by Results is an information-based commissioning system, intended to match finances to specific local population needs. It has been in operation in English acute services for several years; but they have relatively well-defined, coded procedures with predictable costs. In mental health services it is less clear what payment should be made for which results.
What payment?
A Payment by Results tariff is a price, not a cost, and prices reflect the relative values of the commodity to the purchaser and provider. The provider must try to set prices that are higher than costs but less than those of competitors. Yet in mental health trusts there has hitherto been no need for the detailed bottom-up accounting that can attribute costs to standard clinically defined groupings, so the first tariffs will be guesstimates – and that, in a competitive shrinking economy, is risky.
Which results – quantity or quality?
In acute services, activity levels alone are used as the ‘result’, but in our sector we hope to turn the difficulty we have had in defining a costing currency to our advantage, measuring both quantity and quality as a routine.
Because diagnostically defined groups of mental health patients do not have statistically homogeneous costs Reference Elphick and Anthony10 and there is no available intervention classification, the Department of Health chose to use ‘care clusters’ as the costing currency, previously developed to support allocation decisions in community teams. 8 Service users are grouped on the basis of a set of scales now called the Mental Health Clustering Tool (MHCT), comprising 12 HoNOS/HoNOS65+ scales and 6 additional items, assessing both the nature and severity of problems. There are currently 20 clusters, falling into three crude diagnostic categories (non-psychotic, psychotic and organic). Care pathways can be identified, preferably locally, for each. When and how a patients’ cluster might change are subject to a recently issued protocol. Routine collection of care cluster information should therefore provide both the number treated and change in HoNOS/HoNOS65+ scores for each patient in each cluster, and both parameters can eventually be related to finances.
Concerns include the validity and reliability of the MHCT (and the related problem of ‘gaming’ in which patients are inappropriately allocated to clusters that attract greater funding), which we have insufficient space to discuss here, and a lack of reassurance that costs per case within a cluster will be similar enough to support a viable tariff calculation. Also, although cluster allocation is not intended to constrain the clinician in their choices for individual management, more sophisticated electronic care record systems will allow variations to be noted.
How will Payment by Results impact on routine clinical outcomes measurement?
The decision to use HoNOS/HoNOS65+ as the basis of the chosen clustering tool has already had a practical impact on routine clinical outcomes measurement by causing a sharp rise in the numbers of care episodes for which at least one HoNOS/HoNOS65+ rating has been recorded. However, its impact on the genesis of pairs of ratings so necessary for routine clinical outcomes measurement is less obvious. Using the MHCT only for clustering at assessment is simpler than following a protocol for re-allocation at each care transition and discharge, but wastes the potential value – both for routine clinical outcomes measurement and as a costing currency for long-term care. Although there is no evidence for ‘gaming’ of routine clinical outcomes measurement ratings by staff so far, Reference Macdonald and Trauer3 were payment to depend on actual change in HoNOS/HoNOS65+ ratings, it is highly likely to occur. Routine use of patient-reported outcomes measures can act as a check against this.
How will routine clinical outcomes measurement impact on Payment by Results?
Routine clinical outcomes measurement is already central to the Payment by Results approach because it provides the multidomain severity ratings that are used for clustering. Patients in each cluster are therefore necessarily homogeneous in initial HoNOS/HoNOS65+ scores. Then, with the addition of patient-rated or patient-specified outcomes measures, multiple ratings over time should provide a measure of effectiveness, safeguarding services by enabling quality to be publicly visible on the same spreadsheet as financial accounts.
Conclusion
The obstacles to routine clinical outcomes measurement are now being overcome in trusts with significant clinician and management support. There is no doubt that Payment by Results of some sort will have sufficient management support; it will be a financial imperative. But it remains to be seen whether Payment by Results enhances routine clinical outcomes measurement, encouraging clinical involvement in its development, or, through unimaginative or partial implementation, further isolates clinical from management priorities with all that that implies. Whether all this is worthwhile can only be really judged by whether patient outcomes are better than they were before, and at what cost. So if we do it, we had better do it well.
eLetters
No eLetters have been published for this article.