- Nicholas Mays, director of health services researcha,
- Nick Goodwin, research officera,
- Gwyn Bevan, senior lecturer in health economicsb,
- Sally Wyke, senior research fellow the Total Purchasing National Evaluation Teamc
- a King's Fund Policy Institute, London W1M OAN
- b Department of Social Medicine, University of Bristol, Bristol BS8 2PR
- c Department of General Practice, University of Edinburgh, Edinburgh EH8 9DX
- Correspondence to: Mr Mays
- Accepted 19 June 1997
“Total purchasing” by general practitioners who have been delegated a budget by their local health authority to purchase potentially all the hospital and community health services for their patients is the latest and most ambitious extension of the concept of general practitioner fundholding. Several pilots have been established to test the concept, and a national evaluation has been assessing the costs and benefits attributable to total purchasing. We present here a summary description of the English and Scottish total purchasing pilot projects based on the preliminary report of the national evaluation.1 Evaluative data are not yet available on the consequences of the pilot projects.
General practitioner fundholding was introduced into the National Health Service (NHS) as part of the Working for Patients reforms of 1991.2 Under this scheme volunteer general practices are allocated budgets to purchase a restricted range of services for their patients, predominantly elective hospital procedures and community health services. From an early stage, however, several general practitioner fundholders were interested in extending the concept to a wider range of services. Four “pioneer” total purchasing pilot projects began in 1994 with regional support in Bromsgrove, Runcorn, Berkshire, and Worth Valley, West Yorkshire.
The idea behind these local projects was rapidly taken up by the NHS Executive, which in October 1994 invited bids to establish a first wave of total purchasing pilot projects in England and Scotland to start their preparations for total purchasing in April 1995.3 The pilots were to start live purchasing in April 1996 and run until April 1998. Unlike fundholding, and in line with their status as pilots, they were all to be evaluated “to identify the most appropriate models.” Fifty three first wave national projects began in England and Scotland in April 1995 and were joined by 35 second wave projects in April 1996. Pilot projects now also exist in Wales and Northern Ireland.
Total purchasing was defined very broadly as, “where general practitioners in a locality purchase all hospital and community health services for their patients”3 and was to be open only to volunteer fundholding practices. It was to be defined almost exclusively through the process of implementation, although it was clearly presented as the next logical step in the evolution of fundholding. There was also no legislation establishing the scheme, so the budgets for the projects had to remain the ultimate responsibility of the local health authority, whereas fundholders held budgets in their own right.
Total purchasing is the latest and most ambitious extension to the concept of general practitioner fundholding
National evaluation of 53 pilot projects is under way, but it is too soon to be able to assess the costs and benefits
Although the term implies that general practitioners purchase the full range of services for their patients, in practice they do not, and wide variation is seen between the pilot projects in their scope and way of working
Despite the new government's antipathy to fundholding, it is committed to continuing the policy of delegating responsibility for health services purchasing to groups of general practices in ways which bear strong similarities to the existing total purchasing pilots.
The potential of the total purchasing pilots
In principle, total purchasing has the potential to combine the strengths of two very different approaches to using budgetary control to secure cost effective health services—fundholding and health authority purchasing. It also has the potential to develop integrated purchasing and provision, building on the provider role of the general practices. However, at least as many, if not more, potential weaknesses are inherent in the scheme (see box). The evaluation of the English and Scottish pilot projects was designed to shed light, among other things, on these issues.
Combines best of top down, strategic purchasing for the needs of a population and bottom up approach based on responding to demands of individual patients
Scope for service innovation and alternatives to expensive hospital care leading to improvements in cost effectiveness
Projects may act as vanguard to secure service improvements or cost reductions which health authorities can then include in their own contract
Sensitivity to local needs because of general practitioners' knowledge of their patients
Clinicians negotiate on service improvement rather than relying on non-clinicians to purchase care
Fragmentation of NHS priority and purchasing decisions between many smaller purchasing organisations
Higher transaction costs (especially for providers having to deal with increased numbers of purchasers)
Difficulty in managing unpredictable demand and associated costs within budgets set for small populations
Difficulty in finding a fair means of setting budgets
Deepening existing inequity between general practices as participating practices develop but the benefits are not available to all practices
Excessive reliance on expertise and energy of a few lead general practitioners threatening sustainability of projects
Lack of expertise to purchase complex services with which general practitioners are less familiar (for example, mental health services)
No clear incentives for practices to take part or for health authorities to cooperate
Characteristics of the pilot projects
All eight English regions and five of the 15 Scottish health boards have at least one first wave pilot. They are predominantly in rural or suburban areas rather than in the main urban centres which is consistent with the geography of fundholding.4 The second wave of 35 projects does, however, include inner city projects, including two in inner London.
Table 1 summarises the basic characteristics of all the projects and table 2 presents information on the regional distribution and population share of the first wave pilots. Twenty seven of the 53 first wave projects comprised a single general practice and the remainder were multipractice collaborations. Two of the projects included a general dental practitioner to work with the general practitioners, particularly on the purchasing of specialist dental services. One of the dental projects is also distinctive in involving a community pharmacist, who advises the general practitioners on the pharmaceutical aspects of their purchasing, thus departing from a purely general practitioner led approach to purchasing.
In the first 18 months up to October 1996 two of the 53 projects had withdrawn from the scheme and 17 individual practices had withdrawn from specific projects. One of the withdrawals was a large multipractice pilot and the other was a small, single practice project. The main reasons for withdrawal were the large amount of managerial and developmental work required of the general practitioners to prepare for total purchasing, difficulty in making progress when faced with a lack of information to inform purchasing decisions, and differences of view between the project and the local health authority and among the general practitioners.
Organisation and management
All the total purchasing projects are subcommittees of their parent health authorities. As with so many aspects of the projects, however, the ways in which they are organised and managed vary widely. Projects are generally steered by a project board, which usually includes senior staff from the health authority, the lead general practitioners, and the project manager. Day to day decision making is normally in the hands of an executive group, which is most likely to include the project manager and the fundholding managers and general practitioners from each of the participating practices. All the projects are collaborations with the health authority since all rely, to some degree, on health authority staff—for example, in finance, contracting, or public health medicine. However, the nature and extent of interaction varies.
The range of numbers of practices involved in first wave projects and the absence of a management blueprint contributed to considerable variation in direct management costs. In the preparatory period (1995-6) this varied from no additional spending to £40 000 in single practice projects and from £20 000 to £300 000 in multipractice projects. Per capita these costs ranged from £0.26 to £8.05 with a mean of £3.00. The absolute levels and range of spending will probably fall in future because many expenses are incurred in setting up projects (in particular, establishing information systems). Nevertheless, concern exists about the likely increase in NHS transaction costs from of running fundholding, total purchasing, and health authority purchasing in parallel.
Total purchasing suggests the intention of allowing general practitioners to purchase all the hospital and community health services required by their patients. However, none of the projects planned to purchase the full range of services in 1996-7. Instead, the projects have undertaken selective purchasing of services where they had a special interest in change, based on their own local experience, and where change was feasible in the short term. For example, six projects were simply gathering information to guide purchasing in future years. Only 12 projects had more than four specific purchasing objectives for 1996-7 which differed substantially from the objectives of the health authority.
Table 3 shows the main purchasing priorities of the pilots, covering both the services affected and the mechanisms used to bring about service changes which are not necessarily specific to individual service areas. We derived this information from a combination of a content analysis of the projects' purchasing plans and analysis of semistructured interviews with key participants at each project.
Reducing the number of emergency admissions or attendances at accident and emergency departments was the most frequently mentioned priority for service change. Several approaches to this major strategic problem for the NHS were put forward, including developing minor injuries clinics, extending out of hours cover by general practitioners, using general practitioners to assess potential emergency admissions, improving admission procedures at local trusts, patient education, and increasing the availability of intermediate facilities such as community hospitals.
If devolved purchasing is to become a permanent feature of the NHS and the NHS still aims to secure equity of access to services, then some form of capitation must be used to set budgets. If not, all the problems associated with giving fundholders budgets based on previous (inequitable) levels of spending will arise.5 Twenty one of the 30 authorities with total purchasing pilots which provided data reported their intention of using capitation. However, this, in turn, raises the question of the appropriate size of population required for purchasing financed by capitation, given that smaller populations will tend to produce larger random variation in demand for services from year to year.6 Recent studies in the United Kingdom suggest that a risk pool for total purchasing of 10 000 patients over one year would be unmanageable, but that the annual risk of overspending falls sharply when populations reach about 30 000.7 8
Since not all the total purchasing pilots have populations over 30 000, it is not surprising that 14 of the 21 authorities which used capitation found that the sum it produced was so different from past expenditure that they had to set actual budgets between these figures. Agreeing the budget setting method and negotiating the actual amount were the most contentious aspects of implementing total purchasing and appeared to put more strain on the relationships between the practices and their local health authorities than anything else.
In practice the first wave total purchasing pilots are developing as collaborations between health authority and fundholder purchasers, if only because the resources they deploy are still the legal responsibility of the authorities. In the absence of a national specification for total purchasing, however, considerable variation exists in how the pilot projects are organised and in the nature of the interaction between the general practices and their health authorities.1 Some of the projects are implementing total purchasing as an extension of fundholding, emphasising the quest for microefficiency gains for their patients, while others are using it to develop more strategic collective action between practices. A sign that total purchasing may be developing differently from fundholding is the fact that in fewer than half the projects had the practices agreed with their local health authority the right to use any underspends for the benefit of their own patients. This suggests that for most projects this experiment is about developing purchasing to achieve strategic change in close alliance with the health authority.
Despite the new government's commitment to abolishing standard single practice fundholding, it seems increasing likely that the relative merits of different approaches to general practitioner led purchasing will be evaluated9 and that fundholding will only gradually be replaced by Labour's system of “local GP led commissioning groups.” These were envisaged as groups comprising all practices in localities with populations of 50 000 to 150 000 and given their own budgets; this was intended to maintain the involvement of general practitioners but reduce the costs and inequity of single practice fundholding (C Smith, Health Service Journal conference, London, December 1996). Labour's proposals envisage that these groups could build on existing schemes, including total purchasing projects. As a result, the current evaluation remains important for policy development.
It seems increasingly likely that the NHS will be organised around general practitioner purchasing in some form in the future. Yet this raises a number of unanswered questions,10 such as: what balance should there be between health authority and general practitioner leadership in purchasing; should purchasing practices have their own budgets, or are indicative budgets sufficient11; how are general practitioners who are uninterested in purchasing to be included to ensure equity; should different types of services (from community nursing to rare, costly services) be purchased by purchasers responsible for different sizes of population? By watching how the total purchasing pilots progress, we should be able to answer such questions, at least in part.
The evaluation is a collective effort by a consortium of health services researchers led from the King's Fund Policy Institute and comprising staff from the National Primary Care Research and Development Centre, the Universities of Bristol and Edinburgh, the Institute for Health Policy Studies at the University of Southampton, the Health Services Management Centre at the University of Birmingham, and the Health Services Research Unit at the London School of Hygiene and Tropical Medicine. The views expressed in this paper are not necessarily shared by the two funding departments.
Funding: Department of Health and Scottish Office.
Conflict of interest: None.