Taking equity seriously: A dilemma for government from allocating resources to primary care groupsBMJ 1998; 316 doi: http://dx.doi.org/10.1136/bmj.316.7124.39 (Published 03 January 1998) Cite this as: BMJ 1998;316:39
- Gwyn Bevan, reader in policy analysisa
- a Department of Operational Research, London School of Economics and Political Science, London WC2A 2AE
- Accepted 16 September 1997
General practitioner fundholding seemed to threaten equity of access to health care in the NHS in two ways: fundholders' patients have been treated more quickly than patients from non-fundholding practices in the same authority,1 2 3 and there are suspicions that fundholders have been funded more generously than have health authorities, who supply funds to non-fundholders.3 4 5 6 The new Labour government has now announced its intention of abolishing fundholding and has introduced policies that are intended to remedy both types of inequity.2 7 8 Because of endemic variations in medical practice, however, this is not possible—leaving the government on the horns of a dilemma over which type of equity it intends to achieve. It will have to choose either common waiting times or an equitable allocation of resources, and these choices have quite different policy implications.
The equity dilemma
The NHS was created with the objective of reducing inequity,9 10 but this raises questions over which type of equity is being sought.11 The NHS reforms introduced two types of local insurers: general practitioner fundholders and health authorities.12 Fundholding seems to compromise two different types of equity: clinical equity, which is achieved when clinicians use only grounds of clinical priority to decide when patients are to be treated, and financial equity, which is achieved when local insurers within the NHS are allocated equal resources per capita, taking account of the relative risk of their populations.
GP fundholding is criticised for having introduced two types of inequity into the NHS
Fundholders' patients have been treated more quickly and may have been funded more generously than patients from non-fundholding GPs
New government policies are intended to introduce common waiting times and equitable funding for patients from all types of practices
Variations in medical practice preclude achieving both common waiting times and equitable funding across general practice populations
The government will have to choose which type of equity is to have priority for its new primary care groups
The government's policies for the new NHS are to retain the purchaser-provider split, with purchasing based on primary care groups organised around teams of local general practitioners and community nurses.8 Each primary care group will have available their population's share of available resources within a single cash limited envelope. Introducing primary care groups raises the same equity issues as fundholding. Patients of fundholding practices seem to be in the same privileged position as patients with private health insurance, who are able to use their extra resources to secure shorter waiting times. It may thus seem that a policy that is directed at one type of inequity will be directed at the other, and that it is immaterial which is given priority. This example of the introduction of primary care groups shows that it is not.
Example: introducing primary care groups
Consider a health authority that covers two towns (Lowerton and Higherton) served by one (Tonville) NHS trust. Each town has a population of 50 000 with similar needs (and virtually no private health insurance), so each would receive the same budget funded by a capitation formula. There is, however, a consistent difference in referral rates of general practitioners in each town: the general practitioners of Lowerton have half the referral rate of those in Higherton. The local medical school has investigated these differences and failed to find any convincing explanation other than differences in medical practice style. This is in common with most studies of variations in referral rates.13 14 15 Coulter et al showed that higher rates of referral by general practitioners tended to result in higher rates of admission to hospital.16 In numerous studies throughout the world of variations in per capita use (after standardising for age and sex) of hospital services for many conditions, the general conclusion is that variations in use are heavily influenced by variations in medical practice.17 18 19 20
Consider the position now and after the introduction of primary care groups. Now there is clinical equity between Higherton and Lowerton, with common waiting times across the authority. There is financial inequity (which is hidden), with spending per capita of Higherton being twice that of Lowerton. It is possible to achieve clinical equity when the health authority is divided into two primary care groups by funding each on past use of services. This policy was initially applied in setting fundholders' budgets and has since been changed, precisely because it would lead to financial inequity (with Higherton being funded at twice the rate of Lowerton). To achieve financial equity, each primary care group would be funded equally: by a capitation formula that assumes doctors deliver treatment at the same rates for populations in equal need and makes no funding concession for variations in medical practice. This must result in clinical inequity, as each primary care group has access to the same volume of supply, but demand from Higherton is twice that of Lowerton. What differences in waiting times might emerge under these circumstances? The impact is illustrated by applying the basic queuing model.
The basic queuing model is based on three simplifying assumptions: that the trust provides one service only (acts a single server); that patients are seen in the order in which they arrive; and that of rates of arrivals and of service are distributed randomly (with Poisson and negative exponential distributions). Figure 1 shows the relation in the steady state between the expected queue length and the ratio of arrival rate to service rate (known as the traffic intensity). The values shown apply when the arrival rate is lower than the service rate. If the arrival rate exceeds the service rate the queue, and waiting time, becomes longer and longer.
Figure 1 shows the common result of most queuing models with stochastic arrivals and service times: as the arrival rate approaches the service rate, the queue and waiting time become infinitely long.21 When the server is idle, arrivals will be expected later, which will correspond to this idle time, and the server will not then be able to cope. Pressures for efficiency require service rates to be close to arrival rates, which inevitably results in vulnerability to small increases in arrivals. Requiring hospitals to achieve high bed occupancy rates, for example, makes them vulnerable to fluctuations in admission rates resulting in red alerts.
In the example, queues will change drastically. This can be illustrated with an average annual referral rate of about 20%,22 which gives a mean referral rate across the two towns of 180 patients per week. With a current mean service capacity of 200 patients a week there is a common queue of nine patients. With an equitable budget for its primary care group, Lowerton would have a mean referral rate of 60 patients a week, to a service capacity of 100 patients a week—resulting in an average queue of 1.5 patients (and a waiting time a sixth of the current one). Higherton would have a mean referral rate of 120 patients per week, to a service capacity of 100 patients per week—which means that the queue (and waiting time) would just grow and grow.
Hence the government must choose between initial clinical equity or financial equity. Successive governments have sought to remedy financial inequity, once this has been revealed, and the NHS reforms gave this priority over clinical equity.
Identifying and remedying financial inequity
At the start of the NHS, financial inequity for general practitioners was made obvious by massive variations in list sizes by area. Medical practices committees were created, and empowered to secure a more equal distribution through negative direction—that is, stopping general practitioners from moving into over-doctored areas.23
Hospital services were not related to populations until health authorities were created in the 1974 reorganisation. This probably explains why health departments began serious consideration of financial inequity to hospital and community health services only in 1975. In England, the Resource Allocation Working Party recommended capitation formulas to reduce financial inequity.24 Similar approaches were adopted in each country of the United Kingdom.8 This policy has had three problems.
The first problem was, and is, intrinsic to seeking financial equity. Ideally, resource allocation formulas require knowledge of differences in morbidity in populations, how much of this morbidity can be treated effectively, and the relative costs of different types of effective care. As this ideal is unattainable, any resource allocation formula will be open to dispute.25 26 Nevertheless, allocating resources by using formulas is likely to offer greater financial equity than relying solely on past patterns of expenditure.
The second problem was that policies were incoherent. A capitation formula set financial targets for each health authority based on its resident population with an automatic adjustment for cross boundary flows (patients who lived in one authority but who were treated in another).24 Authorities were thus funded as providers on the basis of their catchment populations.27 This policy was only crudely directed at financial equity. An authority with overspending caused by its residents' high use of services was, for example, required to cut its hospitals' budgets.28 Because governments seemed also to pursue clinical equity, authorities lacked an effective means of influencing their populations' use of services: patients could be referred to any hospital, and clinicians decided on treatment without reference to the patients' health authority.
The third problem was that health departments' policies applied to the health authorities only and ignored financial inequity within authorities.29
The rhetoric of the NHS reforms was that these would deliver efficiency in a market based on competition. Fundholding was presented as enabling general practitioners, as informed consumers on behalf of their patients, to put market pressure on hospitals; in turn patients, by choosing their general practitioners, would bring similar pressures to bear on them.11 This required creating local insurers (health authorities or general practitioner fundholders) and giving priority to financial equity, with clinical decisions on admission to hospital having to be related to the ability of the patients' insurer to pay. This resolved the previous incoherence in policy.
General practice populations are those for which clinicians make decisions on use of healthcare resources; all other populations are administrative artefacts. Thus, in principle, general practices offer natural building blocks for funding by capitation.30 31 They also offer a way of addressing financial inequity within authorities. In practice, fundholders' budgets were based on past spending.3 6 This revealed past inequities and created future problems: it created potential for general practitioners who knew they would become fundholders to manipulate increases in spending over the period used to set their future budgets.3 As demands for emergency care increased, health authorities could finance this only by reducing the resources available for elective care for non-fundholders.
The new Labour government has introduced policies directed at securing financial and clinical equity between fundholders and non-fundholders. Fundholders' budgets are intended to be a fair share of the health authority's allocation, and increases in emergency activity will be financed by a common charge.7 32 From April 1998, admission for non-urgent patients must be on the basis of clinical priority alone, with common waiting times for patients from all practices.2 Given the variations in medical practice, it is not possible to achieve both types of equity. The question is, which will be given priority for the new primary care groups?
Priority for clinical equity: running the policy clock backwards
Giving priority to clinical equity has the following implications. Each general practitioner enjoys autonomy over referrals and can refer anywhere in the United Kingdom. Constraints apply, not on individual general practitioners, nor individual practices, nor primary care groups, but only on all GPs using the same provider. This leads into “the tragedy of the commons”33: as no one owns common land, it becomes overused because no individual has an incentive to act responsibly. In this case, providers are the common land. Rationing is through waiting times defined by clinicians, which override purchasers' ability to pay. This implies discarding a policy of financial equity and allowing variations in medical practice to flourish. This is in effect running the NHS policy clock backwards.
Giving financial equity priority: running the policy clock forwards
The economic solution to the tragedy of the commons is to create responsible owners.33 For general practitioner referrals one way of doing this is through budgets for primary care groups and giving priority to financial equity. This need not result in permanent clinical inequity, provided that doctors reduce their variations in medical practice. This might be done by linking epidemiological surveillance of variations in medical care34 with use of evidence based medicine.
Giving priority to clinical equity … leads into “the tragedy of the commons”
Analysis of financial inequity between Boston and New Haven suggested that this was caused by variations in medical practice. Although in Boston 60% more was spent per capita on hospital services than in New Haven, inhabitants were unaware of this, and doctors who had worked in each city were unable to say which had the higher spend.35 Wennberg et al analysed the high spend in Boston by diagnosis related groups and found that Boston's higher utilisation was largely in groups with high variation in admission rates across small areas and for which there is little consensus over the appropriateness of admission to hospital.36 Analysis of general practitioner fundholders in Wales also suggested that high per capita spend was associated with high use of procedures with high variation.37 In our example, analysis may show Higherton's high spend to be concentrated in small areas with high variation in admission, with questionable benefit, and that there are services in Lowerton where rates of efficacious treatment are low and where take up ought to be increased. In this way, a policy of achieving financial equity might go hand in hand with achieving greater efficiency through more effective use of scarce resources. Emphasising this policy may offer a better prospect of real efficiency gains through reducing variation in medical practice than through provider competition.38
Previous Labour governments introduced two main policy instruments for achieving financial equity in the NHS: medical practices committees for general practitioners and capitation formulas for hospital and community health services. The last Conservative government introduced the purchaser-provider split and general practitioner fundholding. These measures remedied the incoherence that undermined achieving financial equity in the unreformed NHS and put financial inequity within authorities on the policy agenda for the first time.
The government must choose initially between seeking clinical equity and financial equity
The government must choose initially between seeking clinical equity or financial equity for the new primary care groups. Because there are variations in medical practice it is not possible to have both. Starting with an emphasis on clinical equity in effect means abandoning the policy of financial equity. It leads into the tragedy of the commons, in which variations in medical practice are allowed to flourish unchallenged. Starting with an emphasis on financial equity addresses the tragedy of the commons by challenging variations in medical practice; reducing these may lead to clinical equity—but this can only be done by constructively involving epidemiologists, general practitioners, and hospital doctors in relating the allocation of resources to variations in use of hospital services.
I am grateful to Nick Mays, Jennifer Dixon, and an anonymous referee for comments on an earlier draft. I am a member of the Total Purchasing National Evaluation Team (TP-NET), which includes researchers from the King's Fund Policy Institute; National Primary Care R & D Centre; Universities of Edinburgh, Bristol, Southampton, York, and Birmingham; the London School of Hygiene and Tropical Medicine, and the London School of Economics and Political Science. More information about the evaluation is available from Nick Goodwin, King's Fund Policy Institute, 11-13 Cavendish Square, London W1M 0AN.
Funding: The national evaluation of total purchasing was commissioned and funded by the Department of Health. However, the views expressed in this paper do not necessarily represent the policy of the department.
Conflicts of interest: None.