Intended for healthcare professionals

General Practice

General PracticeMeeting the information and budgetary requirements of primary care groupsCommentary: Accurate information may be difficult to produce

BMJ 1999; 318 doi: (Published 16 January 1999) Cite this as: BMJ 1999;318:168

Meeting the information and budgetary requirements of primary care groups

  1. Judy Gilley, deputy chair.
  1. General Practitioners Committee, BMA House, London WC1H 9JR
  2. Office for National Statistics, London SW1V 2QQ

    Primary care is experiencing another wholesale reorganisation as the government's “new NHS” is implemented. The intention is to bring general practitioners and other healthcare professionals together at a local level to assess the needs of their shared populations and to ensure that resources are allocated to meet those needs. The mechanisms chosen for England (primary care groups) and Wales (local health groups) have more in common than those for Scotland (primary care trusts and local health care cooperatives). The differences among the three countries represent a worrying fragmentation of “national” health service structures. Primary care groups (PCGs) are not voluntary; all general practitioners are members of a primary care group. Shadow groups started operating at the end of October 1998and go live in April 1999.

    Their three main areas of responsibility include the development of primary care, the commissioning of secondary care services, and a quality agenda delivered under the umbrella of clinical governance. Only level 1and 2groups will exist from 1999; primary care trusts, described by one civil servant as “PCGs in long trousers,” with their wider remit to include community health services, will not start until 2000.Current NHS community trusts, together with primary care groups, will be able to bid to progress to primary care trusts. Level 1groups will have a largely “advisory” role in the commissioning of secondary care services. Level 2groups, in contrast, will take charge of at least 40% of their unified budget to purchase secondary care services. The government's stated aims for all levels are similar: tackling variations in quality of care and distributing NHS cash more fairly. “The healthcare needs of populations, including the impact of deprivation, will be the driving force in determining where the cash goes.”1 Few will argue with these aims, but clearly any redistribution exercise will mean winners and losers, and the pace of change towards fair target shares will be all important if local services are not to be destabilised. After negotiations between health minister Alan Milburn and the General Practitioners Committee (GPC) in June, general practitioners locally were able to choose whether they wanted to form the majority on their primary care group board and have it chaired by a general practitioner; most voted for this high degree of participation.

    Given that primary care groups will work with a cash limited unified budget, derived from the existing separate budgets of the hospital and community health services (£23bn (thousand millions) nationally), prescribing (£4bn), and GMS Cash Limited (GMSCL) (£1bn), it is essential for general practitioners to understand the principles of resource allocation. The idea of a unified budget originally generated immense anxiety. General practitioners were at risk of personal financial loss if there were reductions in the GMSCL budget, which reimburses a large part of general practitioners' committed expenditure on staff, premises, and computers. The GPC successfully negotiated protections for the GMSCL, guaranteed by the health minister.

    Summary points

    • Primary care groups have been set up in England to assess the needs of populations and ensure that resources are allocated to meet those needs; they become live in April

    • Groups are responsible for developing primary care, commissioning secondary care services, and tackling variations in the quality of care

    • Budget setting will at first be based on past year spending and existing formulas, but it will later be determined by formulas based on need; how groups will gather information remains unclear

    • General practitioners on the boards of primary care groups need to be adequately supported for this venture to be a success

    Budget setting for primary care groups

    There are two stages of budget setting. An initial unified baseline for 1999-2000will be based on historic spends, existing formulas for resource allocation, and the health authority population. Health authorities have identified the existing (1997-8) level of services to primary care grouppopulations and calculated the appropriate share of their baseline for each primary care group. This involves policies that lack precision, and it is crucial that these assumptions are tested for fairness with primary care group boards.

    The second stage will be the move to target allocations determined by formulas based on need set by ACRA (Advisory Committee on Resource Allocation) and focused on general practitioners' registered populations. The “distance from target” for each primary care group—that is, the gap between what the group has and what it should have—will vary. There will be a policy on pace of change in that health authorities will determine how fast the distance from target is reduced for each group. This major redistributive process will generate winners and losers.

    The size of the population is the single most important factor in determining fair shares of resources; other needs based formulas have been described as “the icing on the cake.” Which population is used for the purpose of primary care group allocations is therefore important. Traditionally government spending is based on figures from the Office for National Statistics of resident populations derived from census counts. The whole logic of the government's structural reform, however, points to the use of populations registered within a primary care group as the focus for resource allocation.

    Population statistics from the Office for National Statistics have their drawbacks: critics point to the 1991census, which was underenumerated because of the poll tax. Despite constant modifications there is a general consensus that it would be more accurate to move to general practice registered populations. These have also had their problems (such as double counting as patients move between practices), but as computerisation and patient registration on general practice links become the norm these problems can be resolved. The health minister has now established a working group to clear the way towards using general practice registered populations.

    Practicalities of data collection

    How primary care groups will gather information remains something of a mystery; some of the jigsaw pieces are falling into place with successive batches of guidance. Using the national general practice dataset, we can already link socioeconomic data at enumeration district level with general practices, which gives accurate information on deprivation. But how will primary care groups monitor performance of hospital trusts? Instead of individual invoices we are expected to create long term service agreements, which are to be “disease driven” and may place different components of a service with a range of providers. The “league table” of trusts' costs for surgical procedures has recently been published to inform purchasers' decisions.

    The government has announced £1bn to “harness the full potential of the IT revolution” and aims to have every general practice and hospital connected to the NHSnet. Next year over £40m will enable all computerised general practices to be connected and help fund the development of information services for primary care groups.2 These are logical but ambitious plans for the longer term, but it is difficult to see how primary care groups will monitor and collect data in the short term. The GPC believes that the costs have been seriously underestimated.

    Primary care groups will work within a strict accountability and financial framework (adapted from existing health authority arrangements), including responsibity for keeping within budget. Pundits are already suggesting that “failure to keep a tight grip on prescribing will force cuts in other services.” Primary care groups will certainly need to consider a unified drug formulary, shifting more services into primary care if they can be more cost effective in that setting, and discussions with individual practices about significant variations in referral rates. A contentious and daunting list to add to the agenda of developing a new organisation with many new working relationships.

    Finally, the success of primary care groups will be determined to a considerable extent by whether general practitioners who work on the boards of primary care group are adequately supported. They must be properly paid for their work, have their locum costs reimbursed, and be confident that their eventual pension will not be reduced because the pay is not superannuable. General practitioners in 15% of all practices will be acquiring the skills of multidisciplinary working and corporate governance this winter.


    Commentary: Accurate information may be difficult to produce

    1. Azeem Majeed (azeem.majeed{at}, medical epidemiologist
    1. General Practitioners Committee, BMA House, London WC1H 9JR
    2. Office for National Statistics, London SW1V 2QQ

      In her article Gilley gives a good summary of the current position on primary care groups and on what the changes in the NHS might mean for general practitioners. The most important change is that by April 1999all general practices in England will be members of primary care groups.3 As primary care groups develop and take on increased responsibilities they will start to control the bulk of NHS funds and be responsible for monitoring the health status of their population. These changes will have important implications for the methods currently used by the NHS to set budgets, to monitor the performance of purchasers of health care, and to monitor the health status of the population. As Gilley states, one key objective will be the requirement for NHS information systems to generate information relevant to primary care groups as well as for their traditional focus, health authorities; this objective will be difficult to achieve.

      Difficulty in defining a population

      Although Gilley believes that general practice lists can provide better population estimates than official estimates from the Office for National Statistics, there is no published evidence to support this claim. General practice catchment areas do not follow geographical boundaries, overlap with the catchment areas of other practices, and, especially in inner city areas, may also cross health authority boundaries. General practice lists are also often very fluid, can fluctuate rapidly in size, and suffer from list inflation (that is, there are usually more people registered with general practices than live in an area). These problems will make it difficult to use such lists to calculate an accurate denominator population, which in turn makes it difficult to calculate stable and accurate rates for general practices. Aggregating lists by combining the lists of practices in the same primary care group will reduce the magnitude of this problem but will not overcome it entirely.

      Lack of practice based data

      Many of the indicators the NHS Executive proposes to use to monitor the performance of primary care groups are based on data that are not generally available at general practice level4—for example, information on deaths and cancer registrations. Hence, calculation of death rates and cancer registration rates for the populations registered with primary care groups will be difficult with currently available data. Some health authorities have tried to derive death rates for their general practice populations by using locally available information. This can, however, create discrepancies with rates derived from the national data held by the Office for National Statistics, which has very specific rules for coding the underlying cause of death from the causes given on a death certificate.

      Difficulty in attributing data to practices

      Even when data are available at general practice level they can be difficult to attribute to general practices because of inaccurate coding or because the patient has changed general practice since the general practice code on the record was completed. This applies particularly to hospital admissions data and hospital referral data.5 Further problems arise when general practices located in one health authority have a substantial proportion of their patients living in another authority. Because of limitations in the way that information on hospital admissions is sent to health authorities complete admissions data are often unavailable for practices that are located on health authority boundaries. Many such practices therefore often have artificially low referral and admission rates.

      Problems in setting primary care group budgets

      Budgets will initially be based on the current use of services. As Gilley states, however, use of current practice based data to determine the existing level of services used by the population of each primary care group will be imprecise. Hence, setting the budgets will be a difficult task and may lead to arguments between groups in the same authority, particularly when there are wide variations in the use of resources. Ultimately, groups will move from this method of funding, based on the current use of services, to one based on a needs based formula. But this “needs based” method of funding groups also suffers from a number of problems.

      The white paper stated that, “There will be a national formula to set fair shares for the new primary care groups as there is now for health authorities.”1 Several technical problems have to be overcome, however, before this objective can be achieved. The current formula for setting health authority budgets for hospital and community health services includes weightings for total population, age, mortality, and socioeconomic status. If a similar formula is to be used for primary care groups Gilley is correct in saying that the most important issue to be dealt with is what population base should be used in the formula. If general practice list sizes are used without any adjustment for list inflation this will lead to resources being moved from areas with low list inflation to areas with high list inflation. In the longer term the department of health would like to move towards using general practice lists as the population base for resource allocation despite the current limitations of these lists.

      The next issue that needs to be covered is how mortality and socioeconomic data can be includedin the formula for setting the budgets of these groups. As death rates are not routinely calculated for general practices one method of deriving these rates would simply be to attribute the mortality for the locality covered by a primary care group to that group. Because the populations of primary care groups will inevitably overlap, however, this will introduce errors into mortality attributed to the groups.

      Estimated socioeconomic variables for use in the resource allocation formula for primary care groups can be produced for general practices by linking patients' postcodes with census data for enumeration districts.4 As with attributed mortality data, however, these census derived variables will also contain errors. The size of the errors in attributed mortality and census data and their likely effect on funding are all currently unknown.


      Although it will be important for NHS information systems to produce accurate and useful information for primary care groups, technical problems will make it difficult to achieve this objective given the current arrangements for collecting and analysing health related information. Similarly, technical problems may also make it difficult to develop and implement a valid formula for determining the budgets of primary care groups. Gilley also discusses some other problems that could hinder the work of primary care groups. These include the level of payments for general practitioners who sit on primary care group boards and the costs of meeting the proposed future developments in information technology. A wide debate about these problems, what effect they might have on primary care, and how they can be overcome is needed to ensure that primary care groups can be funded fairly and carry out their functions effectively after they come into existence in April 1999.