Intended for healthcare professionals


Better ways of assessing health needs in primary care

BMJ 1995; 310 doi: (Published 25 February 1995) Cite this as: BMJ 1995;310:480
  1. John Shanks,
  2. Sadru Kheraj,
  3. Sue Fish
  1. Consultant in public health medicine Lambeth Southwark and Lewisham Health Commission, London SE1 7RJ
  2. General practitioner Iveagh House Surgery, London SW9 7SE
  3. Practice manager Crown Dale Medical Centre, London SE19 3NY

    Requires adapting conventional methods

    Commissioning authorities increasingly attempt to base their purchasing decisions on systematic, epidemiologically informed assessment of the health needs of their local populations. General practitioners as purchasers usually rely on their own judgment. No one knows which method works better, but a combined approach may capture some of the advantages of both.

    One of the more widely welcomed aspects of the NHS reforms was the requirement that health authorities' decisions on how to use NHS resources should in future be based on a systematic assessment of each local population's needs for health care. This is meant to take account of local demography, the epidemiology of health problems, evidence on the effectiveness of treatments, and the preferences of local people.1 Needs assessment has become an important task for public health doctors and others working in commissioning authorities. This more rational and scientific method is put forward as an improvement over the former approach to allocating health funding, caricatured as “same as last year, plus or minus five per cent for pressure groups.”2

    The most obvious way in which general practitioners can shape decisions on the pattern of purchasing for hospital and community health services is by becoming fundholders. Non-fundholding general practitioners may influence commissioning decisions made by health authorities in various ways,3 but not much has been done to develop a method of systematically assessing health care needs in primary care. General practitioners may therefore find themselves limited to their unaided judgment of which services would most benefit the health of local people.

    An approach that explicitly combines epidemiological analysis with the personal knowledge of primary care practitioners has several advantages. A strength of the epidemiological approach is its ability to look beyond patients who already demand health care to those don't demand it but need it. Homeless people and those with chronic severe mental illness may be among these invisible potential patients. The main advantage of the primary care perspective is the personal knowledge that primary health care teams derive from extended day to day contact with their patients. This informed opinion can be enhanced by the systematic analysis of data that already exist in the practice or are specially collected—for example, by rapid appraisal survey.4 General practitioners' direct experience as referral agents also enables them to assess the quality of service delivered by local providers.5

    What has been described as “living epidemiology” might start with the basic demography of the practice population, extracted from the age-sex register.6 Information on presenting health problems available at the practice level often allows identification of people with chronic diseases (most commonly diabetes, asthma, and hypertension) and some information on those lifestyle and risk factors assessed as part of health promotion activity (for example, cigarette and alcohol consumption). Prescribing data in the practice can be a useful proxy measure for those conditions that are treated by specific drugs (for example, methadone mixture for opioid dependency; insulin for diabetes), assuming that prescribing activity mirrors prevalence. Data from several practices can be aggregated to provide information on a locality.

    With the same aim, data that are held centrally, such as national census information or district mortality, can be disaggregated to small area levels by use of postcodes. Data derived from practices usefully complement centrally held data. Practice data can be very up to date and rich in information about morbidity and qualitative aspects of patients' experience. That they may be of variable quality and available only for registered patients are limitations. Centrally held data are uniformly available for all practices, are more consistent in quality, include people not registered with a general practitioner, are strong on mortality but may be out of date, and miss the qualitative dimension altogether. The combination of practice based and centrally held information for the same locality therefore provides a more complete picture of needs than either could separately.

    Such information can inform locally sensitive commissioning of hospital and community health services. It can also help to guide the allocation of additional resources to practices and can help primary health care teams set priorities for themselves in their role as providers.7 An estimate of relative need in one practice compared with other practices or localities in the commissioning area is generally more realistic and useful than a calculation of absolute need. A practice with a socially deprived population containing a high proportion of young adults is likely to have a greater need for contraceptive and drug misuse services than a similar sized practice whose population is mainly affluent and middle aged. Although this sort of information has been available for some time, the current distribution of services often fails to reflect local variation. An individual contract between each practice and the community health services trust would be one example of matching services to need more closely.

    The volume and quality of data vary,8 9 and complex soft information may be difficult to handle and analyse. Among general practices in Lambeth some enter on computer a Read disease code at each consultation while others record manually only basic minimum information such as data obtained when new patients are registered (T Crayford, personal communication). The extent to which practice based data can be taken as representative of the local population is greater in areas where almost everyone is registered with a general practitioner. In some parts of inner London, however, the proportion may be as low as 70%.10 Registration rates are also lower for some marginalised groups, such as single homeless people, who have special health needs.

    Perhaps the greatest obstacle to the sort of synthesis of information described here has been the historical separation of the family health services authorities from district health authorities. Like the holders of two halves of a prize draw ticket who live next door but don't often speak, these two groups have found that circumstances have prevented either from claiming their winnings. Planned legislative changes should soon permit them to discover the benefits of an arranged marriage.11


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    View Abstract