Effect of fundholding on removing patients from general practitioners' lists: retrospective study

BMJ 1998; 317 doi: (Published 19 September 1998) Cite this as: BMJ 1998;317:785
  1. Dermot O'Reilly (d.oreilly{at}, deputy directora,
  2. Keith Steele, senior lecturerb,
  3. Barry Merriman, research assistanta,
  4. Andrew Gilliland, senior lecturerb,
  5. Scott Brown, general practitionerc
  1. aHealth and Social Care Research Unit, Queen's University of Belfast, BT12 6BJ
  2. bDepartment of General Practice, Queen's University of Belfast
  3. cRoyal College of General Physicians Research Practice, Mountsandel Surgery, Coleraine BT52 1JB
  1. Correspondence to: Dr O'Reilly

    Fundholding by general practitioners was introduced during NHS reforms in 1989. Little is known about its impact on the quality of patient care.1 One measure of this impact is the rate at which practices decide they do not wish to continue to provide general medical services for patients and remove them from their lists (“removal at general practitioner's request”). In Northern Ireland the rate of the removing patients from practitioners' lists increased after the introduction of fundholding in 19932 which suggests that there is a relation between fundholding and removing patients from lists. We report an investigation to determine if becoming a fundholding practice changed the rates of removing patients from practice lists.

    Methods and results

    For the past 15 years the Central Services Agency has maintained a register of all patients removed from lists at a general practitioner's request. The database does not contain patients who have been removed for reasons such as leaving the country or moving outside the practice area. However, records are retained if the patient has died or emigrated. Demographic data were obtained for each patient removed from a list during the study, and each record was coded according to the fundholding status of the practitioner.

    There were four waves of fundholding between 1987 and 1996; the first began in April 1993 and each new wave followed in April of the succeeding year. The data were divided into three phases for each fundholding practice: fundholding, preparatory year (the financial year prior to fundholding), and prepreparatory period (from January 1987 until the start of the preparatory year). The removal of an individual patient or family unit was counted as one decision, and only first time removals within the 10 year period were analysed. Rates for first removal decisions per 10 000 person years were calculated for each period as previously described.2 Rates for non-fundholding practices were also examined using the commencement date of each wave of fundholding to artificially divide the data into before and after periods. In April 1996 there were 419 general practitioners in 114 practices serving 724 104 patients.

    Results of the analysis are shown in the table.There was no increase in the rate of removing patients from non-fundholding practices. Among fundholding practices the rate of removing patients increased from 1.8/10 000 person years in the prepreparatory period to 2.2/10 000 person years during the fundholding period; this was an increase of 21.4% (95% confidence interval 7.4% to 35.5%) Practices that became fundholding practices in later years removed patients more frequently and started removing patients at higher rates during the preparatory year.


    The rates of removing patients from general practitioners' lists are influenced by characteristics of both the practice and population.2 In this analysis practices were compared with their earlier performance obviating these potentially confounding variables. The increases in the rates among fundholders are therefore intrinsically related to fundholding status. The different rates of removal occurring between successive waves of fundholders and between fundholders and non-fundholders could be attributed to differences in socioeconomic and demographic characteristics of practice populations arising from a selection bias in practices that became fundholding practices. Other factors, such as the 1990 contract (which substantively altered the terms of service of all general practitioners in the United Kingdom), cannot explain the increase, as similar changes were not found for the non-fundholding practices.

    Rates of removing patients from general practitioners' lists according to fundholding status and time of becoming a fundholding practice

    View this table:

    It is unclear why the rate of removing patients has increased but “list cleaning” (removing patients who have died or left the practice area from lists) can be discounted. The database used in this study is maintained separately from the patient registration data within the Central Services Agency and contains only removals made at the request of general practitioners. Trained staff undertake validation checks which include contacting practitioners and writing to patients to inform them of the category of removal. Removals because of death are processed differently, and follow up procedures would identify a misclassification.

    The public perception is that financial factors motivate fundholders to remove patients from their lists.3 In the United States where healthcare systems provide financial incentives “adverse selection” is common; it has been suggested that the reforms in the NHS could stimulate similar effects.4 Increased rates of removing patient may, however, reflect the additional workload and pressures of fundholding5 rather than attempts at financial gain.

    The decision to remove patients occurs comparatively infrequently and our results suggest that a fundholding practice with a list size of 5000 patients would be making one additional removal decision every five years. Our findings suggest that other areas of health care that experience large increases in workload, or where the potential for adverse selection exists, should be monitored.


    Contributors: KS and DO'R had the original idea for the study. DO'R, KS, AG, and SB designed the form for data collection. DO'R and BM were responsible for the analysis of the data. DO'R, KS, AG, SB, and BM wrote the paper. DO'R is the guarantor for the paper.

    Funding: This study was assisted by a grant from the Royal College of General Practitioners.

    Conflict of interest: None.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    View Abstract