General Practice

Changing to generic formulary: how one fundholding practice reduced prescribing costs

BMJ 1995; 310 doi: (Published 25 February 1995) Cite this as: BMJ 1995;310:505
  1. Jon S Dowell, clinical research fellowa,
  2. David Snadden, senior lecturera,
  3. James A Dunbar, general practice principalb
  1. a Tayside Centre for General Practice, Westgate Health Centre, Dundee DD2 4AD
  2. b Downfield Surgery, Dundee DD3 8NE
  • Accepted 11 January 1995


Objectives: To observe one general practice's attempt to reduce prescribing costs on becoming third wave fundholders through the introduction of a generic formulary applied to all new and repeat prescribing. To assess the impact on patients and prescribing patterns.

Design: An observational study using interviews with patients and practitioners; questionnaires for patients and prescribing data.

Setting: One urban general practice with five partners in Scotland. It became fundholding in April 1993.

Subjects: 71 searches of the register of repeat prescriptions identified 1274 potential changes in drugs. Questionnaires were sent to a stratified random sample of 280 patients four months after the changes were made; 33 interviews were conducted with 17 patients selected by local pharmacists to represent a wide range of opinion.

Main outcome measures: Changes in prescribing and response and satisfaction of patients.

Results: Of intended changes, 129 (70%) were in place after four months. Thirty three (20%) of the 167 patients who returned questionnaires were “very unhappy,” though interviews suggested that this was primarily with the communication they received rather than the change itself. Generic prescribing rose from one in three (37%) to over a half (58%). The volume of treatment dispensed (as treatment days) fell by 67674 (10.7%), and the average cost per day's treatment fell by 3.1p from 32.3p per day in 1992 (9.4%), producing a total absolute saving of 24% (pounds sterling137712) over the first year.

Conclusion: This practice has achieved a large reduction in prescribing costs rapidly. These were tolerated by patients, none of whom is thought to have left the practice for this reason. Great care must be taken to inform patients appropriately.

Key messages

  • Key messages

  • Financial incentives are increasingly being used to encourage cheaper prescribing

  • It seems that patients will accept large scale changes, at least within the context of fundholding

  • The introduction of such changes should be carefully planned, particularly the way in which patients are informed

  • There is a need for more research into the balance between prescribing expenditure and the provision of other services


Prescriptions issued by general practitioners accounted for pounds sterling3.6 billion of NHS expenditure in 1992-3; this continues to rise at 14% a year.1 It has been shown that fundholding practices are managing to limit this rising cost more successfully than other practices, although interpretation of these studies is complicated by control groups who were about to become fundholding.2 3 4 The Audit Commission advocates financial incentives as a promising way to encourage “rational prescribing” yet there is little information on how practices respond to such incentives or what impact this may have on patients. We describe one practice's experience of reducing prescribing costs while it became a third wave fundholder. The data reported here describe the methods it used, the savings made, reactions of patients, and the resultant changes in prescribing patterns.

We aimed to observe one practice's attempt to reduce prescribing expenditure; to assess the outcome of their actions; and to understand why they succeeded or failed.

Subjects and methods

One urban third wave fundholding practice agreed to cooperate fully in the study. In 1993 the five full time partners shared a list size of 6450. The indicative drug budget of pounds sterling453190 for 1992-3 was overspent, the actual budget for April 1993 to March 1994 was therefore increased to pounds sterling573802; equivalent to pounds sterling89 per capita.

The study was conducted by JSD, who had no previous connections with the practice. All aspects of the changes considered relevant were included, resulting in the four parallel studies outlined below. All appropriate practice information was made freely available.

The steps taken and the reasoning behind them were clarified through interviews with all practice doctors, relevant practice staff, and local pharmacists. Nonparticipant observation was conducted for two sessions at the practice's prescriptions desk and at one practice meeting devoted to refining the new formulary.

A personalised, anonymous questionnaire was sent to a stratified random sample of 280 patients four months after their repeat medication had been altered. This ascertained who had tried the new treatment, who was still taking it, how they had been informed, and their opinion of the change.

The practice prescribing data were converted to defined daily doses for more detailed analysis than that currently available through Scottish prescribing analysis.3 This allowed the volume of treatment dispensed to be accurately compared as well as the cost, irrespective of prescribing interval or changes to alternative preparations within the same therapeutic group.

Semistructured interviews were conducted with 17 patients within one week of their repeat prescription having been altered by the practice. Sixteen of these were interviewed again six months later. These patients were purposefully selected by two local pharmacists to sample a wide range of opinion.5 Patients who showed particular anger, confusion, or acceptance of the change were asked if they could be contacted by the researcher. Interviews were confidential and subjects were strongly reassured that their doctor would not find out who had been interviewed.


The most important background information and identifiable stages in the process were that the practice considered its management of changes to be efficient; all partners agreed to fundhold; success was seen to depend on improving services; improved services required funds to be released; historically high prescribing costs gave an opportunity to save; and the partners accepted that some patients would be unhappy about changes in their prescriptions.

A rapidly introduced change to prescribing habits in the practice was seen as essential to reap a one off saving from the prescribing budget. The following steps were taken to achieve this: a generic formulary was developed; desktop prescribing was started to ease introduction of the formulary; patients receiving repeat prescriptions not included in the formulary were identified (1187 patients were selected for transfer to the formulary alternative and 87 treatments were stopped outright); communication strategies were developed to inform patients; and changes in prescriptions were introduced within two months of fundholding.

This process took no more than six months and produced an acceptable workload for the general practitioners concerned. The development of the formulary proved surprisingly easy, essentially because it was viewed as necessary by every partner, was negotiated between them, and could be modified if necessary. Introduction of the formulary to all acute prescribing was eased by the use of desktop computers. Its application to repeat prescribing produced a considerable administrative burden. Searches were carried out for preparations not listed in the formulary and each reviewed by one partner. The records of patients were then changed to the formulary alternative. At the reviewing partner's discretion some patients were sent an explanatory letter; most were first informed when collecting their repeat prescription from the practice. Those unhappy about the change were asked to consult their doctor. This process created a lot of work for the receptionists and computer operator but did result in some tightening up of the register and some additional recalls for clinical review. No patients are known to have left this general practice because of these changes. Total list size is increasing. Overall the partners considered the effort invested to be well worth while.


Summary of patients' feelings in response to changes in prescribing. Values are numbers (percentages) of patients

View this table:


Of the 280 questionnaires sent, 167 (60%) were returned adequately completed. An additional 18 were marked as having “no change made.” Analysis with χ2 showed no significant differences between those patients whose treatment was stopped, generically substituted, or changed with respect to success rate, how the patient was informed, or the patient's opinion of the change so they are reported together.

The overall success rate was 70% (129/185), success being classed as a patient remaining on the new preparation or having ceased a “stopped treatment” without the substitution of an alternative. Of the 56 remaining patients, 37 (20%) were known to be taking their original treatment, seven were on other prescribed treatment, seven had stopped all treatment, and five bought their own alternative.

Respondents were asked if they were informed about the change and their opinion of it. In total, 122 of 167 (73%) patients could recall being informed of the change in at least one way, 39 (23%) stated they had not been told at all, and six (4%) failed to respond. Table I gives details of patients' satisfaction.

Satisfaction with communication received was closely correlated with satisfaction about the change (Spearman rs=0.81). There was no significant difference, however, in success rate between those who had or had not been informed. Ten patients reported needing to make an extra consultation with their doctor because of the change; eight patients required two consultations. Extrapolated to the practice as a whole, this equals 198 additional consultations. Hence, for every 100 intended changes there were 70 successes and 16 additional consultations were generated.


Scottish prescribing analysis data showed that the practice had been prescribing above the health board average. The practitioners intentionally timed their attempt to reduce their costs to maximise the savings from the 1993-4 drug budget. Table II shows the success achieved in becoming average volume prescribers (8% to 1%), and below average prescribers for cost (20% to -8%) from before to after the introduction of fundholding. The proportion of generic prescribing rose from 36.8% to 57.6%.


Practice prescribing compared with health board averages before and after changes in prescribing

View this table:

Detailed analysis of the data available in defined daily dose form (which covered 90% by cost) revealed more about the nature of the changes which had occurred. Table III shows volume dispensed, cost per defined daily dose, and total cost for the comparable six month periods June to November in 1992 v 1993.


Practice prescribing data for June to November 1992–3

View this table:

The total number of treatment days fell by 67674 (10.7%). Although the repeat prescriptions intentionally stopped could account for 16000 days (2.5%) at most, an additional 51674 days' reduction occurred. Some of the fall was due to the adoption of shorter courses of treatment, especially non-steroidal analgesics for acute injuries, but it seems there has been an unplanned swing towards tighter prescribing in general.

The change in average cost per day's treatment is shown in the third column. This shows the savings made through the use of cheaper preparations after the introduction of a generic practice formulary. The average cost per day's treatment fell by 3.1p from 32.3p per day in 1992 (9.4%). This figure excludes the impact of a reduction in the volume dispensed.

The total cost of dispensed medication included in the defined daily dose calculations fell by pounds sterling43087 (24%) between the two periods analysed. The fourth column shows that the greatest savings were made from ulcer healing (pounds sterling9337), musculoskeletal (pounds sterling8714), and vasodilating drugs (pounds sterling7267). Expenditure in some drug groups increased—notably, hormone and asthma treatments.


A total of 33 detailed interviews were conducted with 17 patients whose treatment had been changed or stopped altogether. These highlighted those aspects of the experience which had been upsetting. Although therapeutic problems were sometimes uppermost (for instance, the stopping of Zovirax cream for cold sores), more commonly it was a failure of communication which underlaid any annoyance. Almost all patients interviewed were aware that saving money was at least part of the reason for the change; none, however, felt that a trial of a cheaper alternative was unreasonable. The outcome was much more likely to be a willing trial of the new treatment if patients thought that they understood the reason for the change.


The practice studied has proved to be an extreme example of the changes in prescribing habits which can be achieved. At a time when NHS drug expenditure is increasing at 14% a year it has achieved an absolute saving of 24%, moved from high to below average cost prescribing, and caused less upset to patients than it expected. These changes are considerably in excess of other reported practices, few of which achieved any absolute saving.2 3 Three questions must be considered: were they abnormally high prescribers before the change; has their prescribing become unsafe; and should others try to emulate them?

It has been suggested that practices may attempt to maximise their drug budgets through generous prescribing in the run up to fundholding.4 6 This practice consciously decided to delay introduction of measures to save costs until becoming a fundholder and knew that as a high prescriber (around 20% above health board average) it had room for manoeuvre. Indeed, this was seen as the prime incentive to join the scheme through which improvements in service could be funded, and the local medical prescribing adviser's previous efforts to reduce costs had been resisted. Historical data are difficult to compare because of changes in the practice's prescribing interval and health board's systems for collection of data so it is unclear whether an unusual rise occurred during the budget setting period. The agreed budget, at pounds sterling88.96 per head, is generous by comparison with others reported but the Tayside average actual expenditure in 1993-4 was pounds sterling81 for fundholders and pounds sterling91 for non-fundholders (Tayside Health Board, personal communication). The absolute saving of pounds sterling137712 should be seen in this light.7 It is impossible to say precisely which service improvements were funded through this saving but additional staff costs included increased psychology (pounds sterling20000), occupational therapy (pounds sterling17000), physiotherapy (pounds sterling13000), and chiropody (pounds sterling12000). About pounds sterling94000 was invested in practice improvements, minor surgery, and ophthalmology and audiology equipment. Future developments include plans for purchasing in 1995 a vehicle for transporting patients.

In the absence of any accepted standards for appropriate prescribing by general practitioners it is not clear whether this practice's prescribing has deteriorated or improved. The number of items prescribed remains 1% above the health board average and the proportion of generic prescribing increased from 36.8% to 57.6% in line with Audit Commission recommendations (Scottish average in 1991, 34.7%). The new prescribing level seems to be clinically reasonable.

The dissatisfaction expressed in the questionnaire was explored at interview and seemed to reflect two related issues. Firstly, power—patients feel impotent if they perceive a change has been forced on them. Secondly, communication—patients want to feel cared for by their general practitioner. They will accept major change if it is delivered in a manner which makes them feel valued. Such an approach also helps to avoid much of the negative feeling generated by the change itself and is a wise investment for the future doctorpatient relationship, which can easily be damaged through these manoeuvres.

Should other practices follow suit? The motivation for this practice to succeed was undoubtedly financial and the workload generated was judged by it to be worth while in the context of becoming a fundholder and producing a considerable saving. Non-fundholding practices should consider carefully the timing of any such moves with this in mind.

This practice may have been atypical. It has been extremely successful in introducing many changes rapidly and is now confident about doing so. As a result a clinical pharmacist is now working within the practice as part of a research project to investigate the impact of such collaboration. The rapid pace of developments has produced a feeling of “being on the winning side,” which has itself enhanced morale and encouraged further progress, such as the introduction of bar codes for patients' notes.

There is increasing evidence that fundholding practices do limit prescribing costs more successfully than non-fundholders, although this effect may be an artefact as a result of other practices maintaining traditional prescribing patterns in case a drug budget is forced on them. There is also some evidence that the quality of prescribing improves with fundholding—for example, greater use of inhaled steroids. The experience of the practice observed in this study has indicated what it is possible to achieve. Considerable financial savings were made at an acceptable cost in terms of effort and upset caused, although the latter could be reduced considerably with care. Without accepted standards, however, it is not clear when rational prescribing would become cheap prescribing if the financial incentives become too great and no standards are set.

We thank the general practitioners and staff at Downfield Surgery for the help they provided, M Maxwell of the department of general practice, Edinburgh University, for converting prescribing data to defined daily dose format, and the pharmacists who recruited subjects.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
View Abstract