Family Practice

Oxford University Press 2000

Vol. 17, No. 1

Printed in Great Britain

A study of factors associated with cost and

variation in prescribing among GPs
Patricia Carthy, Ian Harveya, Richard Brawnb and Chris Watkinsc
Carthy P, Harvey I, Brawn R and Watkins C. A study of factors associated with cost and variation
in prescribing among GPs. Family Practice 2000; 17: 3641.
Background. Inappropriate prescribing has the potential to harm both the individual and
society. Previous research has identified doctor or demographic characteristics that influence
prescribing variation but which were not amenable to change.
Objectives. To identify modifiable factors associated with GP prescribing variance and cost.
Method. Qualitative research methods were used in semi-structured taped interviews with 17
GPs in Avon, South West NHS Region, UK.
Results. GPs considered themselves cautious and conservative prescribers. Prescribing decisions
often were justified by the prescriber, despite conflicting clinical or cost arguments. A personally
developed drug formulary was used to reduce dilemmas potentially associated with prescribing
uncertainty. Willingness to reflect upon, and measure, prescribing habits against set professional
standards varied considerably. The absence of monitoring mechanisms of prescribing decisions,
coupled with under utilization of the community pharmacist, resulted in uncertain prescribing
outcomes. Some GPs found it difficult to keep up to date professionally due to perceived time
constraints. Excessive patient demand was considered to influence their prescribing, but GPs
stated that they were not unduly influenced by the drug representative.
Conclusions. Prescribing makes a considerable impact on health and budgets and yet remains
a contentious issue. Improved partnerships between patient, doctor and pharmacist must be
established. Better prescribing decision monitoring and support through policy development
and educational intervention is needed to reduce prescribing uncertainty. Newly established
Primary Care Groups may need to reflect upon the difficulties facing prescribers, particularly
when prescribing within cash-limited budgets, to avoid discord between prescribing behaviour
and local policy development.
Keywords. Costs, decision making, prescribing variation, support.

There is also concern for suboptimal, or indeed

harmful prescribing which fails to meet the highest
professional standards of care. This is particularly so for
vulnerable groups such as the elderly who receive nearly
half of all prescription items.3
Critics argue that weak regulation of the pharmaceutical industry allows dubious drug marketing practices
to flourish that fuel drug costs and encourage inappropriate prescribing.4

Introduction
Drugs prescribed by the GP account for ~10% of NHS
expenditure and half of the total cost of family health
services.1 It is argued that new technological advances,
particularly those involving expensive medications,
increased public expectation and an ageing population
are largely responsible for a soaring national drugs
bill.2

Aims

Received 29 March 1999; Revised 26 July 1999; Accepted

6 September 1999.
Research and Development Support Unit, Salisbury Healthcare
NHS Trust, Salisbury District Hospital, Salisbury, Wiltshire
SP2 8BJ, aDepartment of Epidemiology and Public Health,
University of East Anglia, bDepartment of Education, University
of Bristol and cBackwell and Nailsea Medical Group, UK.

Despite a wealth of literature, there remains a lack of

real understanding of factors that influence prescribing
decisions. Here we describe the qualitative phase of a
project that aims to identify modifiable factors associated
36

Variation and costs of GP prescribing

with prescribing cost and variation. Findings from a subsequent quantitative phase of the study will be reported
separately.

Method
Semi-structured taped interviews were conducted with
17 Avon GPs to obtain views about their prescribing
habits. The GPs were identified from practice Prescribing Analyses and Cost (PACT) data.
Recruitment of interviewees
Following a pilot exercise with two GP members from
the Department of Social Medicine, 15 GPs were
selected randomly from a sampling frame of 30 GPs. A
letter was sent to selected GPs inviting them to
participate in the study. GPs who agreed to take part in
the study subsequently were interviewed by an experienced researcher (P.C.).
Interview proforma
Key headings of the interview proforma included: the
determinants of prescribing; decisions on whether or not
to prescribe; the choice of drug; and desired decisionmaking support.
Data analysis
Interviews were audio taped, coded and transcribed.
Qualitative research principles5 were used to identify,
log and list emerging themes, that were considered to
influence prescriber decisions, by the lead researcher
(P.C.). Themes were identified, grouped and crossreferenced by the research team in a series of meetings.
Issues included: managerial or prescribing policy; doctor
patient relationship; professional knowledge; and educational support. Additional subgroups were established,
compared and analysed, using criteria to discard factors
that were not measurable or were potentially modifiable
through an educational intervention.

37

Key influences upon prescribing

Prescribing uncertainty and GP knowledge. Prescribers
used a personal head-held drug formulary; a unique,
if somewhat idiosyncratic individual index, to decide
whether and what to prescribe. The formulary was established during medical training and shaped by colleagues,
patients, policy and own experience in general practice.
Prescribing doubts usually were associated with adverse
drug effects: whether the decision to prescribe a particular
drug had a potentially detrimental effect; uncertain or
ill-defined diagnoses; or treatments for children or the
elderly.
Drugs adopted during early training were often
retained, and still perceived to meet the needs of patients
and working practices:
. . . you are influenced by other colleagues and
superiors when youre going through training and I
think gradually as time goes on, you become, you
know, its natural that you become more set in your
ways. I think its harder and harder for pharmaceutical companies and FHSA advisors to influence
you because . . . as time goes on you actually tend to
use the drugs more and more and you become
familiar with, you become familiar with the doses
and so you . . . you tend to write the same drugs over
and over again. So perhaps we become more conservative as we get older as well.
Asked to describe his prescribing:
I would say a conservative prescriber really, I
never really prescribe a drug thats fresh onto the
market, because thats just my nature. I always go
for a tried and tested drug . . . em . . . I think its easy
to be caught out by a new drug that comes onto
the market and finally after it has been used for a
while, problems start to ensue and you then realise
there could be other problems involved . . . Its very
unusual for a new drug to break into my own
personal formulary! (mq66 code)
This cautious watchful, wait and see policy was stated
by many:

Results
Interviewees
Between August 1995 and April 1996, 17 GPs (15
randomly selected and two from a pilot exercise) were
interviewed, one of whom was female. GPs were aged
between 37 and 56 years with an average age of 43 years.
Five city, six urban and six rural practices were
represented, five of which were fundholding and two
single-handed. Seven GPs were from high prescribing
cost practices, four from medium cost practices and the
remaining six from low cost practices (as defined by
overall prescribing cost PACT data). The length of interview ranged from 45 minutes to 1 hour and 20 minutes,
with an average of 1 hour duration.

I dont think that changes that were introduced

post 1990 (new national GP contract) have had
much effect . . . or an impact on my prescribing
habits. As I see it, the situation as far as prescribing
and treatment is concerned, on a one to one basis,
I havent changed my views, irrespective of prescribing figures and so on. I see a patient who has a
need. . . and I address that need and I try to provide
as constructively . . . as comprehensively for that
need and thats the issue that presents to me whenever a patient sits on that chair. (mq296 code)
Generally, GPs expressed a desire to keep up-to-date
professionally, particularly with regard to prescribing.
Some described difficulties experienced during patient

38

Family Practicean international journal

consultations, when they were required to access relevant drug information. The immediacy required to make
a speedy judgement about the best possible course of
action for particular patients was perceived as difficult.
GPs expressed concern about competing demands
on their time (viewed as a potential threat to the quality
of prescribing decisions). Time constraints were cited as
reasons for not translating good intentions into meaningful action. Others described a feeling of unease, on
reflection, about the action they had taken. A balance
appears to have been struck, in terms of desirable continuing professional development, between what was
preferred, feasible and attainable, in a demand-led service. Some GPs found it was simply a case of not knowing
what, if any, educational input was needed.
When asked to describe factors contributing to
perceived prescribing uncertainty or lack of knowledge,
many GPs were unable to describe specific problems.
Some suggested that it was a combination of factors; accessing information, but with insufficient time in which
to keep professionally up-to-date; due to increasing
workload; the changing culture of general practice; with
patients apparently more aware of health issues and
generally more questioning and challenging. However,
when asked about suitable support for their prescribing
decisions, many were unable to give a precise answer.
Most wanted more available time in which to take stock.
Prescribing was not viewed as a priority area for many,
when compared with other aspects of work. The changing face of general practice, the demands of the NHS
reforms and the diversity of general practice loomed far
larger. Keeping pace with these competing demands
resulted in insufficient time available to examine their
day-to-day practice.
Prescribing support. The British National Formulary
(BNF) and Monthly Index of Medical Specialties (MIMS)
were frequently cited as useful support materials. Decision
support computer packages, such as PRODIGY (Prescribing Rationally with Decision support in General
Practice), had been experimented with by some and
found to be helpful, albeit with limited adaptability and
flexibility when patients presented with multiple
conditions.
The community pharmacists role was both valued and
dismissed by GPs. Where a professional partnership had
been established, benefits for both were evident. GPs
acknowledged their fail-safe role in identifying prescription error but, in general, the pharmacists potential
role, in terms of decision support, was rarely exploited.
Perceived pressure to do something in response to a
difficult situation, rather than being viewed as incompetent or not having a ready solution, was highlighted by
some. Difficulties also occurred when seemingly endless
drug regimes were exhausted, repeated advice to patients
went unheeded or where prescribing for heartsink
patients proved particularly onerous.

GPs felt that they generally prescribed appropriately:

prescribing decisions deemed inappropriate were
described as the human side of general practice. Most
acknowledged that from time to time they hadnt got it
right. There seemed to be acceptance of a decision
allowance, a grey area of prescribing, clearly acceptable
to the prescriber, in response to unusual individual
circumstances:
As an individual, I mean there are always exceptions . . . the businessman who is flying off to the
States, whos got a cold in the head and hes terrified
hes going to get sinusitis. If you explain things a bit
more, and if you point out to them, if you (the
patient) think you are going to get better, then you
certainly shouldnt be bothering with them.
(mq496 code)
Prescriptions for antibiotics, with instructions to wait
and see if further symptoms appear, before redeeming
the prescription, was common practice. Whether doctors resisted perceived patient demand depended largely
on the amount of available time or their inclination to
explain (or need to continually repeat) reasons for not
prescribing. Competing work, sustaining effort or interest over time or a poorly established relationship with
the patient exacerbated the problem.
Peer influences. The hospital consultant was viewed
as a valuable source of advice and support. This was
especially true in the use of new drugs, as a direct result
of secondary care patient treatment plans, or from
attendance at a specialist-led GP seminar:
I mean there are terrific numbers (antidepressants)
and theres all these new SSRIs coming out and you
know, youre thinking, well which ones better than
the rest? And you end up sort of peeking through
little summaries of them and you look to people you
respect . . . we met with the consultant psychiatrist
. . . because we wanted to learn more about his
thinking about them. (mq496 code)
Others thought consultants were unaware of prescribing cost differences between primary and secondary
care settings or advocated specific drug regimes without
full patient histories:
A patient with prostatitis, who had been prescribed
Desmopressin by the hospital consultant presented
(to the GP ) . . . you think hang on a minute here,
this chap who is hypertensive was on calcium
antagonists and you look it up and sure enough, its
contraindicated in hypertension. What do you say?
Theres always hassle but I mean, its easy to say to
the patient the consultant said yes . . . heres your
prescription. (mq496 code)
Use of new drugs was considered to be cautious and
conservative but could be influenced by failure of tried

Variation and costs of GP prescribing

and tested therapies. Specific reports or articles in

professional journals could also trigger use of new drugs,
although research studies in professional journals were
criticized for their user unfriendliness, dullness and
length.
Few prescribers met with colleagues to compare their
prescribing. Reluctance to compare prescribing was
due to either confidence in prescribing ability or fear of
criticism.
One GP described hearing a colleague being
criticized, for goodness sake, youre not still using
XXXXXX (type of drug) are you!
The presence of a locum or trainee or registration of a
new patient (with different drug regimens) could also
trigger prescribing change.
Influences of prescribing policy. The concept of the
practice-based formulary was well supported although
implementation appeared problematic. When asked
whether a practice formulary was used, most GPs stated,
we are going to develop one or we keep talking about
introducing one. Prescribing advice from health
authorities was well received by some and dismissed by
others:
I mean I get PACT data and you know, I think it
does influence. I think my goodness, Im . . . you
know . . . theres this huge amount being spent on
say one drug or em . . . I think maybe I could
influence that . . . maybe theres an alternative drug
which would be effective and yet cheaper . . . I think
a move to generic prescribing is for the better and I
am very much in favour of that. Not wholesale
change because there are circumstances when a
brand drug is important to use . . . just that (full
scale) generic prescribing is often detrimental; its
not always to the benefit. So I think a lot of GPs are
probably guilty of em, . . . profligate spending, I
mean, weve just got an open cheque book. (mq6
96 code)
Less enthusiastic GPs referred to advice from the
health authority as political or managerial meddling
they just want to cut costs.
Interestingly, where changes had been made, from
discussions of prescribing with health authority officials,
it was the process of discussion or the external facilitation,
rather than actual outcome, that was considered most
useful.
Influences of the patient. Prescribing for different
socio-economic groups was subject to change. One GP
described prescribing in a sister practice (situated in a
poor district) and compared it with that of the main
surgery (in an affluent middle class area) and acknowledged differences in prescribing between the surgeries.
Reflecting upon the reasons for the disparity, he felt that
certain groups of patient were better at compliance

39

or acted upon GPs advice where circumstances offered

better compliance.
Some GPs devised ways of saving money for poorer
patients. These involved juggling prescriptions against
costs of over the counter drugs, or involved prescriptions of greater volume, over time, to avoid multiple
prescription payments.
Treating children accompanied by difficult family
members could trigger overly cautious defensive reactions from the prescriber or inhibit medical examination
of the child.
Another GP described prescribing for children of
working mothers on a new housing estate:
There was a need for their kids to be fit and well to
be looked after by the childminder. This had a very
definite impact upon the ease or the speed with
which clinical presentation of minor illness takes
place. The GP quoted the mother, doctor fix the
kid up because Ive got to go to work and its got a
snotty nose and doesnt eat very well and I cant go
to work if theres a problem to fix.(mq296 code)
Excessive and unrealistic demands from the patients
frequently were cited. Media-prompted news could
provoke huge responses from patients:
I had somebody . . . who had been reading about
Prozac . . . she hadnt got a relationship that was
working out . . . she thought Prozac would sort out
her problems . . . it was a difficult consultation
because she didnt get the Prozac. (mq196 code)
The Patients Charter was viewed as bothersome, and
the phrase a pill for every ill was frequently quoted by
interviewees.
Prescribing costs. GPs appeared to support costreducing initiatives in principle, although some unease
was acknowledged. Tensions were usually associated
with perceived conflicts between professional boundaries and fiscal policy. All GPs bemoaned increased
administrative chores and excessive paper chasing,
much of which was felt to be time wasting. Several nonfundholding GPs felt that their efforts to reduce prescribing costs had gone unrecognized or unrewarded.
Some expressed regret about fundholding colleagues
and their financial incentive schemes, when compared
with their cost savings:
But basically Im a high cost prescriber . . . but
Im not . . . I dont feel uncomfortable about my
prescribing, I feel comfortable about it. Its difficult
when there are financial restrictions because
Ive seen that there (sic) are a little bit contrary to
patients welfare but there are pressures for cheaper
prescribing and not necessarily in the patients
best interests. For example, in the management of
hypertension, I use a lot of calcium antagonists and

40

Family Practicean international journal

ACE inhibitors, which I think are good drugs but

Im pressurised into using betablockers and thiazide
diuretics which do have problems associated with
them. If youre looking at patients welfare, wellbeing, those things cause lethargy and precipitate
asthma and thiazides precipitate diabetes and gout
and there is dubious effect of betablockers on lipid
profiles and so on . . . I think its quite difficult at
times. (mq196 code)
GPs who used high amounts of generics but who still
exceeded their drug budgets, had different views:
In many respects our prescribing habits tend to be
more modern prescribing, rather than old fashioned
prescribing habits for cheap drugs . . . hypertension
as an example, cardiovascular drugs where we spend
a lot, and the items tend to be expensive . . . ACE
inhibitors . . . calcium blockers and things. Most of
the people who we are prescribing for will have
been reviewed and their medication will be changed
and updated, and in the process of doing that, we
tend to choose, lets say, newer, which tend to be
more expensive products, rather than people . . .
who say . . . have been on the same tablet for fifteen
to twenty years still taking Propanalol or something
which is six pence a bucket, and it may suit them
well, but we wouldnt necessarily initiate very much
prescribing of older drugs. (mq396 code)
Asked why their practice prescribed in this way, the
GP continued:
Well, in the case of hypertension, I mean there are
many drugs which just arent used any more, but
there are still plenty of people out there still on
them . . . if you go to a practice which has been going
for a long while with a stable population, youll find
lots of people on drugs like Methydopa. I havent
used Methydopa in many years . . . sort of went out
with the ark but are cheap prescribing with a significant effect on prescribing costs. I see myself as
somebody who is not afraid to prescribe, and if it
costs a lot of money, well thats just too bad . . . if its
justifiable, but if the condition doesnt require drugs
at all, and people respond to other interventions, or
may be no intervention at all if it is self limiting,
Im not afraid to say no, it doesnt need anything.
(mq396 code)
Apart from routine reviews of chronic conditions, prescribers rarely, if ever, monitored prescribing decisions,
the patient will always come back if it doesnt work.
Monitoring of prescribing decisions was supported, but
the process was deemed difficult to manage.
Influences of the pharmaceutical industry. Most GPs
felt that advice from drug representatives was selective
or contained half truths. Some technical data was

valued, but GPs placed restrictions on access, or number

and duration of visits. Drug company-sponsored study
sessions proved popular, due to the perceived quality of
invited expert speakers.
Most GPs appreciated that marketing techniques
could influence their prescribing but generally expressed
confidence in their ability to withstand commercial sales
pressure:
Well I think he (drug rep) has a very useful
influence. I see drug reps as useful people . . . in a
sense that they bring to my attention the fact that
they have a particular development and I hope they
see my role as being educational, to some extent. I
hope they see their role as being informative and
not simply as peddlers of wares. (mq396 code)

Discussion
Despite some idiosyncratic prescribing, this group of
GPs felt comfortable with their prescribing. However,
use of the personal formulary has obvious drawbacks; it
endorses self-belief in prescribing ability through habit
and familiarity. Without scrutiny, the formulary may
establish and perpetuate poor prescribing patterns.6
Some prescribing habits appeared to defy attempts
from policy and managerial forces to change them, and
prescribing decisions were justified, despite conflicting
clinical or cost arguments.
GPs face numerous difficulties when prescribing
within cash-limited budgets, some of which potentially
influence prescribing behaviour. Newly established
Primary Care Groups may need to reflect upon these
difficulties in order to avoid further discord between
actual prescribing activity and local policy development.
Policy and management forces need to develop locally
appropriate and credible information, to enable prescribers to keep professionally updated. These measures
should be underpinned by adequate support and training that is derived from individual needs assessment.
Use of a visiting pharmacist is known to improve
GP prescribing, and yet support from pharmacists was
generally patchy or absent.7 There is growing recognition of the need for better collaboration between GPs
and pharmacists8 in order to enable the potential role of
the community pharmacist to be fully developed for use
in local training initiatives.
The development of new skills, such as computerized
decision making, should be tailored to individual need
to encourage all prescribers to engage in professional
development. These initiatives must take account of
varying perceptions of need that exist between older and
younger age groups.
Behavioural change is notoriously difficult, and GPs
differ and respond to different types of educational
models. Supportive interventions of rational prescribing,

Variation and costs of GP prescribing

therefore, need to be timed carefully and tailored to

individual prescribing beliefs and attitudes.
GPs who are more critical in their drug choice may
have lower prescribing costs.9 Prescribers knowledge,
selection and use of drugs in this group varied considerably. Overly cautious use of new or expensive drugs may
miss therapeutic opportunities, while high use of generic
drugs reduces prescribing costs but may take scant
account of individual circumstance.10
Critics argue that basic pharmacology, rather than
problem solving and practical application or audit, is
over emphasized during medical training, and largely
responsible for establishing poor prescribing habits that
subsequently prove difficult to change.11
There is cause for concern over professional literature
that is deemed dull and uninspiring. Credible, updated
drug information needs to be both interesting and easily
accessible for doctors to act upon. Prescribers reported
that they were not unduly influenced by the pharmaceutical industry, although this runs counter to the evidence.12
Excessive and unrealistic patient demand was cited
frequently by this group. Critics argue that this professional viewpoint is often exaggerated and misdirected.13
Prescribing decisions require sufficient timeand a
positive attitudein which to access, critically appraise
and synthesize information for the benefit of both patient
and doctor. Lack of critical appraisal potentially provides a licence to prescribe haphazardly. The paucity
of monitoring mechanisms makes it difficult to judge
whether a prescribing decision is appropriate.

their prescribing and compare it with clearly defined

quality outcome indicators.
Without support and monitoring, and encouragement
to forge better decision partnerships between doctor and
patient, and doctor and pharmacist, some prescribing
will, inevitably, remain suboptimal.

Acknowledgements
With grateful thanks to all the doctors who participated
in the study and Elizabeth Robinson, Medical Advisor,
Avon Health, Bristol. The study was supported by funding from the Research and Development Directorate of
the NHS Executive, South and West.

References
1

2
3

Conclusions
Prescribing decisions make a considerable impact on
health and national budgets and require complex personal
and professional judgements to be made about physical,
psychosocial and cost dimensions of health. Is it feasible
to expect clinical need to be determined, managerial and
fiscal policy to be balanced and complex drug information accessed and assimilated, within the confines of an
8-minute consultation, in a demand-led service?
Professional experience, and the use of the personal
formulary, may provide a suitable basis for change
models. Time must be set aside for GPs to reflect upon

41

10

11

12

13

Audit Commission. A Prescription for Improvement Towards More

Rational Prescribing in General Practice. London: HMSO, 1995.
Pallot P. The cost of living. Health Service J 1996; 30 May: 11.
The Royal College of Physicians. RCP Medication for Older People.
2nd edn. London: The Royal College of Physicians, 1997: 10.
Collier J. The Health Conspiracy How Doctors, The Drug Industry
and the Government Undermine our Health. London:
Hutchinson Ltd, 1989.
Berg BL. Qualitative Research Methods. Boston: Allyn and Bacon,
1989.
McGavock H, Webb CH, Johnson GD, Milligan E. Market
penetration of new drugs in one United Kingdom region:
implications for general practitioners and administrators.
Br Med J 1993; 307: 118120.
Avorn J, Sourmerai S. Improving drug therapy decisions through
educational outreach. N Engl J Med 1983; 308: 14571463.
Yeo GT. Educational visiting and hypnosedative prescribing in
general practice. Fam Pract 1994; 11: 5761.
Van de Poel G. A way of achieving more rational drug prescribing.
Int J Pharm Pract 1991; 1: 4548.
Ayanian J, Hauptman PJ, Guadagnoli E, Antman EM, Paslos CL,
McNeil BJ. Knowledge and practices of generalists and specialist physicians regarding drug therapy for acute myocardial
infarction. N Engl J Med 1994; 331: 11361142.
Walley T. Rational prescribing in primary carea new role for the
clinical pharmacologist? Br Med J 1993; 36: 1112.
Kingman S. Fisons investigations of excessive hospitality. Br Med J
1993; 307: 13011302.
Britten N. Patient demand for prescriptions; a view from the other
side. Fam Pract 1994; 11: 6266.