Full text of DOI:10.1093/fampra/18.1.9

Family Practice
© Oxford University Press 2001
Vol. 18, No. 1
Printed in Great Britain
Is it possible to decrease antibiotic prescribing
in primary care? An analysis of outcomes in the
management of patients with sore throats
Claire M Cox and Marion Jones
Cox CM and Jones M. Is it possible to decrease antibiotic prescribing in primary care? An analysis
of outcomes in the management of patients with sore throats. Family Practice 2001; 18: 9–13.
Objectives. The aim of this study was to assess whether it was possible to change clinical practice
through the introduction of an evidence-based protocol for the management of sore throats. The
impact of the changed clinical practice on patients was also assessed.
Method. An observational study was carried out comparing management of patients with
sore throats in a semi-rural general practice (14 000 patients), before and after the introduction
of an evidence-based protocol, using a multidisciplinary approach. All patients over the age of
2 years presenting with sore throats as their chief presenting complaint were included in the
study which ran for two 6-month periods: February–July 1997 (n = 435) and February–July 1998
(n = 350). Antibiotic prescribing rates, consultation and reconsultation rates, duration of sore
throat, analgesia requirements and satisfaction were assessed.
Results. Antibiotic prescribing was significantly reduced in period 2 (56% compared with 19%).
Consultation rates decreased by 19% in period 2. Median number of days to recovery of sore
throats was the same in both groups. Reconsultation rates and dissatisfaction rates were the
same in both groups.
Conclusion. Using a multidisciplinary approach, it is possible to reduce antibiotic prescribing
for sore throats significantly without adversely affecting outcome.
Keywords. Antibiotics, prescribing, primary care, sore throat.
protocol, established in 1992, in our practice (Fig. 1).
Discovering unacceptably high levels of antibiotic pre-
Introduction
Two hundred consultations per 1000 population take
place each year for minor respiratory illnesses, and about
one-third of these are specifically for sore throats.1 Up
to 80% of patients who present with sore throats are
treated with an antibiotic.2 The cost of consultations alone
for sore throats is ~£60 million a year.3 There are also
concerns about increasing bacterial resistance to anti-
biotics which recently has been addressed by the House
of Lords Select Committee on Science and Technology.4
Work has also suggested that prescribing antibiotics
has a medicalizing effect on minor illness and tends to
increase reconsultation rates.5,6
scribing, we then revised the protocol.
The new protocol (Fig. 2) incorporated research
evidence, suggesting little benefit of antibiotics in un-
complicated sore throats.5,7 The aim of the project was
to assess whether it was possible, by these means, to
decrease antibiotic prescribing and adopt evidence-based
practice in primary care. The impact of changing to an
evidence-based, low antibiotic policy on patient accept-
ability, recovery and consultation rates was also assessed.
Method
This project analyses the management of sore throats
by GPs and practice nurses based on a ‘common practice’
Sore throats have been managed in our practice (14 000
patients, semi-rural) by GPs and practice nurses accord-
ing to agreed protocols since 1992. All patients aged
2 years and over presenting to the doctor or practice
nurse in normal working hours with a sore throat as their
chief presenting complaint were recorded for two 6-month
periods (February–July 1997 and February–July 1998).
Received 10 March 2000; Revised 13 June 2000; Accepted
5 September 2000.
The Swan Surgery, Swan Street, Petersfield, Hampshire
GU32 3AB, UK.
9

Family Practice—an international journal
10
The study included patients presenting with typical
pharyngitis and tonsillitis, but not those in whom these
were incidental findings, or not the chief symptom.
The first period (period 1, February–July 1997)
covered the original sore throat protocol, and the second
covered the revised ‘low antibiotic’ evidence-based
protocol (period 2, February–July 1998).
The initial ‘common practice’ protocol (Fig. 1) had been
achieved by a consensus amongst GPs in the practice and
was not evidence based. To achieve the evidence-based
protocol, practice management, practice nurses and GPs
were circulated with data from period 1, along with key
articles from a literature search on the use of antibiotics
in sore throats.3,5,7,8 A practice meeting was then held with
doctors, nurses and managers working in small groups to
devise a new evidence-based protocol. These protocols
were then amalgamated and consensus achieved (Fig. 2).
Patients were asked to sign a consent form prior to
entering the study (period 1 and period 2). Basic patient
details were recorded as well as symptomatology and
duration of illness. All patients were examined and treat-
ment outcomes recorded.
were made to assess the volume of analgesia required by
the patient. Reconsultation rates were also recorded and
any continuing symptoms were noted. Patients who had
continuing sore throat symptoms at this contact were
contacted again at 1 month to assess their recovery and
any serious sequelae.
All patients consulting with sore throats were given
an explanatory leaflet in period 2 explaining the lack of
evidence to support antibiotic usage in uncomplicated
sore throats.
Data were analysed as a case-controlled study. For
categorical outcomes, the two groups were compared by
means of a chi-squared test. For quantitative outcomes,
a two-sample t-test was used to compare mean values. A
non-parametric alternative (Wilcoxon–Mann–Whitney
sum of rank tests) was used for data that could not be
assumed to be distributed normally.
Ethics committee approval was sought and granted.
Results
At 5–7 days, one unblinded researcher initiated contact
with each patient by telephone. Data were elucidated con-
cerning any continuing sore throat symptom. Attempts
A total of 435 patients presented with sore throats as their
chief complaint in study period 1, whilst 350 patients
presented in period 2 (consultation rates of 62 con-
sultations/1000 patients/year and 50 consultations/1000
patients/year, respectively). This is equivalent to a 19%
decrease in demand for appointments for this condition.
In period 1, 43% of patients (n = 188) saw the nurse and,
of those patients, 27% (n = 51) were referred through to
the doctor. In period 2, the respective figures are 36%
(n = 125) and 14% (n = 17).
A similar sex distribution appeared in the groups
in periods 1 and 2 (68% female compared with 70%).
Identical numbers of patients (10%) had tonsillary
exudate in period 1 (n = 42) and period 2 (n = 32).
Ninety per cent (n = 392) of patients were followed up
at 5–7 days by the independent researcher in period 1
compared with 99% (n = 346) in period 2.
During the study, overall antibiotic prescription was
decreased from 56 to 19% (Table 1). Data from our
Primary Care Commissioning Group (East Hampshire)
for the periods March–August 1997 and March–August
1998 (unfortunately February to July data were not
available) showed that our practice’s penicillin V pre-
scribing was reduced by 36.7% more than the average
for our Primary Care Group (Katie Hovenden, personal
communication).
FIGURE 1 Common practice protocol for sore throats
(period 1)
Analysing the antibiotic prescribing in more detail in
period 1, 83% (n = 35) of patients with tonsillary exudate
received antibiotics compared with 25% (n = 8) in
period 2. Nurses independently issued antibiotics to
42% (n = 58) of patients seen entirely by them (n = 137)
in period 1 and to 4% (n = 4) of patients seen entirely by
them (n = 108) in period 2. Of the patients referred by
the nurse to the doctor, 88% (n = 45) received antibiotics
in period 1 compared with 71% (n = 12) in period 2.
FIGURE 2 Evidence-based protocol (period 2)

Reducing antibiotic prescribing for sore throats
11
TABLE 1 Comparative outcomes for periods 1 and 2
Patient characteristic
Period 1
February–July 1997
Period 2
February–July 1998
(n = 350)
Rate difference
(95% confidence
interval)
Probability
(n = 435)
Prescribed
antibiotics
56% (n = 243)
19% (n = 65)
0.373 (0.309–0.433)
P Ͻ 0.0001***
TABLE 2 Follow-up at 5–7 days (and 28 days)
Patient characteristic
Sore throat settled
Period 1
Period 2
(n = 346)
Rate difference
(95% confidence interval)
Probability
(n = 392)
77% (n = 303)
4.0
70% (n = 242)
0.074 (0.010 to 0.137)
P = 0.023*
P = 0.167
Median no. of days for sore throat
to settle (includes data from
28 day follow-up)
4.0
Back to normal health
Required analgesia
58% (n = 227)
80% (n = 315)
52% (n = 205)
2.04
56% (n = 194)
87% (n = 300)
58% (n = 200)
2.55
0.016 (–0.056 to 0.087)
–0.063 (–0.117 to –0.010)
0.055 (–0.126 to 0.017)
P = 0.665
P = 0.021*
P = 0.134
P = 0.019*
Required regular analgesia
Mean no. of days regular
analgesia required
Reconsultation rate
Dissatisfaction rate
4% (n = 16)
5% (n = 21)
63% (n = 249)
6% (n = 19)
4% (n = 13)
87% (n = 301)
–0.014 (–0.047 to 0.017)
0.016 (–0.015 to 0.047)
–0.235 (–0.294 to –0.175)
P = 0.369
P = 0.301
Recollection of advice about
home remedies
P Ͻ 0.0001***
No. of patients who felt they
could ask for an explanation
88% (n = 345)
83% (n = 289)
0.040 (–0.011 to 0.093)
P = 0.124
Significantly fewer patients felt that their sore throat
had settled by 5–7 days in period 2 and more patients
felt the need to take some analgesia (Table 2). However,
analysing the data from both 5–7 days and 1-month
follow-up, the median number of days for the sore throat
to settle was the same for both periods 1 and 2. Patients’
perception of being back to normal health was also the
same in both groups.
Despite the significantly lower usage of antibiotics,
reconsultation rates were unaffected and satisfaction
rates remained high. This is supported by similar num-
bers of patients feeling they could ask for an explanation
of their symptoms and treatment in both groups. Patients
recalled receiving more advice on home remedies during
period 2 when fewer antibiotics were given.
Only two serious sequelae were recorded during the
trial, one in period 1 and the other in period 2. In period 1,
a patient developed quinsy. She had been given penicillin
by the GP at initial contact and settled without hospital
admission, though subsequently underwent tonsillectomy.
A second case of quinsy occurred in period 2. This was in
an 11-year-old boy who was not toxic and had no pus on
his tonsils at presentation. He was not given antibiotics,
but 2 days later presented with quinsy. This settled
with oral penicillin and no hospital admission or other
sequelae developed.
Discussion
Research has shown that antibiotics have little impact on
the recovery from sore throats in non-toxic patients.5–7
In an era when serious sequelae are declining, it has been
stated that there is more chance of dying from anaphylaxis
from antibiotics than from suffering severe consequences
from an untreated sore throat.9 The Standing Medical
Advisory Committee recently has recommended that no
antibiotics should be prescribed for viral throats.10
The aim of this study was to see if it is practicable
to decrease antibiotic prescribing in primary care and,
if so, whether this decrease has any impact on patient
outcomes.
The two main findings from the study were that anti-
biotic prescribing and consultation rates were reduced
in period 2 without there being any adverse affect on
patient outcome. The issue to be addressed is how

Family Practice—an international journal
12
generalizable are these findings and what were the
factors which engendered the change.
protocol. Therefore, we used an educational process,
which has been shown to be effective in other settings,
and our success in reducing antibiotic prescribing may
well be attributable to this.
Our data on recovery rates (median number of days
for sore throat to settle, patients’ perception of being
back to normal health) tend to support previous work, in
that decreasing antibiotic prescribing does not adversely
affect outcome.7 The greater volume of analgesia taken
by patients in period 2 may reflect the advice given on
our patient leaflet about symptomatic treatment for sore
throats.
One of the hurdles to changing clinical practice often
is that we perceive that patients will find the change
unacceptable. However, in our study, satisfaction rates
were consistent. The evidence was that in period 2 more
advice and less medicine was dispensed, and perhaps this
should be the message for engineering change in pre-
scribing patterns in primary care.
The study showed that there was a 19% reduction
in demand for consultations for sore throats following
the introduction of the evidence-based protocol. There
are two possible explanations for this. It could well be
that the incidence of sore throats in the community was
less in 1998 than in 1997, or that media-led education
campaigns resulted in fewer patients with sore throats
choosing to consult the doctor. The other possible ex-
planation is that the different management they received
at the practice affected their decision to consult with
future sore throats. A limitation of the methodological
design of this study is that it results in us not having
enough access to the background incidence of sore
throats in the community. Therefore, any inference from
the reduced consultation rate must be guarded. How-
ever, we do know, from other published work, that the
type of interventions we implemented (e.g. decreasing
antibiotic prescribing) can result in demedicalizing
self-limiting conditions5 and that patient information
leaflets11 are successful in increasing patient knowledge.
Both of these factors may well have influenced the
patients’ decision to consult.
Conclusion
Using a multidisciplinary approach, it is possible to
reduce antibiotic prescribing for sore throats signifi-
cantly without adversely affecting outcome. There is also
evidence that patient education about the role of anti-
biotics may limit their need to consult for self-limiting
illness.
The study showed that there was a 37% reduction
in antibiotic prescribing after the introduction of the
evidence-based protocol. This compares favourably with
other studies which quote figures of 11 and 10% when
doctors set themselves guidelines to reduce antibiotic
prescribing for coughs and wheezy chests.12
Moreover, it appears that practice nurses are adept at
following agreed protocols for minor illnesses, having
decreased their independently initiated scripts from
42% in period 1 to 4% in period 2. The question to be
addressed is what factors were responsible for such a
significant change in antibiotic prescribing.
Acknowledgements
With thanks to the Portsmouth RDSU.
Inevitably, the involvement in a research project may
in itself be an incentive to change doctor’s prescribing
habits—but this would obviously have impacted on the
results of other published work too.13 The importance of
the patient information leaflet on patient knowledge has
been described already,12 and this in itself may have been
helpful in reducing the doctor’s perception of patient
pressure to prescribe. However, we feel that one of the
main reasons for the success in implementing evidence-
based practice is that we employed a multidisciplinary
approach with proven educational methods to engineer
the change. Studies have shown that active educational
methods, such as the small group educational process,
are more effective than passive methods at achieving
sustained change in practice.13 Other work has sug-
gested that peer education and feedback on perform-
ance can promote behavioural change.14 In our project, we
involved the whole team in the development of the
evidence-based protocol, based on a literature review
and feedback on their performance in period 1. Multi-
disciplinary small groups were used to develop the
References
1
Royal College of General Practitioners. Trends in National
Morbidity. Occasional Paper 3, 27. London: Royal College of
General Practitioners, 1976.
Howie JG, Foggo BA. Antibiotics, sore throats and rheumatic fever.
2
J R Coll Gen Pract 1985, 35: 223–224.
Little P, Williamson I. Sore throat management in general practice.
3
Fam Pract 1995; 13: 317–321.
House of Lords Select Committee on Science and Technology. 7th
4
Report (Session 1997–98) Resistance to Antibiotics and other
Antimicrobial Agents. London: HMSO, 1998.
Little P, Gould C, Williamson I, Warner G, Gantley M, Kinmouth A.
5
Reattendance and complications in a randomised trial of pre-
scribing strategies for sore throat. The medicalising effect of
prescribing antibiotics. Br Med J 1997; 315: 350–352.
6
Little P, Williamson I, Warner G, Gould C, Gantley M, Kinmouth A.
Open randomised trial of prescribing strategies in managing
sore throat. Br Med J 1997; 314: 722–727.
Del Mar CB, Galsziou PP. Antibiotics for the symptoms and com-
7
plications of sore throat. In Douglas R, Bridges, Webb C et al.
(eds). Acute Respiratory Infections Module of the Cochrane
Database of Systematic Reviews. The Cochrane Collaboration.
Issue 3. Oxford: Update software, 1997.
8
Drugs and Therapeutics Bulletin 1995; 33(2): 9–12.

Reducing antibiotic prescribing for sore throats
13
9
12
Little P, Williamson I. Are antibiotics appropriate for sore throats?
Br Med J 1994; 309: 1010–1011.
Standing Medical Advisory Committee Subgroup on Antimicrobial
Resistance (Department of Health). The Path of Least
Resistance. Main report. London: HMSO, 1998.
Little P, Griffin S, Kelly J, Dickson N, Sadler C. Effect of educational
leaflets and questions on knowledge of contraception in women
taking the combined contraceptive pill: randomised controlled
trial. Br Med J 1998; 316: 1948–1952.
North of England Study of Standards and Performance in General
Practice. Medical audit in general practice 1; effect on doctors
clinical behaviour for common conditions. Br Med J 1992; 304:
1480–1484.
Lipman T, Price D. Decision making, evidence, audit and education:
case study of antibiotic prescribing in general practice. Br Med J
2000; 320: 1114–1118.
10
13
14
11
Belongia E, Schwartz B. Strategies for promoting judicious use of
antibiotics by doctors and patients. Br Med J 1998; 317: 668–671.