J Antimicrob Chemother

doi:10.1093/jac/dky137

Long-term effect of a practice-based intervention (HAPPY AUDIT)

aimed at reducing antibiotic prescribing in patients with respiratory
tract infections
Carl Llor1*, Lars Bjerrum2, José M. Molero3, Ana Moragas4, Beatriz González López-Valcárcel5, M. José Monedero6,
Manuel Gómez7, Marina Cid8, Juan de Dios Alcántara9, Josep M. Cots10, Joana M. Ribas11, Guillermo Garcı́a12,
Jesús Ortega13, Vicenta Pineda14, Gloria Guerra15 and Susana Munuera16 on behalf of the HAPPY AUDIT 3
Study Team†

1
Via Roma Health Centre, Barcelona, Spain; 2Section of General Practice and Research Unit for General Practice, Department of Public
Health, University of Copenhagen, Copenhagen, Denmark; 3San Andrés Health Centre, Madrid, Spain; 4University Rovira i Virgili, Jaume
I Health Centre, Tarragona, Spain; 5Department of Quantitative Methods for Economics and Management, University of Las Palmas,
Las Palmas de Gran Canaria, Spain; 6Rafalalena Health Centre, Castellón, Spain; 7Mirasierra Health Centre, Madrid, Spain; 8Teis Health
Centre, Vigo, Spain; 9Bollulos Par del Condado Health Centre, Huelva, Spain; 10University of Barcelona, La Marina Health Centre,
Barcelona, Spain; 11Hospital Sant Llàtzer, Palma, Spain; 12La Calzada II Health Centre, Gijón, Spain; 13Rincón de Soto Health Centre, La
Rioja, Spain; 14Serrerı́a I Health Centre, Valencia, Spain; 15Escaleritas Health Centre, Las Palmas de Gran Canaria, Spain; 16Son Pisà
Health Centre, Palma, Spain

*Corresponding author. Tel: !34-935542878; Fax: !34-935542881; E-mail: carles.llor@gmail.com

†Other members are listed in the Acknowledgements section.

Received 8 November 2017; returned 8 January 2018; revised 19 March 2018; accepted 20 March 2018

Objectives: Few studies have evaluated the long-term effects of educational interventions on antibiotic pre-
scription and the results are controversial. This study was aimed at assessing the effect of a multifaceted
practice-based intervention carried out 6 years earlier on current antibiotic prescription for respiratory tract infec-
tions (RTIs).
Methods: The 210 general practitioners (GPs) who completed the first two registrations in 2008 and 2009 were
invited to participate in a third registration. The intervention held before the second registration consisted of dis-
cussion about the first registration of results, appropriate use of antibiotics for RTIs, patient brochures, a work-
shop and the provision of rapid tests. As in the previous registrations, GPs were instructed to complete a template
for all the patients with RTIs during 15 working days in 2015. A new group of GPs from the same areas was also
invited to participate and acted as controls. A multilevel logistic regression analysis was performed considering
the prescription of antibiotics as the dependent variable.
Results: A total of 121 GPs included in the 2009 intervention (57.6%) and 117 control GPs registered 22247 RTIs.
On adjustment for covariables, compared with the antibiotic prescription observed just after the intervention,
GPs assigned to intervention prescribed slightly more antibiotics 6 years later albeit without statistically signifi-
cant differences (OR 1.08, 95% CI 0.89–1.31, P " 0.46), while GPs in the control group prescribed significantly
more antibiotics (OR 2.74, 95% CI 2.09–3.59, P , 0.001).
Conclusions: This study shows that a single multifaceted intervention continues to reduce antibiotic prescribing
6 years later.

Introduction countries with the highest rates of antibiotic prescribing, with a

Overuse of antibiotics has contributed to the development of anti-
microbial resistance.1 As most antibiotics are prescribed in primary
care and many of these prescriptions for mainly respiratory tract
infections (RTIs) are of questionable value, reducing inappropriate
prescribing in this setting is paramount.2–4 Spain is one of the EU
slight increase having been observed over recent years.5,6
According to the Agencia Espan ~ ola del Medicamento y Productos
Sanitarios, the overall antibiotic consumption in 2016 was slightly
over 33 defined daily doses per 1000 inhabitants per day, of which
one-third corresponded to non-reimbursed sales.6 This finding is

C The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved.
V
For permissions, please email: journals.permissions@oup.com.
1 of 8
Downloaded from https://academic.oup.com/jac/advance-article-abstract/doi/10.1093/jac/dky137/4989336
by guest
on 27 April 2018
 Llor et al.
also supported by the results of the latest Eurobarometer on anti-
biotic use carried out in 2016, in which 47% of the Spanish
respondents admitted having taken an antibiotic in the previous
year.7
Several strategies have been developed to reduce unnecessary
antibiotic prescribing in RTIs in the community with little success.8,9
A systematic Cochrane review, which included 39 studies of out-
patient antibiotic prescriptions for several infectious conditions,
found that printed educational materials for general practitioners
(GPs), audits and feedback alone resulted in little or no change in
prescription practices.8 A more recent review, including 58 studies,
found similar results, but in general, interventions based on mul-
tiple initiatives have been found to be more effective than those
focused on only one initiative.9 However, most studies included in
these reviews assessed the short-term effect of interventions, and
few studies have evaluated the long-term effect of a single inter-
vention, and these have also shown controversial results.10–14
The HAPPY AUDIT project was a study financed by the European
Commission aimed at strengthening the surveillance of RTIs
through the development of intervention programmes targeting
GPs and changing people’s habits towards prudent use of anti-
microbial agents.15 The results of this study demonstrated a 50%
reduction of antibiotic prescription 1–3 months after the interven-
tion.16 This follow-up study aimed to examine if there continued to
be a reduction in antibiotic prescribing in the intervention group of
the original study 6 years later.
Methods
Ethics approval
Ethics approval was granted by the Institut d’Investigació en Atenció
Primària Jordi Gol i Gurina, Barcelona, reference number 14/106.
Study design
GPs who had participated in a before-and-after audit-based study in 2008
and 2009 in eight different autonomous regions in Spain were invited to
participate in a new registration study in 2015. Patient registration took
place during 3 week periods in the winter months—January to February—
in 2008 (first registration), 2009 (second registration) and from January to
March in 2015 (third registration) covering a total of 15 working days in
each of these registrations.
Shortly after the first registration, the GPs were invited to follow-up
meetings where they received individual prescriber feedback and identified
potential quality problems from the first registration. These results were
discussed at 2 h follow-up meetings in each of the different areas and were
run by the local co-ordinators of the study. One to three months before the
second registration (November and December 2008) the participating GPs
received the following: (i) a training course on the appropriate use of antibi-
otics for RTIs; (ii) clinical guidelines with recommendations for diagnosis
and treatment of RTIs; (iii) brochures and handouts to patients about pru-
dent use of antibiotics; (iv) posters for waiting rooms, focusing on the appro-
priate use of antibiotics; (v) access to point-of-care tests (POCTs)—rapid
antigen detection tests (RADTs) and C-reactive protein rapid testing (CRP);
and (vi) training in the use and interpretation of POCTs. This intervention
was done during two different 2 h meetings with all the GPs in each of these
areas, led by the local co-ordinators. GPs were then advised to use a POCT in
cases of doubt and not as a stand-alone test: RADT only in patients with
suspected streptococcal pharyngitis (two or more Centor criteria) and
CRP in patients with lower RTI with uncertain aetiology. They were also
2 of 8
Downloaded from https://academic.oup.com/jac/advance-article-abstract/doi/10.1093/jac/dky137/4989336
by guest
on 27 April 2018
instructed to withhold antibiotic therapy for CRP values ,20 mg/L and to
prescribe an antibiotic for values .100 mg/L.
As the availability of rapid tests is scarce in Spain, we again provided
these POCTs to the consultation offices of the GPs allocated to the interven-
tion group in 2015. The tests provided were the Clearview Strep A Exact II
Dipstick, which takes !5 min to obtain a result, and the Alere Afinion CRP,
for which a result is available in ,4 min. The co-ordinators in each area
explained how to perform the procedures with these devices, but they did
not give any further explanation on when to perform and how to interpret
the results of these rapid tests in the third registration. Neither did they give
any other information about guidelines and appropriate utilization of antibi-
otics to avoid any further intervention. A new group of GPs from the same
regions who had never participated in courses on the rational use of antibi-
otics with comparable age, gender, type of job and number of years work-
ing to the GPs assigned to intervention were also invited to participate by
the local co-ordinators and acted as controls.
Measurements and data
The GPs were requested to register adults with RTIs by means of a specific
template providing relevant information about the infectious disease. On
this sheet, the physician attending the patient noted different specific
parameters of medical care, including the age and gender of the patient,
the number of days with symptoms, presenting symptoms and signs, rapid
tests used along with their results, performance of chest radiography, main
diagnosis, antibiotic treatment or not, whether the patient requested an
antibiotic, and referral to another healthcare setting. Three new variables
were added to the template used in 2015: infection severity on a five-item
scale, pulse oximetry and significant comorbidities. Only first-time contacts
for the current disease were registered.
Statistical analyses
The data were analysed with the Stata v0.13 statistical program. Data ana-
lysed in a hierarchical multilevel logistic regression model were estimated
at two levels: contacts with RTIs and GPs. Antibiotic prescription was consid-
ered as the dependent variable (yes/no). The random variation at level 2
(physicians) was accounted for with random intercepts (systematic differ-
ences among physicians in the propensity to prescribe) and a random slope
for the dummy of the intervention group in 2015, which was assumed to
have a non-zero correlation with the random intercept. The model was also
adjusted for the covariables of age and gender, number of days with symp-
toms, symptoms and signs presented, diagnosis and patient demand for
antibiotics. The goodness of fit was assessed using the Wald test of the
model, with the deviance test to compare alternative models. Statistical
significance was considered with P , 0.05.
Results
A total of 123 GPs of the 210 GPs who had participated in the first
and second registrations in 2008 and 2009 agreed to participate in
the third registration; however, one did not send the templates
and another was rejected due to confirmed fraudulent data com-
pletion. No statistical differences were observed in terms of age,
gender, years working and previous antibiotic prescribing rate be-
tween those who completed the third registration and those who
discontinued. Valid data were obtained from 121 GPs (57.6% of all
the GPs who underwent the intervention in 2008). The control
group consisted of 117 new GPs who never had been exposed to
interventions on antibiotic prescribing. Figure 1 shows the general
scheme of the study.
GPs assigned to the intervention group were slightly older than
those allocated to the control group were (mean age of 50.6 years
 Long-term impact of an intervention on antibiotic use
Intervention group
235 GPs were invited to participate in
the first registration, in 2008
11 GPs did not wish
to participate
224 GPs participated in the first
registration, in 2008
14 GPs refused to
continue the study
210 GPs undertook the intervention in
2008
210 GPs participated in the second
registration, 2009
123 GPs participated in the third
registration, 2015
1 GP sent uncompleted registration sheets
1 GP excluded due to unethical completion
121 GPs completed all the registration
sheets in the third registration, 2015
Figure 1. General scheme of the study.
versus 48.9 years, P " 0.04) and there were slightly more men in
the former group (57.8% versus 42.7%, P " 0.21). No statistically
significant differences were observed in the number of years of
professional activity between the two GP groups (21.4 versus
18.9 years, respectively, P " 0.06). The 238 GPs included a total of
22247 patients with RTIs, with a mean age of 47.5 years (SD 20.3),
and 59.8% were female. The infection most frequently registered
was the common cold (9136 cases; 41.1%), followed by acute
pharyngitis (4347 cases; 19.5%) and acute bronchitis (3172 cases;
14.3%). As shown in Table 1, the utilization of POCTs was signifi-
cantly higher among physicians assigned to intervention in 2009
and 2015, while GPs in the control group used the pulse oximetry
significantly more frequently than their counterparts (25.5% ver-
sus 17.3%, respectively).
Antibiotics were given to 5325 patients (23.9%), but this per-
centage was significantly lower among the GPs assigned to inter-
vention just after the intervention took place and was highest
among the new GPs allocated to the control group (16.8% versus
30.6%), whereas the percentage of antibiotics prescribed by the
GPs 6 years after the intervention was 22.3%. The highest antibiot-
ic prescription was observed in pneumonia, acute otitis media and
acute exacerbations of chronic bronchitis and COPD (Table 2).
Data were analysed in a two-level logistic regression model
with patients (n " 22247) allocated to level 1 and GPs (n " 238) to
Downloaded from https://academic.oup.com/jac/advance-article-abstract/doi/10.1093/jac/dky137/4989336
by guest
on 27 April 2018
JAC
Control group
135 new GPs were invited to participate in
this registration, 2015
6 GPs did not wish to
participate
129 GPs accepted to participate
in the study, 2015
10 GPs were trainees
2 GPs sent uncompleted
registration sheets
117 GPs completed all the registration
sheets, 2015
level 2. The Wald test of the model was 912.1 (P , 0.001). Tonsillar
exudate (OR 9.17) and sputum purulence (OR 6.23) were signifi-
cantly associated with antibiotic prescribing. Conversely, cough
(OR 0.73) was associated with low antibiotic prescribing (Table 3).
On adjustment for covariables, compared with the antibiotic pre-
scription observed just after the intervention, GPs assigned to the
intervention group prescribed slightly more antibiotics in 2015, al-
beit without statistical significance (OR 1.08, 95% CI 0.89–1.31),
while GPs allocated to the control group prescribed significantly
more antibiotics (OR 2.74, 95% CI 2.09–3.59).
Discussion
To the best of our knowledge, this is the first study to report the lon-
gest effect of an intervention carried out in the community with
the use of a control group. A single multifaceted intervention
including feedback results of a previous registration, interactive
seminars on RTIs, guideline discussion and workshop on rapid tests
was associated with a slight increase in antibiotic prescribing
6 years later compared with antibiotic prescription observed just
after the intervention, although no statistically significant differen-
ces were observed. Notwithstanding, GPs who had never partici-
pated in an intervention on the appropriate use of antibiotics and
acted as the control group prescribed significantly more antibiotics.
3 of 8
 Llor et al.

Table 1. Baseline characteristics of the contacts with RTIs according to the group of physicians

Intervention group Intervention group Intervention group

in 2008, before the in 2009, just after the in 2015, 6 years after
intervention intervention the intervention Control group, 2015
Variable (121 GPs; 6089 RTIs) (121 GPs; 5199 RTIs) (121 GPs; 5381 RTIs) (117 GPs; 5578 RTIs)

Demographics and characteristics of the RTI

age (years), mean (SD) 46.5 (19.7) 46.0 (20.4) 48.9 (20.4) 48.6 (20.5)
female, n (%) 3601 (59.1) 3081 (59.3) 3228 (60.0) 3383 (60.6)
prior duration of symptoms (days), mean (SD) 4.3 (4.5) 4.5 (4.6) 4.7 (5.2) 4.7 (5.1)
severity of the infection (score), mean (SD) NC NC 1.4 (0.7) 1.4 (0.6)
significant comorbidities, n (%) NC NC 739 (13.7) 970 (17.4)
Symptoms and signs, n (%)
fever 1914 (31.4) 1610 (31.0) 1462 (27.2) 1554 (27.9)
cough 4779 (78.5) 3958 (76.1) 4290 (79.7) 4429 (79.4)
purulent ear discharge 91 (1.5) 78 (1.5) 93 (1.7) 114 (2.0)
odynophagia 2623 (43.1) 2376 (45.7) 2317 (43.1) 2326 (41.7)
tonsillar exudate 265 (4.4) 314 (6.0) 246 (4.6) 185 (3.3)
tender cervical nodes 321 (5.3) 420 (8.1) 328 (6.1) 188 (3.4)
dyspnoea 508 (8.3) 455 (8.8) 542 (10.1) 655 (11.7)
increase in sputum 1135 (18.6) 1011 (19.4) 1076 (20.0) 1037 (18.6)
purulence of sputum 618 (10.1) 431 (8.3) 519 (9.6) 565 (10.1)
Tests performed/ordered, n (%)
RADTs 52 (0.9) 865 (16.6) 663 (12.3) 90 (1.6)
CRP 2 (0.0) 597 (11.5) 565 (10.5) 15 (0.3)
pulse oximetry NC NC 930 (17.3) 1421 (25.5)
chest X-ray 140 (2.3) 147 (2.8) 118 (2.2) 142 (2.5)
Diagnosis, n (%)
common cold 2417 (39.7) 2031 (39.1) 2356 (43.8) 2332 (41.8)
acute otitis media 124 (2.0) 91 (1.8) 127 (2.4) 167 (3.0)
acute sinusitis 160 (2.6) 172 (3.3) 183 (3.4) 185 (3.3)
acute pharyngitis 1231 (20.2) 1210 (23.3) 953 (17.7) 953 (17.1)
acute bronchitis 721 (11.8) 593 (11.4) 859 (16.0) 999 (17.9)
pneumonia 75 (1.2) 71 (1.4) 54 (1.0) 58 (1.0)
acute exacerbations of CB/COPD 271 (4.5) 195 (3.8) 204 (3.8) 233 (4.2)
influenza 658 (10.8) 420 (8.1) 511 (9.5) 517 (9.3)
other RTIs 321 (5.3) 332 (6.4) 134 (2.5) 134 (2.4)
Management of the infection
antibiotic prescribed 1545 (25.4) 872 (16.8) 1201 (22.3) 1707 (30.6)
patient demands antibiotics 110 (1.8) 78 (1.5) 115 (2.1) 93 (1.7)
referral of the patient to hospital 33 (0.5) 30 (0.3) 38 (0.7) 57 (1.0)

CB, chronic bronchitis; NC, not collected.

Strengths and limitations
Our study has several limitations. As in other studies, in the present
study, GPs participated voluntarily and probably their prescribing
habits were not representative of all GPs and this may have limited
the extrapolation of the results obtained. It is possible that the par-
ticipating GPs were more prone to interventions aimed at behav-
iour change than their colleagues who did not participate.17 One of
the striking results of this paper is the considerable difference in
antibiotic prescribing observed in the GPs assigned to intervention
before this took place and the control group recruited in 2015,
which was 5.2% higher among the latter physicians. Although the
GPs allocated to the intervention group could have been more
motivated than their control counterparts, this phenomenon could
also reflect the slight albeit steady increase in antibiotic consump-
tion observed over recent years in Spain.5,6
It is a limitation that we did not register clinical outcomes or
complications. However, some important variables not included in
the first two registrations that might have influenced whether an
antibiotic was prescribed or not such as significant comorbidities,
severity of the infection according to GP judgement or the use of
pulse oximetry were added to the 2015 template. For all patients
included, GPs registered the first reason for encounter when a pa-
tient with suspected RTI was seen in their practice and the pattern
of RTI infections was most probably equal for the different groups
of patients. Other potential confounders not taken into account in
our analysis, such as physician motivation and local campaigns to

4 of 8
Downloaded from https://academic.oup.com/jac/advance-article-abstract/doi/10.1093/jac/dky137/4989336
by guest
on 27 April 2018
 Long-term impact of an intervention on antibiotic use JAC
Table 2. Antibiotic prescription for the different RTIs according to the group of physicians

Intervention group Intervention group Intervention group

in 2008, before the in 2009, just after the in 2015, 6 years after the Control group,
Variable intervention (N " 6089) intervention (N " 5199) intervention (N " 5381) 2015 (N " 5578)

Common cold, n (%) 46 (1.9) 22 (1.1) 51 (2.2) 58 (2.5)

Acute otitis media, n (%) 107 (86.3) 74 (81.3) 90 (70.9) 133 (79.6)
Acute rhinosinusitis, n (%) 135 (84.4) 99 (57.6) 113 (61.7) 147 (79.5)
Acute pharyngitis, n (%) 476 (38.7) 258 (21.3) 344 (36.1) 457 (48.0)
Acute bronchitis, n (%) 396 (54.9) 174 (29.3) 358 (41.7) 628 (62.9)
Pneumonia, n (%) 68 (90.7) 62 (87.3) 46 (85.2) 44 (75.9)
Acute exacerbations of CB/COPD, n (%) 212 (78.2) 133 (68.2) 160 (78.4) 181 (77.7)
Influenza, n (%) 20 (3.0) 1 (0.2) 18 (33.5) 24 (4.6)
Other RTIs, n (%) 60 (18.7) 42 (12.7) 21 (15.7) 35 (26.1)
Total, n (%) 1545 (25.4) 872 (16.8) 1201 (22.3) 1707 (30.6)

CB, chronic bronchitis.

Table 3. Effect of an intervention 6 years later; multivariable predictors regarding antibiotic use were launched in 2007 and 2008, but the
of antibiotic prescription in RTIs results of these campaigns have never been studied. However,
public campaigns on antibiotic awareness in other European coun-
Variable OR SE 95% CI P tries have not been associated with significant reductions in anti-
biotic prescribing in primary care.19 GPs allocated to the control
Age 1.01 0.00 1.01–1.01 ,0.001
group had never participated in courses on the rational use of anti-
Female 0.92 0.46 0.83–1.01 0.08
Prior duration of symptoms 1.03 0.01 1.02–1.04 ,0.001
biotics, but we cannot rule out the possibility that some may have
Fever 3.83 0.23 3.41–4.30 ,0.001
been particularly concerned about the problem of antimicrobial re-
Cough 0.73 0.05 0.65–0.84 ,0.001
sistance. Despite the fact that POCTs are seldom used in primary
Purulent ear discharge 1.76 0.36 1.18–2.62 0.01
care in Spain, we found that 1.6% of all the contacts with RTIs by
Odynophagia 1.08 0.07 0.94–1.23 0.28
GPs assigned to the control group used rapid tests, particularly
Tonsillar exudate 9.17 0.93 7.52–11.18 ,0.001
RADTs. Another potential limitation is the Hawthorne effect, which
Tender cervical nodes 2.83 0.28 2.34–3.42 ,0.001
may have introduced bias because GPs in both the intervention
Dyspnoea 1.52 0.12 1.30–1.78 ,0.001
and the control group, but more importantly in the latter as they
Increase in sputum 2.02 0.14 1.77–2.31 ,0.001
were not familiar with this audit-based registration, may have
Purulence of sputum 6.23 0.49 5.35–7.27 ,0.001
altered their prescription habits. To reduce this risk, GPs in the con-
Common cold 0.09 0.02 0.05–0.15 ,0.001
trol group were specifically instructed to follow their usual care.
Acute otitis media 34.36 9.70 19.76–59.76 ,0.001
Theoretically, the decision to treat should be taken after a diag-
Acute sinusitis 17.77 4.54 10.77–29.32 ,0.001 nosis has been established. In general practice, however, the diag-
Acute pharyngitis 1.72 0.41 1.07–2.75 0.03 nostic procedures and the decision to treat are intricately
Acute bronchitis 3.69 0.88 2.31–5.89 ,0.001 intertwined. The GP may decide whether or not to prescribe an
Pneumonia 19.82 6.23 10.71–36.69 ,0.001 antibiotic at the same time, or even before. After making the deci-
Acute exacerbations of CB/COPD 8.49 2.18 5.12–14.05 ,0.001 sion to prescribe, the GP may thus then adjust the diagnosis to fit
Influenza 0.06 0.02 0.03–0.11 ,0.001 the decision about treatment. This may lead to a diagnostic mis-
Other RTIs 1.06 0.27 0.64–1.75 0.82 classification bias. However, this potential bias might have affected
Patient demands antibiotics 4.36 0.71 3.17–6.00 ,0.001 the validity of the diagnosis both before and after the intervention
Intervention group in 2015 1.08 0.11 0.89–1.31 0.46 and in both groups of GPs (intervention and control groups) and it
Control group in 2015 2.74 0.38 2.09–3.59 ,0.001 is very unlikely to have influenced the effect of the intervention.15
The data corresponding to all the registrations completed in
CB, chronic bronchitis; SE, standard error.
2015 were entered by only one of the authors of this paper (A. M.),
who is both a GP and a researcher. She raised concern of alleged
fraudulent completion of some data by three GPs allocated to the
intervention group on noticing a digit preference in some numeric-
promote appropriate use of antibiotics, were probably evenly dis- al variables.20 Owing to budget constraints we were unable to
tributed between the groups. In all the groups of patients GPs used double-check the information contained in all the templates with
the same type of registration sheet and a potential influence of GP the electronic records, but after checking the information provided
behaviour caused by the registration was most likely similar in by these three suspected cases one of these GPs was excluded
the two groups of GPs.18 Two different nationwide campaigns from the final analysis due to confirmed fraud and all the data

5 of 8
Downloaded from https://academic.oup.com/jac/advance-article-abstract/doi/10.1093/jac/dky137/4989336
by guest
on 27 April 2018
 Llor et al.
entered by this researcher in the three registration periods were
rejected. Conversely, the suspicion of scientific misconduct
observed in the other two cases was finally dismissed, as the infor-
mation provided on the templates was veracious. Therefore, on
concluding the investigation of possible data fraud and imple-
menting the necessary actions to ensure data validity, we can con-
fidently state that all the data entered in the study are valid and
reliable.
This was not a clinical trial, because GPs were not randomly
assigned to the different groups. This might explain why the new
GPs assigned to the control group were slightly younger than their
counterparts, with more female doctors. However, the differences
observed were not clinically relevant and we do not consider that
this difference could have any effect on the results obtained.
A total of 10 trainees who were included in the control group in
one of the areas were finally excluded to make both groups of GPs
more comparable. The greatest strength of this study was the
large number of physicians included. In addition, .50% of the GPs
who had participated in the first and second registrations and had
participated in the intervention accepted to participate in this new
registration 6 years later.
Comparison with other studies
Several randomized controlled trials conducted in primary care
demonstrated a significant effect of various educational interven-
tions on antibiotic prescriptions for RTIs over the subsequent
year,21,22 but few studies aimed at assessing the long-term effect
of interventions have been published. Indeed, studies assessing
effects on antibiotic prescription in primary care after 3 years
wrought conflicting results.10–14 In general, a reduction of the anti-
biotic prescribing rate is still observed in the long term as four of
these studies found a decrease in the proportion of patients with
RTIs who were prescribed antibiotics, but this mainly depends on
the number and type of strategies used in the intervention.10–13 In
a 3.5 year follow-up of a randomized controlled trial carried out in
the Netherlands, an intervention that consisted of the provision of
CRP in the consultations was associated with a 5% reduction of
antibiotic prescribing for patients with acute cough, but the inter-
vention was continuously delivered along the study period.12 In a
French randomized clinical trial, a 2 day interactive educational
intervention was associated with a significant reduction of antimi-
crobials prescribed 30 months after this intervention,23 and contin-
ued showing a lower antibiotic prescribing rate compared with the
control group after 4.5 years of follow-up.13 In contrast, a 5 year
Finnish study showed that guideline dissemination through local
interactive discussion groups failed to decrease significantly the
antibiotic prescriptions, but these results were not compared with
a control group.14
Changing clinicians’ prescribing behaviour is a complex matter.
Several strategies such as the provision of guidelines, poster dis-
play in the waiting rooms and even leaflets have limited effect.
When active strategies are added to these passive educational
sessions, such as the feedback and discussion of previous results,
training in enhancing communication skills in the consultation and
the provision of POCTs, the effect is much greater. The different ac-
tive strategies used in the present study, which required the at-
tendance to 2 h courses and workshops on three different days,
6 of 8
Downloaded from https://academic.oup.com/jac/advance-article-abstract/doi/10.1093/jac/dky137/4989336
by guest
on 27 April 2018
might explain the continued reduction in antibiotic prescribing
observed 6 years after the intervention.
Conclusions
To the best of our knowledge, no study has evaluated the effect of
a single complex intervention more than 5 years later. However, in
this observational prospective study, a single multifaceted inter-
vention carried out 6 years earlier consisting of the discussion of
the first registration results, a training course on appropriate use of
antibiotics and clinical guidelines for RTIs, patients’ brochures,
posters for waiting rooms, workshop on rapid tests and provision
of these POCTs in the consultation, continued to reduce antibiotic
prescription for RTIs.
Acknowledgements
We wish to acknowledge the GPs who voluntarily participated in the
study.
Other members of the HAPPY AUDIT 3 Study Team
Andalucı́a: Idoioa Arrillaga Ocampo, Francisco Atienza Martı́n, Eduardo
Ávila Amat, Marta Bernal Raya, Pablo Cabezas Saura, Pilar Carretero
Castan ~ o, Carmen Dastı́s Bendala, Irene V. de Tena Roger, Pedro
Fernández-Nieto Fernández, Amalia Filella Sierpes, M. Isabel González
Rodrı́guez, Antonio Hernández Alonso, Guillermo Largaespada
Pallaviccini, Beatriz López Fernández, Inés M. López Huerta, Beatriz
Luque López, Yanina Maldonado, M. Luisa Manzanares Torné, Leonor
Marı́n Pérez, M. Mercedes Martı́nez Granero, Rocı́o Martı́nez Pérez, M.
Antonia Máximo Torres, Ainhoa Mestraitua Vázquez, Mercedes Moreno
Labrado, M. Luisa Moya Rodrı́guez, José Oropesa de Cáceres, Irene
Padial Reyes, Antonio Paniagua Galisteo, Vanessa Parra Garcı́a, M.
Encarnación Peláez Gálvez, Juan J. Quero Perabá, M. Jesús Remesal
Barrachina, Miguel Sagristá González, Consuelo Sampedro Abascal,
Jesús C. Sánchez Vázquez, Auxiliadora Sanchı́s Osuna, J. Cristóbal Sendı́n
González, Rosario Serrano Ortega, Miguel Silva Cueto, Guillermo
Velázquez de Cisneros. Baleares: Ester Adelantado Pozuelo, Daniel
Bestard Marı́n, Aberlado Corrales Nadal, M. Teresa Corredor Ibán ~ ez,
Esther Domı́nguez Padilla, Alberto Eek Comas, Margarita Fornés Homar,
Maria E. Garau Miquel, Prado Garcı́a Servera, Salvador J. Gestoso Gaya,
Marı́a J. González-Bals González, Mercedes Gutiérrez Garcı́a, Antoni J.
Jover Palmer, Magdalena Llinàs Suau, Marian Llorente Sanmartı́n,
Blanca Martı́nez Andión, Adriano Mayrata Vicens, David Medina i
Bombardó, Gabriel Moragues Sbert, Joana M. Oliver Gornals, Etel Oliver
Roset, J. Francisco Palmer Simó, Juana Pérez Galmés, Susana Pons Vives,
Lourdes Quintana Torres, José Alfonso Ramón Bauzá, F. Javier Rezola
Gambón, Jaime Ripoll Sánchez, Miguel Román Rodrı́guez, Natalia
Serrano Gomila, Magdalena Servera Trias, Isabel M. Socias Buades, M.
Mar Sureda Barbosa, Joana Tortella Morro, Lucia Ugarriza Hierro, Silvia
Vega Garcı́a. Canarias: Pilar Aguilar Utrilla, M. Carmen Artiles Ruano, M.
Isabel Cardenes Romero, Jonás de la Cruz Cabrera, Margarita Garcı́a
Garcı́a, M. Carmen Guerra, Ricardo Koch, M. Huertas Llamas Martı́nez, M.
Elena Martı́n Santana, Alicia Monzón Guerra, Marta Moreno Ramos,
Caridad Sánchez Artiles, Lucı́a Tejera Pulido, José M. Toscano, Carlos
Prieto, Aurelia Perdomo. Galicia: Francisco Castrillo Villar, Pilar Cobas
~ o,
Pacı́n, M. Jesús Colas Martı́nez, Coral Dı́ez Pérez, Peregrina Eiroa Patin
Ana Fernández Álvarez, Luis Fucin ~ os Cebreiro, J. Antonio Gómez
Villalobos, M. Esther González Garcı́a-Mayor, Ricardo M. Héctor Sanz,
Susana Hernáiz Valero, Esther López Carbajales, J. Luis López Vilar, Elena
Lorenzo Llauger, Elisa Mosquera Gayoso, Concepción Nogueiras Santas,
M. Pilar Pintos Martı́nez, M. Teresa Rı́os Rey, Lucı́a Rodrı́guez Nieto,
 Long-term impact of an intervention on antibiotic use
Santiago Santidrián Arias, Jesús Sueiro Justel. Madrid: Raquel Ban ~ os
Morras, José M. Casanova Colominas, Santiago Castellanos Redondo, M.
Teresa Cobos Hinojal, José Corral Brihuega, M. Canto de Hoyos Alonso,
Belén de la Fuente Martı́n, Silvia de las Heras Loa, Angélica Fajardo
Alcántara, M. Jesús Ferrer Signes, Antonia Garcı́a del Val, M. Mar Garcı́a
Rabanal, Gema Garcı́a Sacristán, Álvaro Garcı́a Vega, M. Elena Gerez
Sánchez-Escribano, Paloma González Escobar, Rosa M. González San
Segundo, M. Elena Hermida López, Paloma Hernández Almarza, Ana
Ibarra Sánchez, Ángeles Lagos Aguilar, Ascensión Lázaro Damas, M.
Eugenia López Delgado, Guillermina López Fernández, Rosario López
Morell, M. José Lucena Martı́n, Jaime Marı́n Can ~ ada, Rosa Martı́n de
Cabo, Blanca Matilla Pardo, Teresa Mazarro Enrique, Cristina Moral
Moraleda, Adela Moreno Mateos, Joaquı́n Morera Montes, Renata Mun ~ oz
Bieber, Javier Mun ~ oz Gutiérrez, Alicia Mun~ oz Nún~ ez, M. Arantzazu
Murciano Antón, Patricia Navarro Gil, Lourdes Pulido Izquierdo, Carmen
Rodrı́guez Fernández, Ana Rodrı́guez Nebreda, M. Elena Rodrı́guez
Quiroga, M. Asunción Rosado López, Ana Ruı́z Molina, Rocı́o Susana Ruiz
Vargas, Irmina Saldan ~ a Alonso, Mar Sánchez Martı́n, Laura Santos
Larregola, Juana M. Torres Martı́nez, Gema Val Morales, Amalia
Velázquez Garcı́a, Lourdes Visedo Campillo, Antonio Yagüe Hernando.
Rioja: Ascensión Aicua Sadia, M. Carmen Aurensanz Sánchez, Carmen
Babace Isturiz, Diego Beni Ruiz, Nieves Calavia Rueda, M. Mar Dı́ez
Casado, M. Ángeles Espeso Fernández Valderrama, Francisco Garcı́a
Marı́n, Tomás Garcı́a Martı́nez, Ana M. Gómez Ruiz, Elvira Moreno
Hernández, Sabino Ornia Rodrı́guez, M. José Palomo Larrieta, Olga
Revilla González, Rosa Ruiz De Austri Fuente, J. Luis Torres Baile, Antonio
Trueba Castillo, Ana M. Vázquez Medrano. Asturias: Carmen Alonso
Alonso, Fernando Álvarez Guisasola, Pablo Belderraı́n Belderraı́n, J.
Carlos Bango Melcón, Teresa Dı́az Mujica, Deopatria Azucena Esteban
Fresno, Rocı́o Fernández Collar, Leonor Fuentes Martin, Alfonso Garcı́a
Viejo, Ramón Macı́a Pérez, Carmen Martı́n López, Ernesto Martı́nez
Estrada, Herminia Menéndez Fernández, Covadonga Monte Llavona, M.
Amor Paredes González, Cristina Pérez Canal, Javier Pérez Fernández, M.
Mar Pizarro López, Rosa Rodrı́guez, J. Ignacio Rodrı́guez-Arias Palomo,
Alicia Rodrı́guez Fernández, Felipe J. Rodrı́guez Gordaliza, Celia T.
Tamargo Garcı́a, Salvador Tranche Iparraguirre, M. Raimunda Vázquez
Aira, Rubén Villa Estébanez. Valencia: Juan Alamar Provecho, M. Teresa
Amorós Barber, Miguel Antón Sánchez, Manuel Batalla Sales, Lourdes
Bermejo Mora, Núria Bosch Girona, M. Pilar Carceller Sales, Rosa Carrión
Villanueva, Teresa Chirivella Raga, Manuela Domingo Ordun ~ a,
Concepción Garcı́a Domingo, Carmen Hernández Espinosa, Blanca Folch
Marı́n, Carmen Gandı́a Moya, Amparo Haya Guaita, Eva M. Izquierdo
Hidalgo, Engracia López González, Nuria Mallén Navarro, Pilar Marı́n
Lluch, Amparo Meneu Ferrer, Helena Molina Raga, Adela Moreno
Fernández, Ángeles Pérez Campos, Belén Persiva Saura, Gloria
Rabanaque Mallén, Remei Raga Marı́, Jazmı́n Ripoll Perelló, Pilar Roca
Navarro, M. Luisa Tregón Moliner.
Funding
This study was funded by TRACE (Translational Research on
Antimicrobial resistance and Community-acquired infections in Europe).
The funding source had no role in study design, data collection, data
analysis, data interpretation or writing the report.
Transparency declarations
C. L. and A. M. report receiving research grants from the European
Commission (Sixth & Seventh Programme Frameworks and Horizon
2020), the Catalan Society of Family Medicine, Instituto de Salud Carlos
III and Alere. All other authors: none to declare.
Downloaded from https://academic.oup.com/jac/advance-article-abstract/doi/10.1093/jac/dky137/4989336
by guest
on 27 April 2018
JAC
Author contributions
The initial idea for this study was proposed by C. L. and L. B.; all authors
contributed to the development of the protocol and to the management
of the study. A. M. entered all the data. C. L., L. B., A. M. and J. M. C. led
the funding application and provided overall co-ordination of the
TRANCE consortium. B. G. L.-V. analysed the data; all authors contributed
to the interpretation of the data and the write-up. C. L., L. B., B. G. L.-V., A. M.
and J. M. C. had full access to all of the data in the study and take responsi-
bility for the integrity of the data and the accuracy of the data analysis.
References
1 Goossens H, Ferech M, Vander Stichele R et al.; ESAC Project Group.
Outpatient antibiotic use in Europe and association with resistance: a cross-
national database study. Lancet 2005; 365: 579–87.
2 Lemiengre MB, van Driel ML, Merenstein D et al. Antibiotics for clinically
diagnosed acute rhinosinusitis in adults. Cochrane Database Syst Rev 2012;
issue 10: CD006089.
3 Smith SM, Fahey T, Smucny J et al. Antibiotics for acute bronchitis.
Cochrane Database Syst Rev 2017; issue 6: CD000245.
4 Kenealy T, Arroll B. Antibiotics for the common cold and acute purulent
rhinitis. Cochrane Database Syst Rev 2013; issue 6: CD000247.
5 Summary of the Latest Data on Antibiotic Consumption in the European
Union, ESAC-Net Surveillance Data, November 2016. https://ecdc.europa.eu/
sites/portal/files/documents/antibiotics-ESAC-Net%20Summary%202016_
0.pdf.
6 Plan Nacional Resistencia Antibióticos. Consumos Antibióticos Atención
Primaria (Receta oficial y Privada en DHD). Agencia Espan ~ ola de
Medicamentos y Productos Sanitarios (AEMPS), 2017. http://www.resisten
ciaantibioticos.es/en/node/332.
7 European Commission. Special Eurobarometer 445, Antimicrobial
Resistance; April 2016. https://ec.europa.eu/health/amr/sites/amr/files/
eb445_amr_generalreport_en.pdf.
8 Arnold SR, Straus SE. Interventions to improve antibiotic prescribing practi-
ces in ambulatory care. Cochrane Database Syst Rev 2005; issue 4: CD003539.
9 van der Velden AW, Pijpers EJ, Kuyvenhoven MM et al. Effectiveness of
physician-targeted interventions to improve antibiotic use for respiratory
tract infections. Br J Gen Pract 2012; 62: e801–7.
10 Mölstad S, Ekedahl A, Hovelius B et al. Antibiotics prescription in primary
care: a 5-year follow-up of an educational programme. Fam Pract 1994; 11:
282–6.
11 Zwar N, Henderson J, Britt H et al. Influencing antibiotic prescribing by
prescriber feedback and management guidelines: a 5-year follow-up. Fam
Pract 2002; 19: 12–7.
12 Cals JW, de Bock L, Beckers PJ et al. Enhanced communication skills and
C-reactive protein point-of-care testing for respiratory tract infection: 3.5-
year follow-up of a cluster randomized trial. Ann Fam Med 2013; 11: 157–64.
13 Ferrat E, Le Breton J, Guéry E et al. Effects 4.5 years after an interactive GP
educational seminar on antibiotic therapy for respiratory tract infections: a
randomized controlled trial. Fam Pract 2016; 33: 192–9.
14 Rautakorpi UM, Huikko S, Honkanen P et al. The Antimicrobial Treatment
Strategies (MIKSTRA) program: a 5-year follow-up of infection-specific anti-
biotic use in primary health care and the effect of implementation of treat-
ment guidelines. Clin Infect Dis 2006; 42: 1221–30.
15 Bjerrum L, Munck A, Gahrn-Hansen B et al. Health Alliance for prudent
antibiotic prescribing in patients with respiratory tract infections (HAPPY
AUDIT) -impact of a non-randomised multifaceted intervention programme.
BMC Fam Pract 2011; 12: 52.
7 of 8
 Llor et al.
16 Llor C, Cots JM, Hernández S et al. Effectiveness of two types of
intervention on antibiotic prescribing in respiratory tract infections in
Primary Care in Spain. Happy Audit Study. Aten Primaria 2014; 46:
492–500.
17 Lervy B, Wareham K, Cheung WY. Practice characteristics associated with
audit activity: a medical audit advisory group survey. Br J Gen Pract 1994; 44:
311–4.
18 Munck AP, Gahrn-Hansen B, Søgaard P et al. Long-lasting improvement
in general practitioners’ prescribing of antibiotics by means of medical audit.
Scand J Prim Health Care 1999; 17: 185–90.
19 Huttner B, Goossens H, Verheij T et al. Characteristics and outcomes of
public campaigns aimed at improving the use of antibiotics in outpatients in
high-income countries. Lancet Infect Dis 2010; 10: 17–31.
8 of 8
Downloaded from https://academic.oup.com/jac/advance-article-abstract/doi/10.1093/jac/dky137/4989336
by guest
on 27 April 2018
20 Al-Marzouki S, Evans S, Marshall T et al. Are these data real? Statistical
methods for the detection of data fabrication in clinical trials. BMJ 2005; 331:
267–70.
21 Gjelstad S, Høye S, Straand J et al. Improving antibiotic prescribing in
acute respiratory tract infections: cluster randomised trial from Norwegian
general practice (prescription peer academic detailing (Rx-PAD) study). BMJ
2013; 347: f4403.
22 Butler CC, Simpson SA, Dunstan F et al. Effectiveness of multifaceted edu-
cational programme to reduce antibiotic dispensing in primary care: practice
based randomised controlled trial. BMJ 2012; 344: d8173.
23 Le Corvoisier P, Renard V, Roudot-Thoraval F et al. Long-term effects of
an educational seminar on antibiotic prescribing by GPs: a randomised con-
trolled trial. Br J Gen Pract 2013; 63: e455–64.