S

A

ara

ctsh

ce©

K

R

bRdd

0

Disponible en ligne sur

ScienceDirectwww.sciencedirect.com

Médecine et maladies infectieuses 44 (2014) 308–314

Original article

Comparison of hospital databases on antibiotic consumption in France, for asingle management tool�

Comparaison des bases de données hospitalières de consommation d’antibiotiques en France, pourun outil de pilotage unique

. Henard a,∗,b, S. Boussat c, B. Demoré b,d, S. Clément e, T. Lecompte b, T. May a,b, C. Rabaud a,b,c

a Service des maladies infectieuses et tropicales, hôpital de Brabois, centre hospitalier universitaire de Nancy, allée du Morvan, 54500 Vandœuvre-les-Nancy,France

b Réseau Antibiolor, CHU Nancy, Vandœuvre-les-Nancy, Francec Centre de coordination de lutte contre les infections nosocomiales–Est, CHU Nancy, Vandœuvre-les-Nancy, France

d Université de Lorraine, SRSMC, UMR, 7565, Nancy, Francee Néanima, Aix-en-Provence, France

Received 2 April 2014; received in revised form 9 May 2014; accepted 2 June 2014Available online 9 July 2014

bstract

Context. – The surveillance of antibiotic use in hospitals and of data on resistance is an essential measure for antibiotic stewardship. Therere 3 national systems in France to collect data on antibiotic use: DREES, ICATB, and ATB RAISIN. We compared these databases and draftedecommendations for the creation of an optimized database of information on antibiotic use, available to all concerned personnel: healthcareuthorities, healthcare facilities, and healthcare professionals.

Methodology. – We processed and analyzed the 3 databases (2008 data), and surveyed users.Results. – The qualitative analysis demonstrated major discrepancies in terms of objectives, healthcare facilities, participation rate, units of

onsumption, conditions for collection, consolidation, and control of data, and delay before availability of results. The quantitative analysis revealedhat the consumption data for a given healthcare facility differed from one database to another, challenging the reliability of data collection. Wepecified user expectations: to compare consumption and resistance data, to carry out benchmarking, to obtain data on the prescribing habits inealthcare units, or to help understand results.

Conclusions. – The study results demonstrated the need for a reliable, single, and automated tool to manage data on antibiotic consumptionompared with resistance data on several levels (national, regional, healthcare facility, healthcare units), providing rapid local feedback andducational benchmarking.

2014 Elsevier Masson SAS. All rights reserved.

eywords: Benchmarking; Antimicrobial resistance; Antibiotic stewardship

ésumé

Contexte. – Surveiller la consommation hospitalière d’antibiotiques et les données de résistance est une mesure indispensable dans le cadre duon usage des antibiotiques. En France, au niveau national, trois recueils de la consommation des antibiotiques existent : DREES, ICATB et ATBaisin. Notre objectif était de comparer ces bases et de formuler des préconisations pour l’élaboration d’une base de recueil optimisée et uniquees consommations d’antibiotiques, utilisable par toutes les personnes concernées : autorités sanitaires, établissements de santé et professionnelse santé.

Méthodologie. – Traitement et analyse des données 2008 DREES, ICATB et ATB Raisin et enquête auprès des utilisateurs.

� This study was presented as a poster at the RICAI in Paris, in November 2013.∗ Corresponding author.

E-mail address: sandrine.henard@wanadoo.fr (S. Henard).

http://dx.doi.org/10.1016/j.medmal.2014.06.001399-077X/© 2014 Elsevier Masson SAS. All rights reserved.

mdud

dp©

M

1

aotdo

siu“cRDtaNR

mpaotHtaoHm

2

a

tsahd(

S. Henard et al. / Médecine et maladies infectieuses 44 (2014) 308–314 309

Observation. – Il existe d’importantes disparités pour les objectifs, le taux de participation des établissements, les unités de consommation, lesodalités de recueil, de consolidation et de contrôle des données et le délai de mise à disposition des résultats. Les données de consommation pour

es mêmes établissements sont différentes d’une base à l’autre, remettant en cause la fiabilité de ces recueils. Nous avons précisé les attentes destilisateurs : pouvoir croiser les données de consommation et de résistance, réaliser un benchmarking, disposer des données traduisant les habitudese prescription au niveau des unités de soin, ou accompagner les interprétations des résultats.

Préconisations. – Il est nécessaire de pouvoir disposer d’un outil fiable, unique et automatisé de pilotage des données de consommation’antibiotiques, croisées avec les données de résistances, à plusieurs niveaux (national, régional, établissement, services de soins) et pouvantermettre une rétro-information locale rapide et un benchmarking pédagogique.

2014 Elsevier Masson SAS. Tous droits réservés.

iotiqu

ot1uAdA

cditprtbtctdlt(Ne

ifurFttm

3

3

i

ots clés : Benchmarking ; Bon usage des antibiotiques ; Résistance aux antib

. Introduction

The misuse of antibiotics is responsible for a particularlylarming increase in the emergence of multi-resistant bacteria,n an individual and collective scale. The available therapeu-ic solutions are very limited, exposing patients to a therapeuticead-end, particularly since the number of new agents in devel-pment is very low [1].

Several “antibiotic plans” have been implemented in Franceince 2001 to promote the rational use of antibiotics. Thesenclude recommendations for intensified surveillance of these of and resistance to antibiotics [2–4]. There are severalnational” databases on antibiotic use in hospitals in France :ollection of data on medicinal products by the Division ofesearch, Studies, Evaluation, and Statistics (French acronymREES), use listed as part of the standard report on action taken

o prevent nosocomial infections (French acronym ICATB),nd surveillance of the Alert, Investigation, and Surveillanceetwork for Nosocomial Infections (French acronym ATBAISIN).

In 2009, confronted with multiple data collection and entryethods imposed on healthcare facilities (HCF), and with the

olitical desire to implement a new single system for collectingnd calculating data on antibiotic use, the Lorraine region antibi-tics network “Antibiolor” conducted a comparative study ofhese databases, after a call for tenders by the French Ministry ofealth (French acronym DGS). The objective of this study was

o compare these databases and to draft recommendations forn optimized, single database to collect information on antibi-tic use, usable by all persons concerned: healthcare authorities,CF, and healthcare professionals (prescribing physicians, phar-acists, microbiologists).

. Method

The study was conducted between May 2010 and April 2011,nd in two stages.

The first stage was dedicated to analyzing and comparinghe existing databases, from a qualitative and quantitative per-pective. This took place from July 2010 to February 2011. Thenalyses were conducted on 2008 data for which data collection

ad been finalized and consolidated for the 3 databases. Sourceata was provided by the 3 institutions running the databasesDREES, DGS, and RAISIN). The quantitative comparison

coa

es

f the 3 databases first required extensive data processing,o harmonize the collection unit in defined daily doses per000 patient-days (DDD/1000 PD) and the level of antibioticse (Anatomical Therapeutic Chemical classification level 3:TC 3). The analyses made it possible to compare shared useata for HCF, in the databases, in pairs, by linear regression.ccess and Excel (Microsoft®) softwares were used.During the second stage, semi-structured interviews were

onducted with the various contributors involved in the pro-uction, management, and/or analysis of these databases, todentify the reasons for participating in these various collec-ion systems, the workload required, any checks and correctionserformed, processing occurring at various levels (national,egional, local), and the expectations and lines of development athese various levels. These interviews were conducted by mem-ers of the Antibiolor team in charge of the study (a 2-personeam with a consultant–infectious diseases specialist, or pharma-ist, or microbiologist), with the 3 national institutions runninghe databases, 4 observatories of medicinal products, medicalevices, and therapeutic innovations (OMEDIT) at a regionalevel, and 11 HCF representative in terms of geographical loca-ion, size, and HCF category. In these HCF, various participantspharmacist, chairman of the Committee for the Prevention ofosocomial Infections (CLIN), microbiologist, infectious dis-

ase specialist) were met either individually or in groups.The data analysis, provided by the different surveys, made

t possible to draw conclusions and to issue recommendationsor the optimized collection and processing of data on antibioticse in hospitals in France, both at a national and local level. Theesults were presented in a detailed report, and briefing with therench Ministry of Health experts in April 2011, Group IV of

he National Surveillance Committee for plan aiming to preservehe effectiveness of antibiotics, in charge of surveillance and

onitoring of prescriptions.

. Results

.1. Qualitative analysis of the various databases

There were major discrepancies in the 3 databases concern-ng the objectives, HCF participation rate, level and units ofonsumption, conditions for collection, control, and validation

f data, and the delay before the results become available. Thesere summarized in Table 1:

310 S. Henard et al. / Médecine et maladies infectieuses 44 (2014) 308–314

Table 1Qualitative analysis of the 3 databases relating to antibiotic use in France (DREES, ICATB, and ATB RAISIN).Analyse qualitative des 3 bases de consommation francaises d’antibiotiques (DREES, ICATB et ATB Raisin).

Objectives Level of consumption andunit

Compulsory/voluntary

Number ofHCF/participationrate

Data checking Length of time beforeavailability of data

DREES Evaluation of theeconomic policy formedicinal products

Recording of unitspurchased, dispensed,sold by hospitalpharmacies, for each SDUcode

Voluntary 115571%

Contractor accordingto checklist

10 monthsno feedback sent toHCF

ICATB In the context of BSALIN ATC 3 class recorded asDDD/1000 PD J01 only

Compulsory 281299.90%

Consolidation andautomatic control bymeans of a web tool

No publication of data

ATB RAISIN Policy on the proper useof antibiotics and controlof resistance

Recording of SDU asATC 5, automaticconversion intoDDD/1000 PD

Voluntary 86131%

Checked by eachCClin

Inter-regional andnational reports withina period of 18 months

ATB RAISIN: Alert, Investigation and Surveillance Network for Nosocomial Infections network; ATC: Anatomical Therapeutic Chemical classification; BSALIN:Standard report on action taken to counteract nosocomial infections; DDD/1000 PD: defined daily doses per 1000 patient-days; DREES: Division of Research,S indicS

•

•

•

3

FsfC

tudies, Evaluation and Statistics (DREES); HCF: Healthcare facilities; ICATB:DU: Standard Dispensing Unit.

the DREES data collection was meant to evaluate the pol-icy on medicinal products (not only antibiotics). A key partof the system is based on monitoring the economic compo-nent. The raw data for the units purchased, dispensed, soldby hospital pharmacies for each SDU (standard dispensingunit) are extracted from the hospital pharmacy inventory man-agement software and transferred to a service provider. Thedata was aggregated by the HCF. 97% of HCF provided ana-lyzable data. Only 24 showed total J01 consumption above1000 DDD/1000 PD (2% of HCF). The data underwent exten-sive processing to be analyzed. The study results proved agood level of data reliability;

the ICATB index was implemented to evaluate the proper useof antibiotics in hospitals. The score resulting from the reportwas published every year. HCF participation was manda-tory. Data on antibiotic use had to be reported and recorded,but was not taken into account when calculating the score.

Consumption was reported by each HCF as DDD/1000 PD,according to the ATC 3 class. The data did not need to beprocessed beforehand. Antibiotic use was not reported by

iR

ig. 1. Total consumption as DDD/1000 PD for healthcare facilities found in all 3 dataystems. Jo1a: tetracyclines, jo1b: phenicols, jo1c: penicillin, jo1d: cephalosporin, mamily, jo1g: aminosides, jo1m: fluoroquinolones, jo1x: other antibiotics including glonsommations totales en DDJ/1000JH des établissements communs aux 3 bases (n

ator for the standard report on action taken to counteract nosocomial infections;

18% of HCF (509 HCF), and 5% of HCF reported totaluse > 1000 DDD/1000 PD (140 HCF). Patient days were notavailable or were equal to zero for 10% of HCF. The observedreliability was very low (empty fields, zero values or valuesnot divided by patient days, for 23% of HCF);

HCF was asked by the coordinating centers against nosoco-mial infections (CCLIN) to take part in the ATB RAISINsurveillance system on a voluntary basis. An Excel file wasused to enter use according to international non-proprietaryname (INN) corresponding to ATC 5. The conversion intoDDD/1000 PD was automatic. The data could be reported forthe whole HCF or by sector of activity. Several anomalieswere observed with discrepancies for patient days indicated,and the sum of DDD compared with total HCF.

.2. Quantitative analysis of the various data collected

The HCF were identified in the 3 databases via the FINESSdentification codes: 431 HCF were identified in the ATBAISIN and ICATB databases, 669 in the DREES and ICATB

bases (n = 424), according to ATC 3 class, based on the different data collectiononobactam, carbapenems, jo1e: sulfamides and trimethoprim, jo1f: macrolideycopeptides.= 424), par classe ATC3, en fonction des différents recueils.

S. Henard et al. / Médecine et maladies infectieuses 44 (2014) 308–314 311

Fig. 2. Linear regression graphs between total antibiotic use, for healthcare facilities found in all of the ATB RAISIN and ICATB (n = 431), ICATB and DREES(n = 669), and ATB RAISIN, and DREES databases (n = 424).G es deD

dasac

au

raphiques de régression linéaire entre les consommations totales d’antibiotiquREES (n = 669) et ATB Raisin et DREES (n = 424).

atabases, 424 in the ATB RAISIN and DREES databases,nd 424 in the 3 databases. Fig. 1 shows the total con-

umption as DDD/1000 PD for all HCF, by ATC 3 class,ccording to the various data collected. The DREES reportedonsumption was usually higher than the one reported by ICATB

aia

s établissements communs des bases ATB Raisin et ICATB (n = 431), ICATB et

nd ATB RAISIN, and the ATB RAISIN consumption rate wassually lower than the ICATB rate. The linear regression graphs

re shown in Fig. 2. The highest correlation was 0.91 for HCFdentified in both the ATB RAISIN and ICATB. It was equivalentnd lower (0.71) among HCF identified in both the ICATB and

3 ladies infectieuses 44 (2014) 308–314

Do

3

pcwi

ciwpctcHiga

3

dtH(

3

a

4

uwi

gpertoorrofhiI

Table 2User expectations for databases on the use of antibiotics.Attente des utilisateurs des bases de consommation d’antibiotiques.

At healthcarefacility level

If possible, avoid multiple data entry, a highlytime-consuming operation with no added valueMonitor antibiotic use and compare with resistance.Healthcare facilities would need to consolidate the dataaccording to FU, department, center, and healthcarefacility, as charts, key indicators, with visualillustrations, items which can be used for presentationsProvide comparative data (benchmarking), bycomparable type of healthcare facility and/orcomparable proprietary medicinal product.Obtain rapid feedback to implement relevant concreteactions. This involves having access to concise feedbackthat can be applied at a local level (presentation toclinical practitioners). Excessively detailed nationalreports should be avoidedObtain support for the analysis of results

At regional level Provide an efficient system for data collection, whichcan easily generate reliable dataSupport healthcare facilitiesIncorporate resistance data to allow initiating alertsImplementation of a tool to monitor data at department

cueo

nitcrrc2oti

aamccsoTDcca

12 S. Henard et al. / Médecine et ma

REES on one hand, and in both the DREES and ATB RAISINn the other hand.

.3. Participation in the various data collection systems

The results of the surveys conducted with HCF proved thatarticipation in a national data collection system was due to itsompulsory nature or to strong incentives. Consumption dataas not very reliable since it was not checked or used by HCF

n their management.The participation was also motivated by HCF’s interest con-

erning ATB RAISIN data collection. Collection via thesenstruments offered a set of indicators, converted into DDD,hich software commonly used in hospital pharmacies did notrovide. The most important healthcare professional for dataollection was usually the pharmacist, and occasionally an infec-ious diseases specialist. The data was consistent for severalollections when the same person carried out data collection.owever, the workload required (time spent by the pharmacist,

nfectious diseases specialist aggregating the data) was muchreater than for other types of data collection: from 0.5 days (ifutomatic requests were created and used) to 10 days per year.

.4. Data analysis

No DREES data was used for data analysis at HCF level (rawata that could not be analyzed directly). The ICATB consump-ion was considered too heavily aggregated to be analyzed. SomeCF used only data from the ATB RAISIN for their meetings

anti-infective drug committee, or CLIN).

.5. User expectations

Specific expectations were mentioned by the participants, andre summarized in Table 2.

. Discussion

This study’s objective was to compare 3 national databasessed to collect hospital antibiotic use data in France. The resultsere that, in addition to the numerous types of data collection

nvolved, the 3 studied databases were heterogeneous.The consumption rate collected by the DREES was usually

reater than the one observed by the ICATB and ATB RAISIN,robably because local hospitals, with low antibiotic use, werexcluded from the DREES data collection. The consumptionates collected by the ATB RAISIN were usually lower thanhe ones collected by the ICATB and DREES, since some HCFnly provided data for 1 or more units, the cumulative totalf which was lower than for the HCF as a whole. The cor-elation coefficients between the 3 databases were low. Theseesults show that the reliability of the data collected was notptimal, due to the collection conditions and the controls per-

ormed. The insufficient reliability of data in the ICATB databasead already been demonstrated by the authors of a study focus-ng on the relationship between antibiotic consumption and theCATB indicator between 2006 and 2008: 21.2% of the values

dw

r

and center level, as part of the implementation of theelectronic patient record

orresponded to zero or missing values, and 23.4% of the val-es were considered aberrant values [5]. To our knowledge, noquivalent data validation studies had been performed for thether 2 databases.

The results of the analysis of the collected data, both atational and local level, varied considerably. Not all of the datan the ICATB database was analyzable; the results obtained fromhe DREES database were not returned to HCF. Only the dataollected in the ATB RAISIN were returned to HCF as an inter-egional and national report, but within 18 months and HCFather than unit-aggregated data. None of this collected dataomplied with the recommendations of the last Antibiotic Plan011-2016, which stipulated analyzing and using the antibi-tics consumption data at local, regional, and national level,ogether with feedback and use of data, and the comparison ofnformation on bacterial resistance and antibiotic use.

Our study results led to recommending the implementation of dedicated instrument for data collection at a national level, andvoiding re-transcription, particularly from hospital pharmacyanagement software. This new instrument should simplify

ollection for participating HCF by allowing them to importonsumption data directly into the instrument from the usualoftware, in the same way as for data on bacterial resistance, notnly for the HCF as a whole, but also for each functional unit.he instrument automatically carries out the conversion intoDD/1000 PD. Various educational and regular pre-formatted

harts and indicators would be available to monitor and manageonsumption more effectively, at department, unit, or HCF level,nd also in terms of regional, inter-regional, and national vali-ation to supply the national databases in a transparent manner,

ith much faster feedback delay.Users of antibiotic consumption data, both at local and

egional level, would like to be able to compare their data

ladies

wpiIphucHaam

rmoieertndmH

5

novl

aibpc

D

c

o

A

Sdpsa

A

ttap(mfoto(MUtstD

R

S. Henard et al. / Médecine et ma

ith that of equivalent HCF or departments, to improve theirractice. This benchmarking practice is effective and helpsmprove antibiotic use, and also reduce bacterial resistance [6].t is also becoming an important aspect of the policy for theroper use of antibiotics in HCF [7–9]. It should also be used inealthcare units, to encourage a greater participation of individ-al prescribers, and to identify other departments to implementorrective measures as soon as possible. It is essential that allCF use a standard method to collect consumption data, with

standard unit, and that the consumption measured be adjustedccording to relevant factors concerning patients, for this bench-arking technique to be valid [10].Up to now, the surveillance of antibiotic use and bacterial

esistance has been performed, both in Europe and worldwide,ainly by means of repeated prevalence surveys, including some

n a very wide scale [11]. These surveys identify targets formprovement in practice, and can be used repeatedly [12]. How-ver, these are still rare, require time and money, and do notnable HCF or prescribers to monitor antibiotic use and bacte-ial resistance in real time. The most effective method to achievehis, is still a dedicated electronic web tool [13]; however, untilow, no such applications have been developed, other than a tooleveloped by CCLIN-Est 2 years ago, currently being imple-ented on a national scale in France, after a pilot phase in a fewCF [14].

. Conclusion

The challenges identified by our study are the need to harmo-ize and rationalize the collection, processing, and interpretationf data on the use of and resistance to antibiotics, to offer a rele-ant, unique, reliable management instrument, on a national andocal (HCF) scale.

An essential goal is also to involve the HCF in the collectionnd use of data, particularly alongside prescribers, providingndicators and charts that can be used by prescribers, and alsoenchmarking data between comparable proprietary medicinalroducts/HCF. The involvement of prescribers in the analysis ofollected data could have an educational role.

isclosure of interest

The authors declare that they have no conflicts of interestoncerning this article.

Funding: This study was supported by the French Ministryf Health (Direction générale de la santé DGS).

uthors’ contributions

SB, SC, BD, TL, CR conceived and designed the study; SC,B, BD, SH, CR participated in the questionnaire collection andata management quality control; SC, SB, BD, SH, CR partici-ated in study coordination; SC, SB, BD and SH performed thetatistical analysis and wrote the manuscript. All authors readnd approved the final manuscript.

[

[

infectieuses 44 (2014) 308–314 313

cknowledgments

At the end of 2009, the French Ministry of Health (Direc-ion générale de la santé DGS) proposed to implement a studyo compare existing national and inter-regional databases onntibiotic use in hospitals (DREES, CClin, and ICATB databasesrovided by HCF, the French Health Products Safety AgencyAFSSAPS) database consisting of declarations made by phar-aceutical companies), to ensure the consistency of results

rom these databases. This study was assigned to the Antibi-lor Network, and was monitored by a steering committee forhe comparative study on national and inter-regional databasesn antibiotic use in hospitals, including Bruno COIGNARDNational Health Monitoring Institute INVS), Didier GUILLE-

OT (INSERM/Pasteur Institute), Florence LIEUTIER (Niceniversity Hospital), Vanessa VAN ROSSEMAGNANI (E2 sec-

ion, French Ministry of Health [Direction générale de l’offre deoins DGOS]), and Jean-Michel AZANOWSKY (RI 1-3 sec-ion, French Ministry of Health [Direction générale de la santéGS]).

eferences

[1] Boucher HW, Talbot GH, Benjamin Jr DK, Bradley J, Guidos RJ, JonesRN, et al. 10 × 20 Progress–development of new drugs active against Gram-Negative Bacilli: an update from the infectious diseases society of America.Clin Infect Dis 2013 Jun;56:1685–94.

[2] Ministère de la Santé. Plan national d’alerte sur les antibio-tiques 2011–2016; 2013 [Available at: http://www.sante.gouv.fr/IMG/pdf/plan antibiotiques 2011-2016 DEFINITIF.pdf, accessed 19 December2013].

[3] Ministère de la Santé. Plan national pour préserver l’efficacité des antibi-otiques 2001–2006; 2013 [Available at: http://www.plan-antibiotiques.sante.gouv.fr/IMG/pdf/Plan national pour preserve l efficacite desantibiotiques.pdf, accessed 19 December 2013].

[4] Ministère de la Santé. Plan national pour préserver l’efficacité des antibi-otiques 2007–2010; 2013 [Available at: http://www.sante.gouv.fr/IMG/pdf/bilan du plan pour preserver l efficacite des antibiotiques 2007-2010.pdf, accessed 19 December 2013].

[5] Henard S, Rahib D, Léon L, Amadéo B, Dumartin C, Cavalié P, et al.Consommation des antibiotiques rapportée via les bilans standardisés delutte contre les infections nosocomiales et relation avec l’ICATB. Med MalInfect 2011;41:197–205.

[6] Fridkin SK, Lawton R, Edwards JR, Tenover FC, McGowan Jr JE, GaynesRP, et al. Monitoring antimicrobial use and resistance: comparison with anational benchmark on reducing vancomycin use and vancomycin-resistantenterococci. Emerg Infect Dis 2002;8:702–7.

[7] Cantón R, Bryan J. Global antimicrobial resistance: from surveillance tostewardship. Part 2: stewardship initiatives. Expert Rev Anti Infect Ther2012;10:1375–7.

[8] Polk RE, Hohmann SF, Medvedev S, Ibrahim O. Benchmarking risk-adjusted adult antibacterial drug use in 70 US academic medicalcenter hospitals. Clin Infect Dis Off Publ Infect Dis Soc Am 2011;53:1100–10.

[9] Vlahovic-Palcevski V, Dumpis U, Mitt P, Gulbinovic J, Struwe J, PalcevskiG, et al. Benchmarking antimicrobial drug use at university hospitals in fiveEuropean countries. Clin Microbiol Infect Off Publ Eur Soc Clin MicrobiolInfect Dis 2007;13:277–83.

10] Ibrahim OM, Polk RE. Benchmarking antimicrobial drug use in hospitals.Expert Rev Anti Infect Ther 2012;10:445–57.

11] Ansari F, Erntell M, Goossens H, Davey P. The European Surveil-lance of Antimicrobial Consumption (ESAC) Point-Prevalence Survey

3 ladies

[

[

[14] Boussat S, Demoré B, Lozniewski A, Aissa N, Rabaud C. Comment

14 S. Henard et al. / Médecine et ma

of Antibacterial Use in 20 European hospitals in 2006. Clin Infect Dis2009;49:1496–504.

12] Willemsen I, Groenhuijzen A, Bogaers D, Stuurman A, Keulen P, van, et al.

Appropriateness of antimicrobial therapy measured by repeated prevalencesurveys. Antimicrob Agents Chemother 2007;51:864–7.

13] Zarb P, Amadeo B, Muller A, Drapier N, Vankerckhoven V, Davey P,et al. Identification of targets for quality improvement in antimicrobial

infectieuses 44 (2014) 308–314

prescribing: the web-based ESAC Point Prevalence Survey 2009. J Antimi-crob Chemother 2011;66:443–9.

améliorer le recueil et l’analyse des consommations hospitalièresd’antibiotiques: résultats préliminaires du déploiement en phase pilote del’outil ConsoRes. Med Mal Infect 2012;42:154–60.