Ciso

G

Journal of Global Antimicrobial Resistance xxx (2015) xxxxxx

G Model

JGAR-140; No. of Pages 8

Contents lists available at ScienceDirect

Journal of Global Antim

jo u rn al h om ep age: w ww.eOtto Cars d,e, Siping Dong c,*aDepartment of Health Science, University Medical Centre Groningen, University of Groningen, Groningen, The Netherlandsb School of Political Science and Public Administration, Wuhan University, Luojiashan, Wuchang District, Wuhan 430072 Hubei Province, Peoples Republic of

ChinacNational Institute of Hospital Administration, National Health and Family Commission of P.R. China, 38 Xueyuan Road, Haidian District, Beijing 100191,

Peoples Republic of ChinadAntibiotics and Infection Control, Public Health Agency of Sweden, 18 SE-171 82 Solna, SwedeneReAct Action on Antibiotic Resistance, Uppsala University, Box 256, SE-751 05 Uppsala, Sweden

1. Introduction

Overprescription of antimicrobials and increasing antimicrobialresistance are major public health issues in China and a signicantchallenge to global health. The World Health Assembly calledmember states to promote rational antibiotic use in order tocontain antimicrobial resistance [1]. As a result, the Chinesegovernment have issued and implemented a number of policies.

To assess the impact of these actions, a study was undertaken aspart of the Plan of Action for Sino-Swedish Health Cooperation20112014. The collaboration aims to apply Swedish expertise inpromoting appropriate use of antibiotics and to motivate theongoing work in China. Sweden has been used as a benchmarkbecause of its well-established and effective policies and legisla-tion on antibiotic use [2]. Sweden was also one of the rst countriesto initiate a cross-stakeholder co-operation for the containment ofantibiotic resistance [Swedish Strategic Programme againstAntibiotic Resistance (STRAMA)] [3]. These long-term andstructured interventions have proved a success as Sweden haslow antibiotic use and low resistance rates [4,5,7].

A R T I C L E I N F O

Article history:

Received 5 August 2014

Received in revised form 20 February 2015

Accepted 1 March 2015

Keywords:

Antibiotic

Antibiotic prophylaxis

Antimicrobial resistance

Anatomical Therapeutic Chemical (ATC)

Dened daily dose (DDD)

Consumption pressure

A B S T R A C T

Changes in patterns of antibiotic use in Chinese hospitals before and after intensive nationwide

interventions are reported and compared with Chinese national targets and antibiotic use in Swedish

hospitals. Chinese data were collected quarterly and yearly from selected patient prescriptions/medical

records and medicines inventory control systems from 15 hospitals (20052012). Swedish data were

extracted from a 2010-point prevalence survey and 20092012 sales data from seven university

hospitals. An interrupted time series with segmented regression analysis was used to measure changes

in patterns of antibiotic use in Chinese hospitals before and after the interventions. Following the 2011

interventions, signicant reductions in antibiotic use in Chinese hospitals were seen: the proportion of

prescriptions with antibiotics decreased 4.7% (P = 0.03) and the proportion of medical records with

antibiotic prescription decreased 7.3% (P = 0.04). The proportions of prescriptions and medical records

with antibiotics in Chinese hospitals in 2012 were 10% and 50%, respectively, and remained much higher

than Swedish hospitals (1.1% in DDD for outpatients and 34% in number of patients for inpatients).

Inpatient consumption in Chinese hospitals dropped signicantly from 910 DDD/1000 inpatient days in

2008 to 473 in 2012 (588 in Swedish hospitals). Antibiotics are being used less frequently in Chinese

hospitals, broad-spectrum antibiotics are still preferred, and overall usage is higher than Sweden. A

signicant reduction in overall inpatient antibiotic consumption was observed after the interventions. It

is not possible to identify whether the changes have resulted in less inappropriate antibiotic use. Further

studies are needed.

2015 International Society for Chemotherapy of Infection and Cancer. Published by Elsevier Ltd. Allrights reserved.

* Corresponding authors. Tel.: +86 10 6232 6305; fax: +86 10 6232 6305.

E-mail addresses: sunjingx@yahoo.com (J. Sun), sipingd@163.com (S. Dong).

http://dx.doi.org/10.1016/j.jgar.2015.03.001

2213-7165/ 2015 International Society for Chemotherapy of Infection and Cancer. Published by Elsevier Ltd. All rights reserved.Changes in patterns of antibiotic use in(20052012) and a benchmark compar

Jing Sun a,*, Xiao Shen b, Meng Li c, Liu He b, ShuyanPlease cite this article in press as: Sun J, et al. Changes in patterns benchmark comparison with Sweden in 2012. J Global Antimicrob Rhinese public hospitalsn with Sweden in 2012

uo c, Gunilla Skoog d, Malin Grape d,

icrobial Resistance

l s evier .c o m/lo c ate / jg arof antibiotic use in Chinese public hospitals (20052012) and aesist (2015), http://dx.doi.org/10.1016/j.jgar.2015.03.001

J. Sun et al. / Journal of Global Antimicrobial Resistance xxx (2015) xxxxxx2

G Model

JGAR-140; No. of Pages 8The health systems of China and Sweden are very different,Sweden being a country with a small population (9.5 million)compared with China (1.3 billion). Swedish hospitals are largelypublicly owned and the primary healthcare system is universal andeffective. Most patients enter the health system via primary care[811]. In contrast, the capacity of primary healthcare providers isstill under development in China, especially in the extensive ruralareas that lack skilled staff and modern equipment. Healthcaredelivery relies predominantly on hospital-based care, with publictertiary hospitals being the main healthcare providers. Primarycare institutions do not play a dominant role in providingoutpatient services, with most patients directly accessing hospitalseven for common diseases, as evidenced by only 41% of outpatientvisits being provided by primary care in 2012 [12,13].

This study measured changes in patterns of antibiotic use andinpatient consumption in Chinese public general tertiary hospitalsfrom 2005 to 2012, which covered the period before and after the2011 interventions. The 2012 observations in Chinese hospitalswere also compared with the Chinese national targets and Swedishhospitals.

Policy impact analysis in China usually uses simple before andafter comparison. Rigorous research studies using more sophisti-cated methods that can make more precise measurement of policyimpact are few [14]. To our knowledge, this is the rst assessmentof the effect of interventions on antibiotic use using a longitudinalstudy with interrupted time series data and segmented regressionanalysis. This methodology is the strongest quasi-experimentaldesign to evaluate longitudinal effects of time-delimited inter-ventions [15].

2. Materials and methods

2.1. Study design

Changes in patterns of antibiotic use in Chinese hospitals beforeand after the 2011 interventions (during March 2005 to December2012) were measured and the 2012 observations in Chinesehospitals were compared with the Chinese national targets [16,17]and with Swedish hospitals.

The 2011 interventions were the most intensive and far-reaching initiative of the Ministry of Health aimed at improvingantibiotic use in hospitals. Following extensive advocacy andmobilisation across the country, all levels of government ofciallyannounced the interventions in written documents and circulatedthe documents to each hospital. The internationally acceptedconcept of antibiotic stewardship was integrated into thesemeasures. Many co-ordinated measures were taken [1618],including: (i) setting targets; (ii) strengthening accountability;(iii) auditing clinical indications for antibiotic use; (iv) building thecapacity of infection control; (v) dening the prescribing of specicantimicrobials for different levels of professionals in clinicalpractice; (vi) restricting the number of agents prescribed; (vii)conducting regular antimicrobial resistance monitoring with analert system for multiresistance; (viii) organising guidelinetraining; (ix) conducting a monthly outpatient prescription andinpatient medical record audit, identifying the prescribers who arecontinuously not able to achieve the targets set by an individualhospital, and revoking prescribing rights for the frequent outliers;(x) setting up monitoring networks across hospitals and regions toshare information both at national and local levels; (xi) carryingout unannounced inspections at the targeted public hospitals; and(xiii) sharing good practice. Specic targets were set by theMinistry of Health as follows: the proportions of outpatient (OP)prescriptions and inpatient (IP) medical records with prescriptionof antibiotics to be

J. Sun et al. / Journal of Global Antimicrobial Resistance xxx (2015) xxxxxx 3

G Model

JGAR-140; No. of Pages 8indicators a, b and g were extracted from 2012 sales data of theseven university hospitals. Data for IP antibiotics and parenteralantibiotic use and the proportion of patients given antibioticprophylaxis for >24 h for indicators c, d and f were obtained fromthe most recent 2010-point prevalence survey.

2.4. Data analysis

The quarterly antibiotic use data collected from the OPprescription and IP medical records data for the Chinese hospitalswere plotted (March 2005 to December 2012) (Fig. 1). Annualantibiotic consumption for Chinese and Swedish hospitals during20052012 for systemic use and for specic ATC categories asDDD/1000 inpatient days and proportions consumed are shown inFig. 2.

Longitudinal time series quarterly data were analysed using asegmented linear regression model to assess changes in levels andtrends after the 2011 interventions in Chinese hospitals. Timeseries data analysis statistical software can control for autocorre-lated errors, which controls for secular trends and can also adjustfor potential serial correlation of the data [15,19]. Although theannouncement of the interventions was issued in April 2011,implementation started in August 2011. Considering that the studycollected quarterly data from Chinese hospitals in 2011, September2011 was adopted as the starting time point of the interventionsfor conducting segmented regression analysis. Changes in trendsbefore and after September 2011 and the change in levels inSeptember 2011 were compared using Stata 12.0 (StataCorp LP,College Station, TX).

Data for China and Sweden were compared using 2012indicators (ag) for Chinese hospitals, and the 2012 Swedish salesdata and 2010-point prevalence survey data.

3. Results

3.1. Changes in antibiotic use in Chinese hospitals before and after

introduction of the 2011 interventions

The proportion of OP prescriptions with parenteral antibioticswas stable at 24 h decreased from 70 to 90% to 60% (Fig. 1C).

Data are also presented on the trends in IP consumption of sometargeted antibiotic ATC classes (Fig. 2). Consumption of antibioticsfor systemic use (J01) was highest in 2008, thereafter decreasing,particularly during 20112012, as did other classes.

Segmented regression analysis of the quarterly data (Table 1)showed that before September 2011, OP and IP antibiotic use(indicators a and c), IP parenteral antibiotic use (indicator d) andpatients given antibiotic prophylaxis for >24 h (indicator f) hadalready decreased signicantly (P = 0.004, P = 0.002, P < 0.001 andP < 0.001, respectively) from the start of the study. OP parenteralantibiotic use (indicator b) steadily increased and antibioticprophylaxis given before incision (indicator e) steadily decreased.

In September 2011, the proportion of OP prescriptions given forat least one antibiotic (indicator a) and the proportion of IP medicalrecords with at least one antibiotic (indicator c) signicantlydropped by 4.7% [P = 0.03; 95% condence interval (CI) 8.9% to0.6%] and 7.3% (P = 0.04; 95% CI 14.2% to 0.4), respectively. Theproportion of IP medical records with at least one parenteralantibiotic prescribed (indicator d) and the proportion of antibioticprophylaxis of duration >24 h (indicator f) both droppedPlease cite this article in press as: Sun J, et al. Changes in patterns benchmark comparison with Sweden in 2012. J Global Antimicrob Rsignicantly by 8.6% (P = 0.03, 95% CI 16.1% to 1.1%; andP = 0.003, 95% CI 13.9% to 3.3%, respectively). The proportion ofOP prescriptions given with at least one parenteral antibiotic(indicator b) and the proportion of antibiotic prophylaxis givenbefore incision (indicator e) steadily decreased.

After September 2011, antibiotic prophylaxis of duration >24 h(indicator f) signicantly decreased (P = 0.002), with the otherindicators not changing signicantly.

3.2. Comparison of antibiotic use in Chinese and Swedish hospitals

Table 2 shows that 10% of OP prescriptions and 50% of IPmedical records include a prescription for antibiotics in Chinesehospitals. These are below the Chinese national targets (

(A)

0%

10%

20%

30%

Mar

-05

Jun-

05

Sep-

05

Dec-

05

Mar

-06

Jun-

06

Sep-

06

Dec-

06

Mar

-07

Jun-

07

Sep-

07

Dec-

07

Mar

-08

Jun-

08

Sep-

08

Dec-

08

Mar

-09

Jun-

09

Sep-

09

Dec-

09

Mar

-10

Jun-

10

Sep-

10

Dec-

10

Mar

-11

Jun-

11

Sep-

11

Dec-

11

Mar

-12

Jun-

12

Sep-

12

Dec-

12

% outpaent prescripon with anbiocs % outpaent prescripon with parenteral anbiocs

(B)

40%

50%

60%

70%

80%

Mar

-05

Jun-

05

Sep-

05

Dec-

05

Mar

-06

Jun-

06

Sep-

06

Dec-

06

Mar

-07

Jun-

07

Sep-

07

Dec-

07

Mar

-08

Jun-

08

Sep-

08

Dec-

08

Mar

-09

Jun-

09

Sep-

09

Dec-

09

Mar

-10

Jun-

10

Sep-

10

Dec-

10

Mar

-11

Jun-

11

Sep-

11

Dec-

11

Mar

-12

Jun-

12

Sep-

12

Dec-

12% Inp aent medical records with anbiocs % Inp aent me dical records with parenteral anbiocs

(C)

30%

40%

50%

60%

70%

80%

90%

100 %

Mar

-05

Jun-

05

Sep-

05

Dec-

05

Mar

-06

Jun-

06

Sep-

06

Dec-

06

Mar

-07

Jun-

07

Sep-

07

Dec-

07

Mar

-08

Jun-

08

Sep-

08

Dec-

08

Mar

-09

Jun-

09

Sep-

09

Dec-

09

Mar

-10

Jun-

10

Sep-

10

Dec-

10

Mar

-11

Jun-

11

Sep-

11

Dec-

11

Mar

-12

Jun-

12

Sep-

12

Dec-

12

% surgery with AB prophylaxis befor e incision % AB proph ylaxis duraon >24 hou rs

Fig. 1. Proportions of (A) outpatient prescriptions of antibiotic and (B) inpatient medical records with antibiotic, and (C) proportions of surgical patients given antibioticprophylaxis before incision and for a duration of >24 h in Chinese hospitals. Data source: selected outpatient prescriptions and inpatient medical records from 15 Chinese

public tertiary general hospitals.

J. Sun et al. / Journal of Global Antimicrobial Resistance xxx (2015) xxxxxx4

G Model

JGAR-140; No. of Pages 8

Please cite this article in press as: Sun J, et al. Changes in patterns of antibiotic use in Chinese public hospitals (20052012) and abenchmark comparison with Sweden in 2012. J Global Antimicrob Resist (2015), http://dx.doi.org/10.1016/j.jgar.2015.03.001

J. Sun et al. / Journal of Global Antimicrobial Resistance xxx (2015) xxxxxx 5

G Model

JGAR-140; No. of Pages 8subtle trends in indicator g. IP consumption of antibiotics forsystemic use (J01) decreased in 2008 and further decreasedfollowing the 2011 interventions. Until December 2012, itquickly dropped to almost one-half of the highest level. Broad-spectrum antibiotics always dominated prescribing during20052012 and reached a peak in 2012. Conversely, consump-tion of narrow-spectrum antibiotics continuously decreasedduring 20052012.

Fig. 2. Inpatient antibiotic (J01 systemic use) consumption by Anatomical Therapeutic Chmedicines inventory control systems of 15 Chinese public general tertiary hospitals; (2) S

not measured (J01A, J01B, J01DF, J01DI, J01 E, J01FF, J01FG, J01G, J01MB and J01X) are

calculated as dened daily doses (DDD) per 1000 inpatient days.

Please cite this article in press as: Sun J, et al. Changes in patterns benchmark comparison with Sweden in 2012. J Global Antimicrob R4.2. Comparison of antibiotic use in Chinese and Swedish hospitals

IP antibiotic consumption of Swedish hospitals has been wellcontrolled at a stable level over time and, like most other EUcountries, penicillins (J01C) are always the most frequently usedantibiotics [25]. In contrast, IP antibiotic consumption in Chinesehospitals dropped sharply from an unacceptably high level toequivalent to the 2012 Swedish level. The absolute values of each

emical (ATC) categories in (A) Chinese and (B) Swedish hospitals. Sources: (1) China:

weden: sales data of seven Swedish university hospitals. Notes: (1) The ATC classes

included as Others in the bar; (2) Inpatient antibiotic consumption pressure is

of antibiotic use in Chinese public hospitals (20052012) and aesist (2015), http://dx.doi.org/10.1016/j.jgar.2015.03.001

fte

J. Sun et al. / Journal of Global Antimicrobial Resistance xxx (2015) xxxxxx6

G Model

JGAR-140; No. of Pages 8Table 1Segmented regression analysis results comparing antibiotic use before, during and a

Independent variable (indicator) Coefcienta

(a) % of OP prescriptions with antibiotic b0b1studied ATC category all decreased. The proportion of third-generation cephalosporins (J01DD) prescribed to all prescribedantibiotics for systemic use (J01) did not change and it was alwaysthe dominant class, whereas penicillins (J01C) fell throughout theperiod.

b2b3

(b) % of OP prescriptions with parenteral antibiotic b0b1b2b3

(c) % of IP medical records with antibiotic b0b1b2b3

(d) % of IP medical records with parenteral antibiotic b0b1b2b3

(e) % of antibiotic prophylaxis before incision b0b1b2b3

(f) % of antibiotic prophylaxis duration >24 h b0b1b2b3

SE, standard error; CI, condence interval; OP, outpatient; IP, inpatient.a b0 is the intercept of the pre-intervention trend line (value of the variable at the star

trend); b2 is the immediate post-intervention absolute change (immediate effect of the inof the post-intervention trend line.

b Bold signies statistically signicant coefcient (P < 0.05).

Table 2Comparison of antibiotic use patterns and consumption pressure in Chinese public genera

Outcome measure 15 Chinese public general tertiary hos

(a) Proportion of OP prescriptions with

antibiotic

10% (n = 18,000, S.D. = 1.1, 95% CI [8.0,

(b) Proportion of OP prescriptions with

parenteral antibiotic

2% (n = 18,000, S.D. = 0.6, 95% CI [1.0, 3

(c) Proportion of IP medical records

with antibiotic

50% (n = 5400, S.D. = 2.0, 95% CI [38.0,

(d) Proportion of IP medical records

with parenteral antibiotic

46% (n = 5400, S.D. = 2.0, 95% CI [42.3,

(e) Proportion of antibiotic prophylaxis

before incision

55% (n = 1708, S.D. = 7.5, 95% CI [39.3,

(f) Proportion of antibiotic prophylaxis

duration >24 h

56% (n = 1708, S.D. = 8.1, 95% CI [38.5,

(g) Total IP consumption of antibiotics

for systemic use (J01)

473 DDD/1000 inpatient days (n = 15,

S.D. = 29.0, median 464.4)

IP consumption of penicillins (J01C) 51 DDD/1000 inpatient days (10.8% of

(n = 15, S.D. = 7.7, median 40.0)

IP consumption of quinolones (J01 M) 51 DDD/1000 inpatient days (n = 15, S

median 47.9)

IP consumption of rst-generation

cephalosporins (J01DB)

44 DDD/1000 inpatient days (n = 15, S

median 44.8)

IP consumption of second-generation

cephalosporins (J01DC)

82 DDD/1000 inpatient days (n = 15, S

median 73.7)

IP consumption of third-generation

cephalosporins (J01DD)

105 DDD/1000 inpatient days (22.2% o

(n = 15 hospitals, S.D. = 14.7, median 93

IP consumption of fourth-generation

cephalosporins (J01DE)

19 DDD/1000 inpatient days (n = 15,

S.D. = 5.1, median 17.0)

OP, outpatient; IP, inpatient; S.D., standard deviation; CI, condence interval; N/A, nota Ministry of Health of PR China.b In DDD.c 2012 sales data.d 2010 Swedish Strategic Programme against Antibiotic Resistance (STRAMA) point

Please cite this article in press as: Sun J, et al. Changes in patterns benchmark comparison with Sweden in 2012. J Global Antimicrob Rr the 2011 interventions in China, 20052012.

SE P-valueb 95% CI

22.8 0.8 0.000 21.2, 24.5

S0.2 0.05 0.004 S0.3, S0.06The signicant reductions in OP and IP antibiotics do notaddress more appropriate use. However, there are achievementsmeasured against a background of an unacceptable baseline.Although the proportions of OP and IP antibiotic use achieved thenational targets following the 2011 interventions (there were no

S4.7 2.0 0.03 S8.9, S0.6S0.4 0.5 0.4 S1.4, 0.62.6 0.7 0.001 1.1, 4.10.04 0.05 0.4 S0.06, 0.1

S0.7 1.1 0.5 S3.0, 1.6S0.1 0.3 0.7 S0.8, 0.574.7 1.4 0.000 72.0, 77.5

S0.3 0.09 0.002 S0.5, S0.1S7.3 3.4 0.04 S14.2, S0.4S1.5 0.8 0.07 3.2, 0.275.2 1.5 0.000 72.2, 78.2

S0.4 0.1 0.000 S0.6, S0.2S8.6 3.7 0.03 S16.1, S1.1S1.3 0.9 0.1 S3.1, 0.542.8 3.7 0.000 35.3, 50.4S0.06 0.2 0.8 S0.5, 0.45.0 6.0 0.4 S7.3, 17.41.7 1.6 0.3 S1.5, 4.9

95.6 1.1 0.000 93.4, 97.8S0.8 0.07 0.000 S0.9, S0.7S8.6 2.6 0.003 S13.9, S3.3S2.2 0.6 0.002 S3.4, S0.9

t of the observation); b1 is the slope of the pre-intervention trend line (the baselinetervention); b3 is the change of the post-intervention trend; and b1 +b3 is the slope

l tertiary hospitals with the Chinese national targets and the Swedish levels in 2012.

pitals Chinese national targetsa Seven Swedish university

hospitals

12.8])

J. Sun et al. / Journal of Global Antimicrobial Resistance xxx (2015) xxxxxx 7

G Model

JGAR-140; No. of Pages 8targets for parenteral antibiotic use), antibiotic prophylaxis usageand practice still falls short of the national targets as well asSwedish usage.

It is concerning that Chinese hospitals still use more broad-spectrum antibiotics than narrow-spectrum antibiotics and thatsuch preferences did not change. Broad-spectrum antibiotics havebeen generally regarded by many Chinese doctors and the generalpublic as the more efcient treatment to address all kinds ofinfections. There is also a reluctance to use diagnostics and a heavyreliance on empirical treatment. The second factor is that theChinese guideline requires an allergy test for penicillins (excludinga few oral formulations) [26]. Chinese doctors tend to avoid skintesting procedures and are averse to any potential risk of disputewith the patient due to allergy, which contrasts with most othercountries where a negative allergy history is accepted. Resistancerates in China are much higher than Sweden and prescribersbelieve immediate use of broad-spectrum agents give the bestempirical cover. Greater availability of resistance data shouldencourage the use of narrow-spectrum agents.

Although overuse of antibiotics had been reduced in Chinesehospitals through the joint efforts during 20022011, theChinese national targets are yet to be met and there is asignicant gap compared with Swedish hospitals. The primarycare system in China was not as effective as in Sweden,particularly for patients referred to hospital. Despite efforts tobuild capacity in primary care in China, patients still prefer todirectly seek hospital care due to a lack of trust, no compulsoryreferral mechanism and a weak referral capacity in primary care.Approximately 60% of the medicines market is dominated bytertiary hospitals in China [27,28], and Chinese patients willalways go directly to tertiary hospitals for common diseases. InSweden, primary care forms the foundation of the healthcaresystem. Although there is no formal gate-keeping role forprimary care in most county councils, it plays an important rolein guiding the patient to the right care within the health system[2,11]. It might be inferred that Chinese patients require lessintensive antibiotic treatment than patients in the equivalentSwedish settings, which should lead to a lower consumption inChina. Alternatively, Swedish IP consumption data were collect-ed from sales data, which might be higher than the data collectedfrom medicines procurement and inventory managementsystems leading to the apparent anomaly.

Chinese and Swedish hospitals have opposing trends for the useof narrow- and broad-spectrum antibiotics but similar IP antibiotic(for systemic use) consumption. Such a contradiction could beexplained by the 2011 interventions causing a rapid signicantreduction in overall antibiotic consumption. The high rate ofantibiotic use might be due to unnecessary treatment of minorillnesses and non-bacterial infections, combined with under-treating serious bacterial infections owing to tight restrictions onantibiotic use. Further studies are needed in China to investigateguideline compliance and hospital-acquired infection rates. Inaddition, as our consumption measurement does not include OPand community pharmacies, we do not know whether IP antibioticuse was shifted to OP or over-the-counter sales.

Another reason, which should be considered is that the Chineselongitudinal data of antibiotic use indicators (af) might be anunderestimate as they did not include data from departments withpotential high use (e.g. infectious diseases).

Ideally, we would have used the existing Chinese nationalmonitoring network data to conduct this study. However, due tothe unavailability of national data, we had to conduct a samplingstudy whilst social, demographic and economic characteristics ofdifferent geographic areas were considered in selecting hospitals.The hospitals were all part of a special national monitoringnetwork and may be atypical.Please cite this article in press as: Sun J, et al. Changes in patterns benchmark comparison with Sweden in 2012. J Global Antimicrob RThe internationally recognised ATC/DDD system was used inour comparisons of patterns of antibiotic use between China andSweden. However, the Chinese medicines classication system isdifferent from the ATC system and different prescribed doses areused in different countries. We only targeted the antibioticcategories that were similar between the Chinese system andthe ATC system to attempt to reduce this bias.

There are also some limitations in the Swedish data. Unlike theIP consumption data collected in China, sales data of antibiotics tothe hospital wards were extracted, but not what was actually usedby the patients. This may result in an overestimate of antibiotic usein Sweden. In addition, due to the limitation of Swedish dataavailability (consumption data only in 20092012, IP antibiotic useonly in 2010), these do not exactly match with the years of theChinese data (consumption data in 20052012, IP antibiotics usedata in 2012).

Segmented regression analysis of interrupted time series data isthe strongest quasi-experimental design to evaluate longitudinaleffects of time-delimited interventions. This evidence would bestronger if a control setting was available. Due to the restriction ofdata availability, we were not able to obtain control data.

Data from this study do not enable us to understand whetherthe changes in China have resulted in less inappropriate antibioticuse or whether antibiotic use has been shifted from IP to othersources. Some intervention strategies such as revoking prescrip-tion rights of the frequent outliers may lead to appropriateantibiotic use being denied. Sophisticated policies are needed toguide antibiotic use rather than simple restriction of use. There isalso a clear need to obtain better surveillance of both antibioticresistance and usage to guide future interventions.

Funding

The Swedish International Development Agency providedfunding for this study [Sidas insatsnummer: 5504000101]; theywere not involved in the study design, data collection, analysis andinterpretation of data, nor were they involved in the writing of thereport or in the decision to submit the article for publication.

Conict of interest

The authors declare that they have no conict of interest.

Ethical approval

Not required.

Acknowledgments

The authors would like to thank the National Health and FamilyPlanning Commission of China who endorsed the study. Theauthors also thank the 15 Chinese hospitals (The First Hospital ofChina Medical University, Jilin University First Hospital, BeijingXuanwu Hospital, Hebei Medical University Second Hospital,Afliated Hospital of Jining Medical University, Anhui ProvincialHospital, Nanjing Drum Tower Hospital, Wuhan Third Hospital,Tongji Hospital, West China University Hospital, West ChinaSecond University Hospital, Sun Yat-sen University First Hospital,Zhuhai Peoples Hospital, Xijing Hospital and General Hospital ofNingxia Medical University) and the seven Swedish tertiaryhospitals (Karolinska University Hospital, Sahlgrenska UniversityHospital, Skane University Hospital, Uppsala University Hospital,Linkoping University Hospital, Orebro University Hospital andUniversity Hospital of Umea) who provided the outpatientprescriptions and inpatient medical records data as well as theof antibiotic use in Chinese public hospitals (20052012) and aesist (2015), http://dx.doi.org/10.1016/j.jgar.2015.03.001

consumption data. Special thanks to Prof. Jinhu Wu (Wuhan ThirdHospital), Prof. Yufeng Ding (Tongji Hospital), Dr. Zhaoyi Yang(Anhui Provincial Hospital), Dr. Liang Huang (West China SecondUniversity Hospital) and Dr. Haixia Zhang (Nanjing Drum TowerHospital) who provided valuable technical support to the study.The authors also thank Prof. Peter Hawkey (University ofBirmingham and Public Health England) for providing valuablecomments and editorial support.

References

[1] World Health Organization. World Health Assembly Resolutions: 51.17/1998,emerging and other communicable diseases: antimicrobial resistance; 58.27/2005 improvingthe containment of antimicrobial resistance; 60.16/. In: Progressin the Rational Use of Medicines, Including Better Medicines for Children; 2007.

[2] Anell A, Glenngard AH, Merkur S. Sweden: health system review. Health SystTransit 2012;14:1159.

[3] Public Health Agency of Sweden. Swedish work on containment of antibioticresistance tools, methods and experiences. Stockholm, Sweden: Public HealthAgency of Sweden; 2014 , http://www.folkhalsomyndigheten.se/publicerat-material/publikationer/Swedish-work-on-containment-of-antibiotic-resistance/ [accessed 13.02.15].

[4] Molstad S, Erntell M, Hanberger H, Melander E, Norman C, Skoog G, et al.Sustained reduction of antibiotic use and low bacterial resistance: 10-yearfollow-up of the Swedish Strama programme. Lancet Infect Dis 2008;8:12532.

[5] Strama. Swedish point prevalence studies on antibiotic use in hospitals. 20032010. Available at: http://www.folkhalsomyndigheten.se/amnesomraden/statistik-och-undersokningar/antibiotikastatistik/sverige/slutenvard/

[7] The Swedish Institute for Communicable Disease Control and National Veteri-nary Institute. SWEDRES/SVARM. Use of antimicrobials and occurrence ofantimicrobial resistance in Sweden. Uppsala, Sweden: SMI & SVA; 2012.

[13] Center for Health Statistics and Information, Ministry of Health, P.R. China.Research on health service of primary care facilities in China 2008. Report No.1 of the 4th National Health Service Survey. Beijing, China: China UnionMedical College Publishing House; 2009.

[14] Sun J. Systematic review of interventions on antibiotic prophylaxis insurgery in Chinese hospitals during 20002012. J Evid Based Med 2013;6:12635.

[15] Wagner AK, Soumerai SB, Zhang F, Ross-Degnan D. Segmented regressionanalysis of interrupted time series studies in medication use research. J ClinPharm Ther 2002;27:299309.

[16] Xiao Y, Zhang J, Zheng B, Zhao L, Li S, Li L. Changes in Chinese policies topromote the rational use of antibiotics. PLOS Med 2013;10:e1001556.

[17] Ministry of Health of P.R. China. The announcement of carrying out intensivenationwide interventions on antimicrobial clinical use and its implementationstrategy. Document No. 56; 2011, http://www.moh.gov.cn/mohyzs/s3586/201104/51376.shtml [accessed 13.02.15].

[18] Xiao Y, Li L. Legislation of clinical antibiotic use in China. Lancet Infect Dis2013;13:18991.

[19] Lagarde M. How to do (or not to do). . . assessing the impact of a policy changewith routine longitudinal data. Health Policy Plan 2011;1:7683.

[20] Ministry of Health and State Administration of Traditional Chinese Medicine ofP.R China. Provisional regulation on pharmacy administration of the healthfacility. Document No. 11; 2011, http://www.gov.cn/zwgk/2011-03/30/content_1834424.htm [accessed 13.02.15].

[21] Ministry of Health of P.R. China. Regulation on prescriptions. Document No.269; 2004, http://www.moh.gov.cn/mohbgt/pw10409/200804/18354.shtml[accessed 13.02.15].

[22] National Health Family Planning Commission of P.R. China. Guideline forantimicrobials clinical use. Document No. 285; 2004, http://www.nhfpc.gov.cn/zhuzhan/zcjd/201304/2da63e01325e4df8a235590112321403.shtml[accessed 13.02.15].

[23] Ministry of Health P.R. China. Announcement of setting up the monitoringnetwork for antimicrobials clinical use and resistance; 2005, http://www.nhfpc.gov.cn/zhuzhan/zcjd/201304/b0da0ebfc7b3428f98d435a29cfa4250.shtml [accessed 13.02.15].

[24] National Health Family Planning Commission of P.R. China. Guideline for

J. Sun et al. / Journal of Global Antimicrobial Resistance xxx (2015) xxxxxx8

G Model

JGAR-140; No. of Pages 8[8] Goossens H, Ferech M, Stichele RV, Elseviers M. Outpatient antibiotic use inEurope and association with resistance: a cross-national database study.Lancet 2005;365:57987.

[9] OECD. OECD Reviews of Health Care Quality. Sweden: OECD; 2013, http://www.oecd-ilibrary.org/social-issues-migration-health/oecd-reviews-of-health-care-quality-sweden-2013_9789264204799-en[accessed 13.02.15].

[10] Masseria C, Irwin R, Thomson S, Gemmill M, Mossialos E. Primary care inEurope. Policy Brief. Directorate-General Employment, Social Affairs andEqual Opportunities Unit E1Social and Demographic Analysis. Brussels,Belgium: European Commission; 2009.

[11] Albin B, Hjelm K, Zhang W. Health care systems in Sweden and China: legaland formal organization aspects. Health Res Policy Syst 2010;8:20.

[12] National Health Family Planning Commission of P.R. China. China HealthStatistics; 2013, http://www.nhfpc.gov.cn/ewebeditor/uploadle/2014/04/20140430131845405.pdf [accessed 13.02.15].Please cite this article in press as: Sun J, et al. Changes in patterns benchmark comparison with Sweden in 2012. J Global Antimicrob Rprescription audit. Document No. 28; 2010, http://www.nhfpc.gov.cn/zwgkzt/glgf/201306/094ebc83dddc47b5a4a63ebde7224615.shtml [accessed13.02.15].

[25] European Centre for Disease Prevention Control. Summary of latest data onantibiotic consumption in the European Union-highlights on antibiotic consump-tion; 2011, http://www.ecdc.europa.eu/en/eaad/documents/eaad-2011-summary-antimicrobial-consumption-data.pdf [accessed 13.02.15].

[26] Chinese Pharmacopoeia Commission Clinical drug information. Beijing, China:Chinese Medical Science and Technology Press; 2011.

[27] Health I.M.S. Review of the Chinese pharmaceutical market in 2012. http://blog.sina.com.cn/s/blog_59834aba0102eqpk.html [accessed 13.02.15]

[28] Sino-Health Ltd Report of Chinese pharmaceutical retail market development20112012. http://blog.sina.com.cn/s/blog_70b99cd80102e7zv.html [accessed13.02.15]of antibiotic use in Chinese public hospitals (20052012) and aesist (2015), http://dx.doi.org/10.1016/j.jgar.2015.03.001

Changes in patterns of antibiotic use in Chinese public hospitals (20052012) and a benchmark comparison with Sweden in 20121 Introduction2 Materials and methods2.1 Study design2.2 Outcome measures2.3 Study population, setting and data source2.3.1 China2.3.2 Sweden

2.4 Data analysis

3 Results3.1 Changes in antibiotic use in Chinese hospitals before and after introduction of the 2011 interventions3.2 Comparison of antibiotic use in Chinese and Swedish hospitals

4 Discussion4.1 Changes in antibiotic use in Chinese hospitals before and after introduction of the 2011 interventions4.2 Comparison of antibiotic use in Chinese and Swedish hospitals

FundingConflict of interestEthical approvalAcknowledgmentsReferences