What Is Driving the Epidemiology of Methicillin-Resistant Staphylococcus aureus Infections in Europe?

Abstract

Introduction:

The burden from infections caused by methicillin-resistant Staphylococcus aureus (MRSA) in the European Union/European Economic Area (EU/EEA) has increased in recent years, especially in the higher prevalence southern and eastern countries. Addressing this challenge requires a clear knowledge of the factors driving this epidemiology to inform effective national interventions and campaigns.

Methods:

We identified national quality metrics for antibiotic use and hand hygiene from the 2016/2017 European Centre for Disease Control (ECDC) Point Prevalence study as well as structural, cultural, and governance indicators from other sources. We then utilized general linear modeling to identify parsimonious correlations with national MRSA proportions reported by the European Antimicrobial Resistance Surveillance Network (EARS-Net).

Results:

The main process predictor of MRSA prevalence in EU/EEA countries was the use of more than one concurrent antibiotic per patient. The impact of hand hygiene was less clear, possibly because consumption of alcohol hand-rub was suboptimal throughout Europe. Hospital and infection control structural factors did not appear relevant at overall national level. Culture and governance were collectively key predictor groups; uncertainty avoidance, masculinity, and corruption indices strongly correlated with MRSA prevalence.

Discussion:

Our results suggest that the critical antibiotic stewardship factor associated with MRSA in EU/EEA countries is the quality of antibiotic prescribing, especially spectrum of activity, rather than overall consumption levels in hospitals or proportion of patients treated. Above all, MRSA hyperendemicity is correlated with a set of sociocultural behavioral constructs that typically manifest themselves in lack of urgency to address risk and normalization of deviance in relation to noncompliant practices.

Introduction

Infections caused by methicillin-resistant Staphylococcus aureus (MRSA) continue to pose a major challenge to health care systems in many European countries. A recent landmark publication by Cassini et al. estimated that almost 150,000 MRSA infections occur every year in countries of the European Union (EU) and the European Economic Area (EEA), resulting in >7,000 attributable deaths.¹ The socioeconomic burden of these infections ranges from 29.8 to 35.6 disability-adjusted life years per 100,000 population. This is second only to infections caused by Escherichia coli resistant to third-generation cephalosporins. More importantly, although the European Antimicrobial Resistance Surveillance Network (EARS-Net), run by the European Centre for Disease Control (ECDC), reported a decrease in pan-European MRSA proportions from 26.6% in 2007 to 16.8% in 2015, Cassini et al. estimated that the incidence of MRSA infections in the EU/EEA actually increased by 1.28 times (95% Uncertainty Interval 1.11–1.47) during the same period.¹

It has been known for more than two decades that MRSA prevalence in the EU/EEA varies significantly, with a clear north to south/east gradient.² Indeed, every country reporting above-median MRSA proportions to EARS-Net lies within the south and east of the region. This situation has remained largely unchanged over the past decade. Even the two countries (Portugal and Romania) that reported a statistically significant reduction still continue to exhibit very high levels (39.2% and 44.4%, respectively in 2018).³ It is clear that noteworthy MRSA control is still far from becoming a reality in highly endemic European regions.

Effective interventions aimed at prevention and control can only take place if the drivers of MRSA infections in these countries are fully identified. MRSA spread in hospitals has been historically attributed to cross-transmission through the hands of health care workers and to selection pressure exerted by antibiotic use.⁴ More recently, structural and socioeconomic drivers have been recognized as key predictors of antibiotic quality indicators in the same European region,⁵ and could therefore also be additionally promising avenues to investigate.

Methods

IRB approval not indicated since only public data were used.

For the purpose of the study, the proportion of MRSA in invasive blood culture isolates within EU/EEA countries was extracted from the report Surveillance of Antimicrobial Resistance in Europe 2018, published by the ECDC.⁶ This surveillance network incorporates all EU/EEA countries except Lichtenstein. We then proceeded to source possible predictive variables from the ECDC report of the 2016/2017 Point Prevalence Survey (PPS) of Healthcare-Associated Infections and Antimicrobial Use in European Acute Care Hospitals.⁷ This survey was undertaken using standardized methodology in all EU/EEA countries except Denmark and Sweden; no data from these two countries were therefore included in any analyses. We also sought to identify structural and socioeconomic variables that could potentially influence MRSA prevalence within EU/EEA countries. Antibiotic use has been cited as an important driver of MRSA acquisition at hospital level.⁸ We included, per country, the total average antibiotic use in hospitals as defined daily doses (DDD) per 100 patient-days and the average DDD per patient receiving an antibiotic prescription. We used average proportion of hospital patients concurrently receiving more than one antimicrobial as a marker for broad-spectrum antibiotic use and proportion of antibiotics administered for surgical prophylaxis in excess of 24 hours as an indicator of prescribing practices that were not evidence based. Because hand hygiene has also been reported to impact on the prevalence of MRSA in health care settings,⁹ we included median alcohol hand rub (AHR) consumption per country (in L/1,000 patient-days) from the same PPS report.

Structural and system factors, such as overcrowding and understaffing, have also been cited as important drivers of MRSA acquisition in hospitals.¹⁰ We extracted median percentage of occupied hospital beds at midnight reported in the 2016/2017 PPS as well as the incidence of practicing nurses and doctors per 1,000 population and the hospital spending in Euros per capita for each EU/EEA country, from the OECD/EU Health at a Glance report 2018.¹¹ Because MRSA is primarily spread in a clonal manner, effective infection prevention and control (IPC) infrastructures and activities should reduce the risk of cross-transmission and therefore overall incidence.¹² From the 2016/2017 PPS report, we identified the number of IPC nurses and doctors working as full-time equivalents per 250 hospital beds, together with the median percentage of single-room beds among the total number of hospital beds in each participating country within the ECDC PPS.

Sociocultural factors have been quoted to be possible drivers for inappropriate behavior facilitating antimicrobial resistance in hospitals, including MRSA.¹³ These studies have utilized the anthropological model developed by Geert Hofstede, which proposes that national cultures vary along consistent, fundamental, dimensions.¹⁴ Scores for the Hofstede behavioral constructs of: power distance (PDI), individualism (IDV), masculinity (MAS), uncertainty avoidance (UAI), long-term versus short-term orientation (LTO), and indulgence versus restraint (IVR) for all EU/EEA countries under study (except Cyprus) were available and collated from his publicly available resource. Poor governance has also been linked to antimicrobial resistance.¹⁵ We obtained governance indicators for: Voice and Accountability, Political Stability and Absence of Violence, Government Effectiveness, Regulatory Quality, Rule of Law and Control of Corruption from the 2018 Worldwide Governance Indicators published by the World Bank.

The respective data per EU/EEA country were input into Excel (Microsoft Excel 2016; Microsoft Corp.) for preliminary evaluation. The predictors were grouped into the mentioned four groups: process, structural, sociocultural, and governance. The distribution of MRSA proportions was right skewed and did not satisfy normality assumption. Therefore, a generalized linear model (GLM) was adopted, assuming a Gamma distribution and a reciprocal link function, to relate MRSA proportions of the EU/EEA countries to possible predictors collectively. A GLM was fitted for each group of predictors in a univariate manner. For each model fit, the deviance was used as the measure of goodness of fit. A backward procedure was then used to identify the significant predictors that best explain the variation in the MRSA proportions, assuming a 0.05 level of significance. Because some of predictors were highly correlated, the backward elimination procedure was able to remove redundant predictors and reduce multicollinearity problems. The parsimonious model was determined by including solely significant predictors in the model fit and the predictor combination that yielded the lowest deviance. Subsequently, all the predictors that were found to be significantly related with the MRSA proportions in each individual parsimonious GLM were then reanalyzed together in the same manner. All the GLMs were fitted using the facilities of the Statistical Package for the Social Sciences, version 25 (SPSS, Inc., Chicago, IL).

Results

Average antibiotic use per patient (DDD) and prescribing more than one antimicrobial per patient correlated significantly with MRSA proportions (Table 1). MRSA prevalence was higher in countries with increased average antibiotic use per patient and higher number of patients prescribed more than one antimicrobial. However, no significant relationship could be established between MRSA prevalence and overall antibiotic consumption as DDD/100 bed days (BD) or surgical prophylaxis in excess of 24 hours. AHR consumption ranged from 6.4 L/1,000 BD in the least consuming country to 59.7 L/1,000 BD in the highest (interquartile range: 13.3–37.0). It failed to show any relationship with the dependent variable. The only significantly associated structural indication was the number of nurses per 1,000 inhabitants (Table 2). Three cultural variables—PDI, MAS, and UAI—were retained in the GLM (Table 3). In particular, UAI showed a very strong relationship with MRSA (Wald chi-square = 21.47; p < 0.001). All three cultural dimensions were positively related with MRSA proportions, implying that the prevalence of MRSA increases with an increase in these three cultural dimension scores. Of the World Bank governance indices, only control of corruption achieved a significant relationship with MRSA (p < 0.001) although accountability had a p-value that almost reached significance (Table 4). Control of corruption was negatively related to MRSA proportion; the prevalence of MRSA increased with reduced corruption control. When all the individual significant predictors were analyzed collectively, the GLM retained as significant predictors of MRSA proportions in EU/EEA countries: use of more than one antimicrobial per patient, control of corruption, masculinity, and uncertainty avoidance. Table 5 provides the parameter estimates and corresponding standard errors and 95% confidence intervals for these predictors. Because a reciprocal link function was used, a negative parameter estimate indicates a positive relationship between the prevalence of MRSA and the predictor, whereas a positive parameter estimate indicates a negative relationship. The change in deviance was used to investigate the goodness of fit of this 4-predictor GLM compared with the minimal model. This change in deviance (15.758) had a chi-square distribution with 4 degrees of freedom and was sufficiently large to indicate a significantly better model fit (p = 0.003).

Table 1.

Generalized Linear Model Relating Methicillin-Resistant Staphylococcus aureus Prevalence in European Union/European Economic Area Countries to Process Predictors

Antibiotic and IPC process predictors	n	Univariate		Parsimonious
Antibiotic and IPC process predictors	n	Wald χ²	P	Wald χ²	p
Overall antibiotic consumption (DDD/100 BD)	27	1.374	0.241	—	—
Average antibiotic consumption per patient (DDD)	27	4.382	0.036	6.425	0.011
Patients with >1 antimicrobial (%)	27	4.412	0.036	14.337	<0.001
Surgical prophylaxis in excess of 24 hours (%)	28	0.071	0.790	—	—
Alcohol hand run consumption (L/1,000 BD)	28	0.252	0.616	—	—

—, Factor rejected by parsimonious model; DDD, defined daily doses; IPC, infection prevention and control; BD, bed days.

Table 2.

Generalized Linear Model Relating Methicillin Resistant Staphylococcus aureus Prevalence in European Union/European Economic Area Countries to Structural Predictors

Structural and system predictors	n	Univariate		Parsimonious
Structural and system predictors	n	Wald χ²	P	Wald χ²	p
Hospital spending per capita (€)	28	0.593	0.441	—	—
Doctors per 1,000 population	28	0.789	0.374	—	—
Nurses per 1,000 population	28	5.110	0.024	10.233	0.001
Infection control nurses FTE per 250 beds	28	0.027	0.869	—	—
Infection control doctors FTE per 250 beds	28	0.163	0.686	—	—
Median of occupied hospital beds at midnight (%)	28	1.321	0.250	—	—
Median of single-room beds from total hospital beds (%)	27	0.095	0.759	—	—

—, Factor rejected by parsimonious model; FTE, full-time equivalents.

Table 3.

Generalized Linear Model Relating Methicillin-Resistant Staphylococcus aureus Prevalence in European Union/European Economic Area Countries to Cultural Predictors

Cultural predictors (scores)	n	Univariate		Parsimonious
Cultural predictors (scores)	n	Wald χ²	p	Wald χ²	p
Power distance	27	2.055	0.152	7.315	0.007
Individualism	27	0.198	0.656	—	—
Masculinity	27	7.081	0.008	6.483	0.011
Uncertainty avoidance	27	7.101	0.008	21.470	<0.001
Long-term orientation	27	0.926	0.336	—	—
Indulgence	27	0.278	0.598	—	—

—, Factor rejected by parsimonious model.

Table 4.

Generalized Linear Model Relating Methicillin-Resistant Staphylococcus aureus Prevalence in European Union/European Economic Area Countries to Governance Predictors

Governance predictors (scores)	n	Univariate		Parsimonious
Governance predictors (scores)	n	Wald χ²	p	Wald χ²	p
Accountability	28	3.160	0.075	—	—
Stability	28	2.382	0.123	—	—
Effectiveness	28	0.036	0.850	—	—
Regulatory	28	2.048	0.152	—	—
Rule of law	28	1.171	0.279	—	—
Control of corruption	28	7.763	0.005	18.450	<0.001

—, Factor rejected by parsimonious model.

Table 5.

Parsimonious Generalized Linear Model Relating Methicillin-Resistant Staphylococcus aureus Prevalence in European Union/European Economic Area Countries to All the Significant Predictors Identified in the Four Individual Models

Predictor	Wald χ²	B	Standard error	95% CI of Wald χ²		p
Predictor	Wald χ²	B	Standard error	Upper	Lower	p
Intercept	28.061	0.355	0.0671	0.224	0.487	<0.001
More than one antimicrobial per patient	11.482	−0.005	0.0014	−0.008	−0.002	0.001
Control of corruption	9.244	0.037	0.0121	0.013	0.061	0.002
Masculinity	9.093	−0.001	0.0003	−0.002	<0.001	0.003
Uncertainty avoidance	10.432	−0.002	0.0005	−0.003	−0.001	0.001

CI, confidence interval.

Discussion

The literature is replete with studies that have linked antibiotic use as a major risk factor for MRSA, including in the Mediterranean region.¹⁶ This fits in well with the findings of our study, where both average antibiotic use per patient and proportion of patients prescribed more than one antimicrobial were correlated with MRSA prevalence. Although the overall data within the 2016/2017 PPS report does not make it possible to establish the exact nature of these antibiotics, the overriding reasons why patients are treated with more than one antibiotic is to access a wider antimicrobial spectrum than would be the case with receiving only one drug. This would in turn suggest that this practice would be a driver, rather than a consequence, of high MRSA prevalence. Indeed, numerous publications have illustrated how antibiotic use drives MRSA by disrupting the resident microbiome and allowing proliferation of methicillin-resistant strains.¹⁷ Our results suggest that it is the duration and type of treatment, at individual patient level, that correlates with MRSA prevalence. This association between prolonged treatment and resistance has already been made for other drug-bug combinations.¹⁸

For more than two decades, hand hygiene has been advocated as a critical intervention to prevent MRSA cross-transmission.¹⁹ Furthermore, AHR consumption has been found to be a reliable indicator of hand hygiene compliance at hospital level.²⁰ It was therefore quite surprising to find no association between MRSA prevalence and AHR consumption in EU/EEA countries. At face value, this would suggest that the impact of hand hygiene on MRSA transmission may not be as high as previously thought. Indeed, a recent review concluded that whereas many published studies have reported a link between improved hand hygiene compliance and reduction in MRSA acquisition and infections, several issues remain unanswered including the temporal relationship between hand hygiene enhancement strategies and decrease in MRSA rates.⁹ There could however be another reason for the lack of association between AHR consumption and MRSA prevalence in our study. The 75th percentile of AHR consumption among EU/EEA countries participating in the 2016/2017 PPS was 37 L/1,000 BD. Assuming a standard 3 mL AHR dispenser, this would be equivalent to only 12 doses per patient per day. If AHR was at least applied before and after the patient contact and taking into account the possibility of some hand washing being additionally done, hand hygiene would have been performed for only ∼10 patient opportunities per day, even in the better performing countries. This is a far cry from the 70 contact opportunities reported by the HOW2 Benchmark Study as happening on a daily basis during health care delivery in an average hospital ward. In turn, this would suggest that hand hygiene performance still remains very much suboptimal throughout EU/EEA, even in the countries reporting higher AHR consumption.

Other than number of practicing nurses, our results suggest that structural factors do not seem to play a major role in the epidemiology of MRSA in the EU/EEA, contrary to previous individual studies.²¹ Understaffing has been well documented to facilitate MRSA acquisition, especially in hospital outbreaks.¹⁰ Nevertheless, the elimination of this parameter from the final regression model would suggest that it may not be highly relevant.

On the contrary, our results continue to evidence the significance of sociocultural factors on the epidemiology of antimicrobial resistance (AMR) within the European region. It is quite telling that cultural dimensions showed some of the highest correlations with MRSA prevalence, both in the individual and final models. At face value, the association with masculinity would seem to be incongruous. Masculine societies tend to be very much focused on outcomes. Indeed, centrally determined targets were deemed to be a critical contributor toward the massively successful reduction of MRSA bacteremia in the United Kingdom.²² This country scores high for masculinity in Hofstede's cultural indices. However it also scores low for UAI. It does not seem coincidental that it is the combination of masculinity and UAI scores that correlates with increased MRSA prevalence.²³ Hofstede emphasizes that uncertainty avoidance is not the same as risk avoidance. Individuals in high UAI societies will only address risk if it fosters a feeling of discomfort and stress. This is rarely the case with inappropriate antibiotic prescribing and noncompliance with IPC; both are health care practices that drive MRSA. Self-perception by health care personnel that their poor practices are leading to health care-associated infections is often absent and unlikely to cause discomfort or stress.²⁴ A “crisis management” approach tends to be the norm in many countries that score high for UAI and masculinity. The constant experience of routine MRSA infections, which is typical of hyper-endemic countries, will not be viewed as a crisis. It therefore tends to go unchallenged and attention is more likely to be directed at more tangible challenges, such as waiting lists. It is not surprising to read the report of the ECDC AMR country visit to Italy stating that the visiting team “often gained the impression that these high levels of AMR appear to be accepted by stakeholders throughout the health care system, as if they were an unavoidable state of affairs.”²⁵ In the Hofstede indices, Italy has one of the highest scores for masculinity and UAI in Europe. The inclusion of power distance in the univariate sociocultural regression model, adds a further element to this hypothesis. Societies that score high in power distance often encounter challenges in fostering ownership, accountability, and the ability to work in multidisciplinary teams; all are critical to an effective IPC and antibiotic stewardship program.²⁶

It is something of a struggle to associate corruption with the delivery of health care in a group of countries brought together by shared values of dignity, equality, and rule of law. Yet the repeated association between this index of poor governance and AMR refuses to go away.¹⁵ However, it is difficult to believe that nurses in high MRSA prevalence countries expect bribes to perform hand hygiene or that physicians prescribe antibiotics unnecessarily to gain favors from the pharmaceutical industry.²⁷ We posit an alternative hypothesis. Countries where control of corruption is lax are characterized by a general tolerance of deviant social norms, which go unchecked and unchallenged by the mainstream.²⁸ That same normalization of deviance will apply throughout society, including in hospitals and in relation to policies and guidelines aimed at controlling multi-drug resistant organisms (MDRO). In other words, although most health care professionals would be well aware of the importance of practices such as hand hygiene or contact precautions for the prevention of health care-associated infections, non-compliance becomes a deviance that is accepted and rarely challenged, neither by peers nor by management.

The main limitation of the study was that the dependent variable comprised solely of the MRSA proportions of 28 countries; therefore, fitting a GLM could have caused severe overfitting. It was for this reason that it was decided to first analyze the predictors in a univariate manner. Those predictors that were not significant were automatically excluded from the final analysis. Data usage of the accessed European reports might cause gaps for some process and structural factors. For example, although IPC staffing and isolation room capacities are valuable information, they say little about the infection prevention policies on MRSA adopted by the different European countries. The outcome indicator depends on the frequency of blood testing and the selection of the included results within the countries participating in EARS-Net; this may not always be a random sample or complete survey of all blood cultures. Similar biases may act on the predictors for the antimicrobial consumption and AHR, as collected in the PPS, which also constitutes a sample of EU/EEA hospitals and, owing to its design, over-selects patients with long hospital stays. Yet they are, at the moment, the “best available” data sources in the EU/EAA.

Our results are important because they continue to highlight the relevance of behavioral drivers to the MDRO situation within EU/EEA countries. Indeed, the only significantly associated factor not related to culture and governance was the prescribing of more than one antibiotic per patient. Even here, it should be commented that the only explanation as yet provided for a similar variability in prescribing broad-spectrum antibiotics in community practice across Europe is cultural and again related to UAI.²⁹ It does not take much of an extrapolation to suggest that uncertainty avoidance could also be a driver for broad-spectrum antibiotic prescribing within hospitals.

The past decade has seen concerted campaigns to reduce AMR through improved hand hygiene and antibiotic stewardship, including the World Health Organization's Clean Care is Safer Care and the European Commission's annual European Antibiotic Awareness Day activities.^30,31 Yet these initiatives do not seem to have achieved the desired impact on MRSA in high-prevalence European countries. “More of the same” is unlikely to result in any significant change in the status quo. An alternative approach appears urgently needed; one that is better informed by behavioral sciences and, above all, is compatible with the cultural and governance realities in these countries.

Footnotes

Acknowledgments

The assistance of Dr. C. Seutens (ECDC) in the preparation of this manuscript is gratefully acknowledged. PPS statistics were established by national coordinating centres and hospitals in the respective countries.

Disclosure Statement

No competing financial interests exist.

Funding Information

This study was supported by internal funding.

References

Cassini

, Högberg

L.D

, Plachouras

, Quattrocchi

, Hoxha

, Simonsen

G.S

, Colomb-Cotinat

, Kretzschmar

M.E

, Devleesschauwer

, Cecchini

, Ouakrim

D.A

, Oliveira

T.C

, Struelens

M.J

, Suetens

, and Monnet

D.L.

. 2019. Attributable deaths and disability-adjusted life-years caused by infections with antibiotic-resistant bacteria in the EU and the European Economic Area in 2015: a population-level modelling analysis. Lancet Infect. Dis. 19:56–66.

Stefani

, and Varaldo

P.E

. Epidemiology of methicillin-resistant staphylococci in Europe 2003. Clin. Microbiol. Infect. 9:1179–1186.

European Centre for Disease Prevention and Control. 2018. Surveillance of antimicrobial resistance in Europe—Annual report of the European Antimicrobial Resistance Surveillance Network (EARS-Net) 2017. ECDC, Stockholm.

Harbarth

. 2006. Control of endemic methicillin-resistant Staphylococcus aureus—recent advances and future challenges. Clin. Microbiol. Infect. 12:1154–1162.

Gaygısız, Ü., Lajunen

, and Gaygısız

. 2017. Socio-economic factors, cultural values, national personality and antibiotics use: a cross-cultural study among European countries. J. Infect. Public Health, 10:755–760.

European Centre for Disease Prevention and Control. 2019. Surveillance of antimicrobial resistance in Europe 2018. ECDC, Stockholm.

European Centre for Disease Prevention and Control. 2019. Point prevalence survey of healthcare-associated infections and antimicrobial use in European acute care hospitals, 2016–2017. ECDC, Stockholm.

Dancer, S.J. 2008. The effect of antibiotics on methicillin-resistant Staphylococcus aureus. J. Antimicrob. Chemother. 61:246–253.

Marimuthu

, Pittet

, and Harbarth

. 2014. The effect of improved hand hygiene on nosocomial MRSA control. Antimicrob. Resist. Infect. Control, 3:34.

10.

Clements

, Halton

, Graves

, Pettitt

, Morton

, Looke

, Whitby

. 2008. Overcrowding and understaffing in modern health-care systems: key determinants in methicillin-resistant Staphylococcus aureus transmission. Lancet Infect. Dis. 8:427–434.

11.

OECD/EU. 2018. Health at a glance: Europe 2018: state of health in the EU cycle. OECD Publishing, Paris.

12.

Garvey

M.I.

, Bradley

C.W

, and Wilkinson

M.A.C

. 2019. The value of the infection prevention and control nurse led MRSA ward round. Antimicrob. Resist. Infect. Control, 8:53.

13.

Borg, M.A. 2014. Lowbury lecture 2013. Cultural determinants of infection control behaviour: understanding drivers and implementing effective change. J. Hosp. Infect. 86:161–168.

14.

Hofstede

, Hofstede

G.J

, and Minkov

. 2010. Cultures and organisations: software of the mind. McGraw-Hill, New York.

15.

Collignon

, Athukorala

P.C

, Senanayake

, and Khan

. 2015. Antimicrobial resistance: the major contribution of poor governance and corruption to this growing problem. PLoS One, 10:e0116746.

16.

Borg

M.A.

, Zarb

, Scicluna

E.A

, Rasslan

, Gür

, Ben Redjeb

, Elnasser

, and Daoud

. 2010. Antibiotic consumption as a driver for resistance in Staphylococcus aureus and Escherichia coli within a developing region. Am. J. Infect. Control, 38:212–216.

17.

Tacconelli

, De Angelis

, Cataldo

M.A

, Pozzi

, and Cauda

. 2008. Does antibiotic exposure increase the risk of methicillin-resistant Staphylococcus aureus (MRSA) isolation? A systematic review and meta-analysis. J. Antimicrob. Chemother. 61:26–38.

18.

Ruhe

J.J.

, and Hasbun

. 2003. Streptococcus pneumoniae bacteremia: duration of previous antibiotic use and association with penicillin resistance. Clin. Infect. Dis. 36:1132–1138.

19.

Pittet

, Hugonnet

, Harbarth

, Mourouga

, Sauvan

, Touveneau

, and Perneger

T.V.

. 2000. Effectiveness of a hospital-wide programme to improve compliance with hand hygiene. Lancet, 356:1307–1312.

20.

Borg

M.A.

, and Brincat

. 2018. Addressing the controversy of 100% hand hygiene compliance: can alcohol rub consumption data serve as a useful proxy validator? J. Hosp. Infect. 100:218–219.

21.

Borg, M.A. 2010. Are healthcare economics a factor behind European MRSA rates? Eur. J. Clin. Microbiol. Infect. Dis. 29:477–479.

22.

Duerden

, Fry

, Johnson

A.P

, and Wilcox

M.H.

. 2015. The control of methicillin-resistant Staphylococcus aureus blood stream infections in England. Open Forum Infect. Dis. 2:ofv035.

23.

Borg

M.A.

, Camilleri

, and Waisfisz

. 2012. Understanding the epidemiology of MRSA in Europe: do we need to think outside the box? J. Hosp. Infect. 81:251–256.

24.

Nicol

P.W.

, Watkins

R.E

, Donovan

R.J

, Wynaden

, and Cadwallader

. 2009. The power of vivid experience in hand hygiene compliance. J. Hosp. Infect. 72:36–42.

25.

European Centre for Disease Prevention and Control. 2017. ECDC country visit to Italy to discuss antimicrobial resistance issues. ECDC, Stockholm.

26.

Gould

D.J.

, Hale

, Waters

, and Allen

. 2016. Promoting health workers' ownership of infection prevention and control: using Normalization Process Theory as an interpretive framework. J. Hosp. Infect. 94:373–380.

27.

Rönnerstrand

, and Lapuente

. 2017. Corruption and use of antibiotics in regions of Europe. Health Policy, 121:250–256.

28.

Takács

, Csapodi

, and Takács-Györg

. 2011. Corruption as a deviant social attitude. Publ. Financ. Quart. 56:27–43.

29.

Borg

M.A.

, and Camilleri

. 2019. Broad-spectrum antibiotic use in Europe: more evidence of cultural influences on prescribing behaviour. J. Antimicrob. Chemother. 74:3379–3383.

30.

Pittet

, Allegranzi

, and Boyce; World Health Organization World Alliance for Patient Safety First Global Patient Safety Challenge Core Group of Experts

. 2009. The World Health Organization Guidelines on Hand Hygiene in Health Care and their consensus recommendations. Infect. Control Hosp. Epidemiol. 30:611–622.

31.

Earnshaw

, Monnet

D.L

, Duncan

, O'Toole

, Ekdahl

, and Goossens; European Antibiotic Awareness Day Technical Advisory Committee; European Antibiotic Awareness Day Collaborative Group

. 2009. European Antibiotic Awareness Day, 2008—the first Europe-wide public information campaign on prudent antibiotic use: methods and survey of activities in participating countries. Euro Surveill. 14:19280.