Introduction. N. J. Chaves, 1 A. C. Cheng, 2 N. Runnegar, 3 J. Kirschner, 4 T. Lee 4 and K. Buising 1. Abstract

Transcription

1 Analysis of knowledge and attitude surveys to identify barriers and enablers of appropriate antimicrobial prescribing in three Australian tertiary hospitals N. J. Chaves, 1 A. C. Cheng, 2 N. Runnegar, 3 J. Kirschner, 4 T. Lee 4 and K. Buising 1 1 Victorian Infectious Diseases Service, Melbourne Health, 2 Department of Epidemiology and Preventive Medicine, Monash University, 4 Pharmacy, Alfred Health, Melbourne, Victoria and 3 Infection Management Services, Princess Alexandra Hospital, Brisbane, Queensland, Australia Key words anti-infective agent, attitude of health personnel, attitude of health personnel, clinical competence, inappropriate prescribing, questionnaire. Correspondence Nadia Chaves, Victorian Infectious Diseases Service, 9 North, Royal Melbourne Hospital, Grattan Street, Parkville, Vic. 3052, Australia. nadia.chaves@mh.org.au Received 10 November 2013; accepted 20 January doi: /imj Abstract Background: Antimicrobial stewardship programmes aim to optimise use of antibiotics and are now mandatory in all Australian hospitals. Aim: We aimed to identify barriers to and enablers of appropriate antimicrobial prescribing among hospital doctors. Methods: Two paper-based and one web-based surveys were administered at three Australian university teaching hospitals from March 2010 to May The 18-item questionnaire recorded doctors level of experience, their knowledge regarding the use of common antimicrobials and their attitudes regarding antimicrobial prescribing. Local survey modifications allowed inclusion of specific questions on: infections in intensive care unit patients, clinical microbiology and use of local guidelines. Results: The respondents (n = 272) were comprised of 96 (35%) registrars, 67 (25%) residents, 57 (21%) interns and 47 (17%) consultant hospital doctors. Forty-one per cent were working in a medical specialty. Identified barriers included: gaps in antimicrobial prescribing knowledge (especially among interns), a lack of awareness about which antimicrobials were restricted and a reliance on senior colleagues to make antimicrobial prescribing decisions. Enablers of optimal prescribing included: an acknowledgement of the need for assistance in prescribing and reported readiness to consult national prescribing guidelines. These results were used to help guide and prioritise interventions to improve prescribing practices. Conclusion: A transferable knowledge and attitudes survey tool can be used to highlight barriers and facilitators to optimal hospital antimicrobial prescribing in order to inform tailored antimicrobial stewardship interventions. Introduction Antimicrobial resistance is an increasing problem worldwide. A strategic global response is required to preserve current antibiotics and to reduce development of resistance. 1 3 Antimicrobial stewardship programmes aim Funding: N. J. Chaves is a recipient of a National Health and Medical Research Council Translating Research into Practice Fellowship. Conflict of interest: A. C. Cheng, N. Runnegar, J. Kirschner, T. Lee none. N. J. Chaves, K. Buising a web-based decision support system to assist antimicrobial stewardship is a product owned by Melbourne Health. The authors are employed by Melbourne Health but declare no other direct personal financial interest in the system. The funding bodies had no influence on the study design, data collection and analysis or interpretation and writing of this manuscript. The authors declare no competing interest with regard to the material published in this article. to optimise antimicrobial use and are now required as part of accreditation of many hospitals worldwide, including Australia. 4 Effective antimicrobial stewardship interventions include audit and feedback, education, academic detailing, antimicrobial restriction and approval and computerised decision support. 5,6 However, the optimal model of antimicrobial stewardship is likely to differ depending on the context, and a challenge for all institutions is to find an antimicrobial stewardship model that fits the prescribing culture of that particular institution to best facilitate sustainable change. 7,8 Interventions that are tailored to address specifically identified barriers are more likely to change behaviour and to improve professional practice. 9 Prompted by the experience at an initial centre, 10 three hospitals adapted a survey tool to identify particular barriers to and enablers of appropriate antimicrobial prescribing. We present the results of surveys performed at these three institutions and their 568

2 Antimicrobial KAP survey three hospitals approaches to addressing the barriers and enablers identified. We aimed to assess the knowledge and explore the attitudes of Australian hospital doctors to antibiotic use in order to identify barriers to and enablers of appropriate antimicrobial prescribing. Methods Setting This study involved three Australian tertiary adult hospitals, with beds each, in Melbourne and Brisbane. Each hospital had infectious diseases (ID) registrars who provided telephone and in-person consultations as required for complex patients. The hospitals had formularies and policies that required approval from the infectious diseases service for the use of nominated restricted, mainly broad-spectrum antimicrobials. All hospitals had some local prescribing guidelines as well as access to the Therapeutic Guidelines: Antibiotic, 11 a nationally available antibiotic prescribing guideline available through the intranet. At the time of the survey, there were no antimicrobial stewardship teams providing post-prescription review. All three hospitals provided antimicrobial education in the form of a lecture at intern orientation. Hospital A wards, but not their intensive care unit (ICU), had a web-based decision support system at the time of the survey. Survey design and administration This study was approved as a quality assurance project by the ethics committees at all three hospitals as it was low risk as per National Health and Medical Research Council guidelines 12 and, therefore, did not require full ethics approval. The survey had been developed by a multidisciplinary team of infectious diseases and microbiology doctors and a pharmacist and had been piloted at another site. 10 The original questionnaire had 18 questions: three on participant level of experience and current role, 10 multiple choice questions regarding knowledge about appropriate antimicrobial use and five Likert-type questions assessing attitudes to antimicrobial stewardship. There were several additional free text boxes for comments. All questions testing knowledge about antimicrobial use reflected the recommendations in the Therapeutic Guidelines: Antibiotic (Appendix). Six of the 10 multiplechoice questions on antimicrobial knowledge were common across all three hospitals. Investigators at individual institutions modified some questions to identify specific barriers to and enablers of appropriate antimicrobial prescribing of particular interest at their site. At hospital A, ICU and ward doctors received different surveys. Examples of modifications included: questions on empiric meropenem use for ICU doctors, in an effort to identify potential knowledge gaps that might explain high levels of meropenem use in that department (hospital A); questions on Gram s stain characteristics of common bacterial pathogens in response to observed knowledge gaps among junior doctors (hospital B); and, questions on attitudes towards specific antimicrobial stewardship interventions prior to planned implementation of a web-based decision support and approvals programme (hospital C). Survey administration varied by institution, but all surveys were anonymous and voluntary. Participants were doctors employed at the three hospitals. Doctors in Australian hospitals are defined by their level of training as follows: interns first year post-medical school, resident years two to three post-medical school, registrar year four and above, and usually enrolled in a specialty training programme (medical or surgical), consultants vocationally registered and have completed specialty training. Participants self-identified and were purposively sampled at intern, resident and medical registrar education sessions (hospital A, B and C), ICU meetings (hospital A) and grand rounds (hospital B). Hospital A and B s paper-based surveys were distributed at the start or end of meetings, were generally completed within 10 min and were collected immediately without time to refer to other resources. Hospital C s medical staff received the survey through , followed by one reminder , and had a movie ticket provided as an incentive on completion. All ed surveys were completed within 1 month of initial distribution. Statistical analysis All three hospitals surveys were administered between March 2010 and 13 May All analyses were conducted in Stata version 12 (StataCorp, College Station, TX, USA). Differences in responses between groups (doctors levels of experience) were summarised using mean and standard deviation (SD) and were analysed using a non-parametric Kruskal Wallis (KW). Pairwise differences between experience levels were analysed using a post-hoc Dunn test with Bonferroni correction for multiple simultaneous comparisons. Where a Bonferroni deflation of the P-value significance cut-off was not required, P < 0.05 was considered significant. Likert-type items were analysed using a 5-point scale. 13 A barrier was defined as process or situation that would potentially impede or obstruct optimal antimicrobial prescribing, whereas an enabler or facilitator was a process or situation which would allow optimisation of antimicrobial therapy. These definitions were established through author consensus. 569

3 Chaves et al. Table 1 Survey participants level of experience Level of experience Number of doctors Hospital name Total (%) A B C Not specified (2) Intern (21) Resident (25) Registrar (35) Consultant (17) Total Question number a Results Surveys were submitted by 313 doctors in three hospitals, but 41 (14%) were excluded as they had submitted blank surveys. Results were therefore analysed on 272 surveys (Table 1). Exact participation rates were unavailable, but the opportunistic sampling at hospital A and B captured more than 90% of those attending the education sessions and lunchtime meetings. Hospital A s survey captured 80% of their intensive care consultants, 30% of their medical registrars and 15% of their interns. Hospital B s survey captured 50% of their medical registrars and 30% of their interns. Hospital C s ed survey estimated a 30% response rate, and also captured 30% of their registrars. Forty-five per cent of all participants were from a medical specialty. Differences between experience levels in knowledge and attitudes to antimicrobial prescribing Regarding common knowledge questions, registrars were found to have the highest number of correct responses, followed by consultants, then residents and interns (Fig. 1). There was a significant difference between groups (KW P = 0.001), with consultants, registrars and residents scoring significantly higher than interns (P < 0.001), and registrars scoring significantly higher than residents (P < 0.001). Consultants did not score significantly differently to residents (P = 0.19). With respect to attitudes questions, interns reported requiring a significantly higher level of assistance in choosing antibiotics (mean 4.0 (SD 0.96)) than the other levels of experience (mean = 3.2 (SD 0.1)) (P < 0.001) (Fig. 2). Identified barriers to and enablers of appropriate antimicrobial prescribing Results of Likert-scored attitudes questions are displayed in Figure 3. Enablers identified included: the consultation Interns = 57; Residents = 67; Registrars = 96; Consultants = 47 of pharmacists and infectious diseases units, the use of Therapeutic Guidelines and the belief that prudent antimicrobial use will result in reduced resistance. s included: a poor awareness of local unit and hospital guidelines, prescribers belief that they were aware of the management of most common conditions (despite suboptimal knowledge scores) and a lack of awareness of which antimicrobials were restricted. Results of specific hospitals Hospital A Percentage correct Figure 1 Percentage of correct answers to knowledge questions asked at all three hospitals by level of experience. ( ), Intern; ( ), Resident; ( ), Registrar; ( ), Consultant. Consultant Registrar Resident Intern 0% 20% 40% 60% 80% 100% Figure 2 I often feel I need assistance to choose an appropriate antibiotic and dose for a specific clinical situation. ( ), Strongly disagree; ( ), disagree; ( ), equivocal; ( ), agree; ( ), strongly agree. Of the 74 staff surveyed at hospital A, 27 worked in the ICU, and this group was asked to complete a specifically modified questionnaire (Appendix). The question on 570

4 Antimicrobial KAP survey three hospitals Figure 3 Survey results of 5-point Likertscale attitudes questions. ( ), Strongly disagree; ( ), disagree; ( ), equivocal; ( ), agree; ( ), strongly agree. A, B, C, hospital names; A-ICU, hospital A intensive care unit. appropriate indications for empiric prescription of meropenem was answered correctly by 8/27 (30%) of surveyed ICU staff. Four per cent (1/27) chose the duration of antibiotic therapy for uncomplicated ventilatorassociated pneumonia that was consistent with the recommendations in the National Guidelines. Twelve out of 27 ICU doctors (44%) reported that they were aware of which antimicrobials were restricted. One hundred per cent of ICU medical staff agreed or strongly agreed that prudent antimicrobial use was important to reduce resistance in their ICU. Hospital B Hospital B interns, registrars and consultants completed a table describing the Gram s stain (Gram-positive or negative, bacilli or cocci) of common organisms (Staphylococcus aureus, Eschericia coli, Enterococcus, Pseudomonas, Listeria, Neisseria meningitidis, Enterobacter). Consultants had significantly higher Gram s stain scores compared to other levels of experience (mean 5.4 (SD 1.5) versus nonconsultants mean 4.0 (SD 1.1) (P = 0.001)). Forty-one per cent of respondents (27/66) reported that they were aware of which antimicrobials were restricted. Hospital C At hospital C, 61% (72/117) of respondents answered because my consultant asked me to when asked about reasons for inappropriately commencing broad-spectrum antimicrobials or for not de-escalating antimicrobial therapy. The second most common answer in 48% (more than one option allowed) was I am worried about missing something if I chose a narrow agent. Seventy-three per cent of respondents reported that they were aware of current national guideline recommendations for the management of common clinical conditions. Overall 53% were equivocal or disagreed with the comment I find contacting ID to be easy and effective. One-third of respondents reported that they thought web-based approval would be a better process (than the current ID registrar approvals system). Fortynine per cent responded it would depend on how easy the system is to use. Qualitative responses in survey Although this survey was primarily quantitative, the free text boxes in hospital C s survey allowed participants to express views in their own words. Table 2 contains 571

5 Chaves et al. Table 2 Quotations from hospital C s free text boxes illustrating themes Theme Support for antibiotic restriction Fear of consequences (of antibiotic restriction through web-based support or ID registrar) Communication difficulties and time constraints (in trying to contact ID for approvals) Quote Absolutely necessary so we don t end up like the US. I think this is good because it...protects the world from resistant organisms bred by our laziness. Imposing any further restrictions may cause detriment to patients [who require] broad spectrum antibiotics. If there is difficulty in obtaining approval (e.g. system crash, ID reg busy) that this should not hold up the delivery [of the antibiotic] to the patient. Valuable learning opportunities are embedded in the process [ID phone referrals], hopefully this won t be lost [with the web-based system]. As a junior doctor the lack of time in the day is the restricting part of this process [of obtaining approvals]. They are often very busy and take some time getting back to you. My interns often have to spend 15 min on the phone just to get an antibiotic approval. Depends on the [ID] registrar but I generally find this to be a very difficult and time consuming task. Almost always needs multiple pages over hours. I feel bad for having to contact [the ID registrar] for things which are obvious. [Text] is added by the author for clarification. We have made every effort not to alter the original intent of these statements. ID, infectious diseases. quotations illustrating identified themes. Three main themes emerged from hospital C s free text box comments: fear of consequences (of inappropriate antibiotic prescribing and of a web-based decision support programme), time management issues (with obtaining approval for restricted antimicrobials) and general support for antibiotic restriction. Analysis of results using a theoretical framework An understanding of strategies to address identified barriers and enablers can be achieved in part through application of behaviour change theory. A recent consensus document formulated the theoretical domains framework (TDF) several key themes relevant to changing healthcare behaviour. 14 These include: knowledge, skills, social/professional role and identity, beliefs about capabilities, optimism, beliefs about consequences, intentions, goals, memory, attention and decision processes, reinforcement, environmental context and resources, social influences, emotion and behavioural regulation. The TDF has been used to identify barrier and enablers and inform behaviour change strategies in several different healthcare disciplines, including medication safety 15 and hand hygiene. 16 Although we did not design the survey with prior knowledge of the TDF, we found that this formed a useful framework to understand our results, to inform targeted interventions and to feed back the results to medical staff, key groups and executive. A variety of interventions were proposed as a result of the feedback in each hospital (Table 3). Discussion This study reports on the results of three similar surveys that aimed to explore barriers to and enablers of appropriate hospital antimicrobial prescribing in order to inform antimicrobial stewardship interventions. Identified barriers include: gaps in antimicrobial prescribing knowledge (especially among interns), a lack of awareness about which antimicrobials were restricted and a reliance on senior colleagues to make antimicrobial prescribing decisions. Enablers of hospital prescribing include: an acknowledgement of the need for assistance in prescribing and reported readiness to consult national prescribing guidelines. Our findings support the results of other antimicrobial knowledge and attitude surveys. Other studies of hospital doctors have identified barriers of antimicrobial prescribing, such as antimicrobial prescribing knowledge gaps, the influence of senior prescribers on junior doctors antimicrobial choices and the utilisation of national over local guidelines. 23 Identified enablers included a support of antimicrobial restriction 24 and an awareness of the threat of antimicrobial resistance. 18,25 All studies suggested that knowledge and attitude surveys could improve antimicrobial stewardship programme implementation but did not necessarily specify what programmes should be implemented. This study was designed to be a simple modifiable quantitative survey, which meant it was by necessity quite short and did not explore topics in great detail. Of note, one of the TDF domains which was not identified on the survey, emotion was clearly relevant when analysing the free text box comments. Development of the survey 572

6 Antimicrobial KAP survey three hospitals Table 3 Survey results categorised into behaviour change domains, barriers or enablers and proposed interventions Themes Survey results (hospital identifier) or enabler Proposed interventions Knowledge Knowledge-gap especially among junior staff (A, B, C) Increase junior staff education Clinical microbiology knowledge poor (B) Include microbiology training at medical school Appropriate indications for empiric meropenem use unclear (A-ICU) Clarify meropenem indications; develop guideline with ICU Skills Vancomycin dose incorrectly adjusted (C) Pharmacist to dose vancomycin to reduce errors Social professional role/identity Junior doctors influenced by consultants when choosing antibiotics (C) AS team to provide consultant to consultant advice, education and feedback Beliefs about capability Doctors need assistance choosing antimicrobials (A, B, C) Enabler Web-based decision support and AS teams to provide assistance to prescriber Familiar antimicrobials in use more often (A) Enabler Develop lanyards with traffic light indications to improve familiarity Optimism (or pessimism) A web-based approval system is time consuming (A-ICU) Further stakeholder engagement was required before implementing Beliefs about consequences Intentions Environmental context and resources Antimicrobials often continued because doctors worried about missing something (C) Belief that prudent use of antimicrobials will reduce resistance (A-ICU) Difficult to know which antibiotics are restricted, so I just wait for a pharmacist to tell me (A) Obtaining approvals through ID registrar is time consuming (C) A web-based system needs to be easy to use (C) Enabler Point-of-care audit and feedback to doctors by AS team to support timely antimicrobial de-escalation Identify and engage local ICU champions of prudent antimicrobial use Incentives (chocolate) provided to improve intentions Ensure web-based decision support is time efficient and easy to use Therapeutic guidelines used regularly Enabler Use therapeutic guidelines for any antimicrobial stewardship programme Intranet guidelines difficult to find Increase accessibility through home page links Behavioural regulation Most doctors did not dislike the idea of web-based approval Enabler Implement web-based approval A, B, C, hospital names; A-ICU, hospital A intensive care unit; AS, antimicrobial stewardship; ID, infectious diseases. questions through a priori use of the TDF, and more detailed exploration of themes through focus groups and qualitative interviews would improve the validity of results. 15,26,27 Moreover, modification of the survey at each site meant that it was difficult to compare the results between institutions. The low moderate response rate may have resulted in biases, and we have no detailed information on non-respondents. Additionally, not all respondents answered all questions, and we suspect that some respondents chose not to respond rather than respond incorrectly. Nevertheless, this survey did provide information that could be used to help guide interventions. Conclusion All hospitals worldwide currently face significant problems with inappropriate antimicrobial use and the growing problem of antimicrobial resistance among pathogens. Identification of barriers to and facilitators of appropriate antimicrobial use can be part of an evidencebased strategy for developing effective antimicrobial stewardship programmes. 9,28,29 We have used a modifiable transferable survey tool in three Australian tertiary hospitals to identify locally relevant barriers and enablers with which to develop tailored antimicrobial stewardship programmes. Acknowledgements We thank Dr Saliya Hewagama (Alice Springs Hospital) and Mr Osbert Cotta (Melbourne Health) for use of their survey tool and Dr Tim Spelman (Victorian Infectious Diseases Service) for his statistical assistance. References 1 Holmes AH, Sharland M. The Chennai Declaration: India s landmark national commitment to antibiotic stewardship demonstrates that truth alone triumphs. J Antimicrob Chemother 2013; 50: Hayashi Y, Davis L, Paterson DL. Why can t I prescribe that antibiotic? The role of antimicrobial stewardship programmes in modern medicine. Intern Med J 2009; 39: Tacconelli E. Antimicrobial use: risk driver of multidrug resistant 573

7 Chaves et al. microorganisms in healthcare settings. Curr Opin Infect Dis 2009; 22: Australian Commission on Safety and Quality in Health Care. Hospital Accreditation Workbook. Sydney: ACSQHC; Dellit TH, Owens RC, McGowan JE Jr, Gerding DN, Weinstein RA, Burke JP et al. Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America guidelines for developing an institutional program to enhance antimicrobial stewardship. Clin Infect Dis 2007; 44: Davey P, Brown E, Fenelon L, Gould IM, Holmes A, Ramsay CR et al. Interventions to improve antibiotic prescribing practices for hospital inpatients. Cochrane Database Syst Rev 2005; CD Hulscher ME, Grol RP, van der Meer JW. Antibiotic prescribing in hospitals: a social and behavioural scientific approach. Lancet Infect Dis 2010; 10: Duguid M, Cruickshank M. Antimicrobial Stewardship in Australian Hospitals. Sydney: ACSQHC; Shaw B, Cheater F, Baker R, Gillies C, Hearnshaw H, Flottorp S et al. Tailored interventions to overcome identified barriers to change: effects on professional practice and health care outcomes. Cochrane Database Syst Rev 2005; CD Hewagama S, Cotta O. What do junior doctors know about antibiotics? In: Australasian Society of Infectious Diseases, ed. Annual Scientific Meeting of Australasian Society of Infectious Diseases. Darwin (NT, Australia): Australasian Society of Infectious Diseases; 2010; Antibiotic Expert Group. Therapeutic guidelines: antibiotic. Melbourne: Therapeutic Guidelines Limited; The National Statement on Ethical Conduct in Human Research 2007 (Updated December 2013). The National Health and Medical Research Council, the Australian Research Council and the Australian Vice-Chancellors Committee. Commonwealth of Australia, Canberra 2013 [cited 2013 Dec 29]. Available Supporting information from URL: _files_nhmrc/publications/attachments/ e72_national_statement_ pdf 13 Boone H, Boone D. Analyzing Likert data. JOE 2012; 50: 2. [cited 2013 Dec 29]. Available from URL: 14 Cane J, Connor D, Michie S. Validation of the theoretical domains framework for use in behaviour change and implementation research. Implement Sci 2012; 7: Duncan EM, Francis JJ, Johnston M, Davey P, Maxwell S, McKay GA et al. Learning curves, taking instructions, and patient safety: using a theoretical domains framework in an interview study to investigate prescribing errors among trainee doctors. Implement Sci 2012; 7: Boscart VM, Fernie GR, Lee JH, Jaglal SB. Using psychological theory to inform methods to optimize the implementation of a hand hygiene intervention. Implement Sci 2012; 7: Abbo L, Sinkowitz-Cochran R, Smith L, Ariza-Heredia E, Gomez-Marin O, Srinivasan A et al. Faculty and resident physicians attitudes, perceptions, and knowledge about antimicrobial use and resistance. Infect Control Hosp Epidemiol 2011; 32: Srinivasan A, Song X, Richards A, Sinkowitz-Cochran R, Cardo D, Rand C. A survey of knowledge, attitudes, and beliefs of house staff physicians from various specialties concerning antimicrobial use and resistance. Arch Intern Med 2004; 164: Lucet JC, Nicolas-Chanoine MH, Roy C, Diamantis S, Papy E, Riveros-Palacios O et al. Antibiotic use: knowledge and perceptions in two university hospitals. J Antimicrob Chemother 2011; 66: Mol PG, Rutten WJ, Gans RO, Degener JE, Haaijer-Ruskamp FM. Adherence barriers to antimicrobial treatment guidelines in teaching hospital, the Netherlands. Emerg Infect Dis 2004; 10: Charani E, Castro-Sanchez E, Sevdalis N, Kyratsis Y, Drumright L, Shah N et al. Understanding the determinants of antimicrobial prescribing within hospitals: the role of prescribing etiquette. Clin Infect Dis 2013; 57: De Souza V, MacFarlane A, Murphy AW, Hanahoe B, Barber A, Cormican M. A qualitative study of factors influencing antimicrobial prescribing by nonconsultant hospital doctors. J Antimicrob Chemother 2006; 58: Ali MH, Kalima P, Maxwell SR. Failure to implement hospital antimicrobial prescribing guidelines: a comparison of two UK academic centres. J Antimicrob Chemother 2006; 57: Bannan A, Buono E, McLaws ML, Gottlieb T. A survey of medical staff attitudes to an antibiotic approval and stewardship programme. Intern Med J 2009; 39: Pulcini C, Williams F, Molinari N, Davey P, Nathwani D. Junior doctors knowledge and perceptions of antibiotic resistance and prescribing: a survey in France and Scotland. Clin Microbiol Infect 2011; 17: Patel SJ, Salman L, Duchon JM, Evans D, Ferng YH, Larson E. Development of an antimicrobial stewardship intervention using a model of actionable feedback. Interdiscip Perspect Infect Dis 2012; 2012: French SD, Green SE, O Connor DA, McKenzie JE, Francis JJ, Michie S et al. Developing theory-informed behaviour change interventions to implement evidence into practice: a systematic approach using the Theoretical Domains Framework. Implement Sci 2012; 7: Thakkar K, Gilchrist M, Dickinson E, Benn J, Franklin BD, Jacklin A et al. A quality improvement programme to increase compliance with an anti-infective prescribing policy. J Antimicrob Chemother 2011; 66: Elligsen M, Walker SA, Simor A, Daneman N. Prospective audit and feedback of antimicrobial stewardship in critical care: program implementation, experience, and challenges. Can J Hosp Pharm 2012; 65: Additional Supporting Information may be found in the online version of this article at the publisher s web-site: Appendix S1 Survey questions used in the three hospitals. 574