Skip to main content

Prevention of Hospital Infections by Intervention and Training

Final Report Summary - PROHIBIT (Prevention of Hospital Infections by Intervention and Training)

Executive Summary:
Healthcare-associated infections (HAIs) are the most frequent adverse events in healthcare delivery. They are associated with morbidity and mortality, longer hospital stay, and increase costs. The four major types are catheter-related bloodstream infection (CRBSI), surgical site infection (SSI), ventilator-associated pneumonia (VAP), and catheter-related urinary tract infection (CAUTI). They share one characteristic: they are preventable. A growing body of evidence suggests that so-called “best practice” procedures reduce HAI. In parallel, there has been a proliferation of clinical practice guidelines in the field of infection prevention and control (IPC). However, the scope and quality of such documents remain unknown, including how consistently they are actually used in clinical practice. Given the increasing evidence base, it is often assumed that hospitals adopt practices and procedures with proven effectiveness, but no effort to explore the implementation of IPC practices has been undertaken on a large scale and across different sociocultural backgrounds. In addition, it is not known whether it is possible to implement a multimodal infection prevention strategy in culturally diverse countries, such as in Europe. CRBSI is a prototype of causality and prevention of HAI and represents an ideal scenario to study the effectiveness of such a programme on a European level. Emphasis should be given to the role of hand hygiene in this context since hand hygiene has been widely promoted in Europe (and throughout the world) over the past two decades.
The aim of the Prevention of hospital infection by intervention and training (PROHIBIT) project was to provide a global perspective of IPC activities in Europe on several levels, i.e. availability of published guidelines and recommendations, management and organisation of infection control, the capacity of hospitals to implement a multimodal intervention programme, and identification of barriers and facilitators in implementing IPC programmes in European hospitals. Four scientific work packages addressed these topics using a mixed-methods approach combining a systematic review, a large-scale survey, state-of-the art qualitative research, and a randomised controlled trial.
Guidelines serve as a reference for good practice. The challenge in Europe is language and the many different healthcare systems. Variation in scope, update, and quality of the many documents analysed in PROHIBIT was not unexpected. Some websites were not particularly user-friendly and the documents proved difficult to retrieve, which we assume works against accessibility and the wide dissemination of recommendations. To provide the same level of evidence for all patients in Europe, it is necessary to ensure that the offered documents are of good quality, regularly updated, and an accurate reflection of scientific evidence. It is no longer acceptable that healthcare professionals have to rely on outdated or incomplete documents solely because English is not their first language. Access to updated state-of-the-art guidelines in their own language is the sine qua non of improvement in quality of care.
Twenty-four countries provided data from more than 300 acute care hospitals related to the organisation of IPC and the implementation of evidence-based best practice procedures. Hospitals employed a median of four IPC nurses and one doctor per 1000 beds. When compared to previous surveys, the staffing level has not changed over the past years and remains at the level of the seminal United States (US) study on the efficacy of nosocomial infection control (SENIC) conducted 30 years ago. However, current expert opinion suggests that the SENIC staffing key is outdated and the per-bed rate of IPC professionals should be approximately one per 100 beds in acute care and one per 150-250 beds in long-term care. Two-thirds of the hospitals have introduced a link-nurse system. The vast majority (91%) have established an interdisciplinary IPC committee to define infection control objectives, predominantly addressing hand hygiene compliance, reduction of one or more HAIs and, to a lesser extent, antibiotic stewardship. Most hospitals offer training for hand hygiene. Education addressing other IPC topics are organised less often. Not unexpectedly, we observed a substantial variation of specific IPC procedures. For example, pre-surgical shaving is still common for hair removal and skin antisepsis for surgical procedures is still performed using aqueous iodine or alcohol alone in some hospitals. This raises concerns about the way in which hospitals decide to adopt practices. Procedures for the prevention of other HAIs, such as catheter-related infection, show less variation. The reported consumption of alcohol-based handrub (AHBR) is low: 21 ml per patient-day, thus confirming the findings of the most recent European point prevalence survey (19 ml per patient-day). Assuming that 2-3 ml are necessary for a single handrub action, the use of AHBR in Europe is largely insufficient. Despite the many initiatives in past years, more must be done to improve hand hygiene. Interestingly, countries that spend more than the European average of 6.5% of the gross domestic product (GDP) for healthcare consumed significantly more handrub per patient-day than those spending less than the European average.
The randomised controlled trial to reduce CRBSI by promoting hand hygiene and/or a multimodal CVC bundle intervention was a success. Fourteen centres in 11 European countries (Austria, Belgium, Greece, Hungary, Ireland, Italy, Latvia, Poland, Romania, Slovenia, and Spain) participated in this intensive care unit (ICU)-centred study. An already relatively low infection rate of 2.4 CRBSI per 1000 catheter-days was further reduced to 0.9 infections per 1000 catheter-days. Compliance with hand hygiene recommendations and best practice at catheter insertion improved almost everywhere. Hospitals in Latvia, Hungary, and Ireland were among those with high baseline infection rates. This raises concerns about the role of public healthcare expenditure in quality of care as these three countries spend the lowest portion of health expenditure for public health according to the World Bank. The average baseline hand hygiene compliance of 49% was similar to the recent Mastering hospital antimicrobial resistance (MOSAR) study, which was done in similar clinical settings. The hand hygiene programme, based on the strategy of the World Health Organization (WHO), alone or in combination with the catheter bundle, significantly improved hand hygiene compliance. The greatest effect was achieved in the group of hospitals that were randomised to apply the hand hygiene programme only. However, improved hand hygiene improvement was observed also in some of the hospitals that were allocated to implement only the catheter bundle. Similarly, compliance with our catheter insertion checklist improved in hospitals that were randomised to implement the catheter bundle intervention, either alone or in combination with a hand hygiene promotion strategy. The catheter bundle and the insertion checklist were based on a successful prevention programme conducted at the University of Geneva Hospitals (Geneva, Switzerland). Some hospitals in the group randomised to promote only the hand hygiene strategy also improved compliance with the checklist, but the effect was low.
Multimodal intervention strategies or even a simple hand hygiene intervention are not hard-edged programmes. Hospitals are likely to adopt procedures other than the ones prescribed by a programme such as PROHIBIT, e.g. they improve hand hygiene when the intervention is the implementation of a catheter bundle, or they improve catheter insertion and care when the intervention is a hand hygiene promotion programme. We were aware that this would happen when the intervention was designed. For ethical reasons, we did not prevent hospitals from improving procedures other than the ones they were randomised to and we deliberately left room for local adaptation of the intervention by centrally training local stakeholders. Although we cannot distinguish formally the effectiveness between the hand hygiene programme and the more comprehensive catheter bundle implementation, we learnt much about the dynamics of implementing a multimodal strategy. The programmes always focused on a core intervention, i.e. the WHO hand hygiene strategy or the Geneva central-line infection prevention programme, but the boundaries of the interventions were not sharply defined. Encouraged by participating in a European study, local stakeholders developed a range of activities around the core intervention that met the local organisational culture aspects and finally guaranteed success. The study confirmed that multimodal infection prevention interventions are feasible and effective in culturally-different European countries if room is left for local adaptation.
Six themes emerged as the most relevant facilitators and barriers to implementation of the PROHIBIT hand hygiene and catheter bundle programme. However, we observed a considerable variety in the implementation narratives and pivotal drivers of the implementation process. Combinations of factors led to implementation success rather than any one single facilitator. People always made a difference with their talent being important, as well as their position in the institution, and if they succeeded in being boundary spanners between the different professional groups. Surprisingly and noteworthy, success occurred in organisational and cultural environments that appeared to be the most challenging prior to the intervention. The context of an international, evidence-based initiative, but still leaving room for local adaptation, was a major driving and enacting force. Providing hospitals with 50% of a nurse’s salary helped to generate the necessary focus and, together with social bonding among change agents, generated a commitment to succeed. While the halo effect of PROHIBIT was instrumental in under-resourced settings, the provided resources were integrated into pre-existing, locally-branded programmes in institutions that viewed themselves as leaders in the field. Prior to PROHIBIT, all institutions demonstrated knowledge on international good practice guidelines and even local protocols for CRBSI prevention, but not all hospitals had established best practice procedures in their ICUs. We believe that a theoretical framework of social norm activation within and across institutional boundaries, rather than diffusion of innovation, applied as a model in most cases. Establishing a link between the qualitative analysis and the quantitative results remained challenging due to the small number of cases, the already low baseline rate in some hospitals, and the relatively short follow-up time. An important message comes from the fact that the highest baseline rate reduction was most prominent in the most challenging pre-implementation contexts. Unsurprisingly, implementation relied heavily on the commitment of engaged leaders who succeeded to reduce CRBSI even in hostile environments. Of note, if such commitment is not supported by good IPC structures, it may not be sustainable.

Recommendations to the European Commission

The overarching objective of PROHIBIT was to provide the vision for policy-makers, managers, and healthcare workers (HCWs) to prevent HAI by improving the understanding of European guidelines and hospital policies and practices, by overcoming common obstacles for implementing evidence-based best practices, and by testing the effectiveness of two interventions to prevent CRBSI. As a consequence and based on the findings of PROHIBIT, we suggest a number of actions to the European Commission (EC) as to how IPC can be made more relevant in Europe, as well as tangible recommendations to improve patient safety.
- Establish a European IPC strategy.
- Establish a European-wide surveillance for outcome and/or process indicators (a prospective initiative or repeated point prevalence surveys as done at present).
- Create and sustain an IPC platform that serves as a European IPC network.
- Consider train-the-trainer workshops for hospital staff on a European level.
- Consider providing dedicated salaries for a limited time period to get IPC programmes off the ground
- Establish an authoritative body for state-of-the-art guideline production.
- Encourage the adoption of evidence-based IPC strategies in European hospitals through authoritative guidelines and prevention projects.
- Improve hand hygiene in hospitals through a European initiative.
- Raise awareness of the importance of IPC among European HCWs, particularly doctors.
- Foster a positive organisational culture and teamwork in European hospitals.
- Encourage (or mandate) hospitals to establish IPC standards of staffing and organisation/structure (IPC goals, indicators, committee).
- Encourage the establishment of link nurses and doctors (or liaisons) at the hospital level to span institutional boundaries and facilitate multidisciplinary collaboration.
- Encourage countries to provide a budget for IPC activities; in particular, ensure that IPC budgets in hospitals are separate.
- Fund high-quality effectiveness research in the field of IPC emphasizing implementation and behaviour change

Project Context and Objectives:
Background
Healthcare-associated infections (HAIs) are the most frequent adverse events in healthcare delivery and result in increased morbidity and mortality, longer hospital stay and disability, and significant extra-costs. The main HAI categories are catheter-associated urinary tract infection (CAUTI), surgical site infection (SSI), ventilator-associated pneumonia (VAP), and catheter-related bloodstream infection (CRBSI). During the past decade, there has been a proliferation of clinical practice guidelines in healthcare in general, and in infection prevention and control (IPC) in particular. Each of these guidelines recommends numerous practices for HAI prevention. However, it is not well understood which guidelines and specific practices are actually being used and to what extent recommendations of the national documents are evidence-based. In addition, it remains unknown how factors at different levels of health systems, ranging from national policy to individual practitioner knowledge, affect compliance with guidelines. These observations suggest that it would be most useful to have data linking the implementation of evidence-based practices to infection control outcomes, ideally rates of specific HAIs. CRBSI is a prototype of causality and prevention of HAI and represents an ideal scenario to study as it is the most reliably measured and most consistently observed HAI among European hospitals. Evidence suggests that CRBSI rates are amenable to substantial improvement. Studies suggest that the promotion of hand hygiene and a bundle of targeted interventions are effective to prevent CRBSI, but the effectiveness of such interventions remains to be established in the wide cultural diversity of European hospitals. Although multimodal interventions have been shown to be effective on the outcome level, determinants of compliance and non-compliance are not completely understood. Demonstration of the feasibility and effectiveness of these interventions and investigation of the determinants of success could provide suggestions for best practices and lead to the reduction of HAI across European hospitals, particularly CRBSI rates.
Objectives
The aim of the Prevention of hospital infection by intervention and training (PROHIBIT) project was to provide a global perspective of IPC activities in Europe at the following levels: availability of published guidelines and recommendations; IPC practices applied in hospitals as a reality check; the capacity of European hospitals to implement multimodal IPC strategies; and identification of barriers and facilitators in implementing IPC programmes in European hospitals. The specific objectives were:
1. to conduct a systematic review of current guidelines for the prevention of HAI and schemes for the surveillance and public reporting of HAI in European countries (Work Package [WP] 2);
2. to conduct a large survey of European Union (EU) hospitals on current practice for HAI prevention, enabling factors and barriers to compliance with guidelines, and the relationship of these factors to infection control outcomes (WP3);
3. to conduct an in-depth study of factors influencing the adoption and implementation of infection control practices at multiple levels in a sample of hospitals (WP4);
4. to conduct a randomised effectiveness trial of two interventions (World Health Organization [WHO] hand hygiene promotion strategy and a catheter bundle) in a sample of hospitals, focusing on determinants of compliance with the intervention and the incidence of CRBSI (WP5).

Project Results:
SYSTEMATIC REVIEW OF EUROPEAN GUIDELINES FOR THE PREVENTION, SURVEILLANCE, AND PUBLIC REPORTING OF HEALTHCARE-ASSOCIATED INFECTIONS – WORK PACKAGE 2

Background
There has been a remarkable expansion and proliferation of IPC guidelines worldwide. According to existing knowledge, it can be assumed that the variability among countries in the presence or absence of national guidelines is considerable. In many European countries, there are national guidelines for the prevention of HAI established by national committees, such as the Dutch Working Party on Infection Prevention (WIP) in the Netherlands, or the Robert Koch Institut (RKI) in Germany, or commissioned by governmental bodies, like the United Kingdom (UK) Department of Health. In some European countries, the United States (US) Centers for Disease Control and Prevention (CDC)/Healthcare Infection Control Practices Advisory Committee (HICPAC) guidelines are followed entirely (http://www.cdc.gov/HAI/prevent/prevent_pubs.html accessed 18 August 2014). Professional societies and the ECDC have issued also specific guidance or recommendations.4,5 To date, there has been no systematic overview of guidelines issued in Europe.

Objectives
The objective of WP2 was to perform a systematic review of current clinical guidelines and recommendations for HAI prevention developed at the European, national, and subnational (regional) levels with the following defined aims:
• to describe the characteristics and specific elements of these guidelines;
• to identify similarities and differences and potential conflicts among current guidelines;
• to identify key elements of current guidelines that could be analysed for a potential impact on patient outcomes;
• to identify process measures that could be used to construct practical checklists to improve practice;
• to identify and compare schemes for surveillance of HAI rates in European countries;
• to identify schemes for public reporting of HAI information in European countries.

Methods
The dimension of the existing national, regional, and local guidelines allowed a thorough investigation at the national and regional levels focusing on the prevention of SSI, catheter- associated bloodstream infection (CABSI), CAUTI, VAP, and Clostridium difficile infection (CDI). Documents from the following countries were scrutinised: all 27 European Commission (EC) member states (England, Northern Ireland, Scotland, and Wales were analysed as separate entities), Croatia, Iceland, Norway, and Switzerland. A questionnaire was sent to the ECDC HAI surveillance NCPs and other HAI experts in Europe in August 2010 to obtain a first inventory of guidelines for CABSI, CAUTI, CDI, SSI, and VAP. Based on the results, the internet was searched for documents; when no online version was available, the NCPs were invited to send guidelines in print. The most recent versions of national or subnational/regional guidelines published before 31 March 2011 were eligible for review. National guidelines were defined as those published by a nationally recognised committee, such as the WIP in the Netherlands, or a public institution, like the US CDC or the RKI in Germany. Guidelines in languages other than English, French, or German were translated into English and back-checked for validity by the NCPs. A detailed analysis was performed using the research software (QSR NVivo 8). The documents were reviewed for specific recommendations and compared to the topics of the matrix, which was based on the German RKI recommendations and the US CDC guidelines, and adjusted if new aspects were identified. The initial review process was validated by cross-checking for each of the five topics by a second reviewer. The results of the evaluation were fed back to the NCPs for verification.

Results
Retrieving the guidelines was difficult. Not all websites linked to the questionnaire of interest and locating them was possible only by direct contact with the NCP for some countries. In total, 101 guidelines from 34 European countries were identified. Eight countries reported no availability of national or subnational guidelines, four countries had national guidelines on one topic, one country had national guidelines on three topics, and six countries had national guidelines on four topics. Fourteen countries reported guidelines on all five topics (Figure 2).

Figure 2. Number of guidelines available in European countries


Countries with only one guideline offered a document on CDI prevention. Malta uses non-adjusted US CDC guidelines. Italy and Norway provide guidelines exclusively at a subnational/regional level. In the UK, Northern Ireland follows the England guidelines, while Scotland and Wales have developed their own documents. Guidelines are published by a public institution and/or the ministry of health in most countries, except for four countries where the documents are published by a healthcare society or an expert panel. National guidelines have the status of “recommendations” without legal consequences, except in seven countries (Bulgaria, Denmark, England, Hungary, Lithuania, Portugal, and Scotland) where the guidelines or parts of the guidelines are legally authoritative. Scope and detailing differ substantially between the documents. Thus, provision of a guideline on one of the defined topics (CABSI, CAUTI, CDI, SSI, and VAP) does not necessarily mean that all relevant aspects are covered. Particularly in the field of SSI, a number of accessory documents were identified. Most national guidelines do not provide categorised levels of evidence or strengths of recommendations, neither do they document the source of information on which the guidance is based. There was an unexpected variation about how the quality of the scientific evidence was categorised in the different countries.

Prevention of CDI
Eighteen countries have national guidelines for CDI prevention (Austria [2007], Belgium [2008], Denmark [2011], England [2008], Finland [2009], France [2010], Germany [2009], Ireland [2008], Italy [2009], Latvia [2007], Luxembourg [2007], Malta [2008], Netherlands [2006], Northern Ireland [2008], Scotland [2009]), Sweden [2006], Switzerland [1995], and Wales [2008]). Northern Ireland and Wales follow the English guideline. In 2008, the ECDC issued a European guidance document on the topic.6 Five of the 16 European guidelines (Austria, England, Italy, Scotland, Sweden), the Society of Healthcare Epidemiology of America (SHEA)/Infectious Diseases Society of America (IDSA) guideline (used in Malta), and the ECDC guidance use a grading system for the strength of recommendations. Five guidelines (Belgium, Denmark, Finland, Italy, Scotland) are based on the ECDC document. All but one guideline recommend placing patients with symptomatic CDI in a single room if possible. Cleaning and disinfection of the patient environment is considered a key element to prevent transmission. Most guidelines recommend to use a chlorine-based product, but with variable frequencies. The German guideline offers oxygen-releasing compounds as an alternative. All guidelines recommend gloves for patient care, but there is no consensus about how to clean hands appropriately. All guidelines recommend washing hands with soap and water, but some include the use of alcohol-based handrub (ABHR) in varying sequences. Ten guidelines recommend that isolation precautions can be discontinued 48 hours after resolution of clinical symptoms. Some do not comment on the topic and others require a longer period of isolation (up to 72 hours). Only nine documents comment on antibiotic stewardship as a preventive measure for CDI prevention.

Prevention of CABSI
Twenty-one countries have national guidelines for CABSI prevention (Belgium [2000], Bulgaria [2010], Denmark [2010], England [2007], France [2010], Germany [2002], Hungary [2009], Ireland [2010], Italy [2010], Lithuania [2010], Malta [2002], Netherlands [2010], Northern Ireland [2008], Norway [2009], Poland [2002], Portugal [2009], Scotland [2008], Slovenia [2009], Sweden [2006], Switzerland [1997], and Wales [2010]). Eight guidelines (Bulgaria, England, Germany, Hungary, Norway, Portugal, Sweden, Slovenia) and the SHEA/IDSA guideline (used in Malta) use a grading system for the strength of recommendations. The search for guidelines in the UK revealed papers in different formats, one being the epic2 guideline (published in 2007),7 and two short papers entitled “High Impact Intervention” (HII), both referring to the epic2 guideline. Scotland developed practical “cause-and-effect-charts” with few references to the epic guideline. Nineteen documents were evaluated for central venous catheter (CVC) insertion and care.
Six guidelines recommend handwashing and/or surgical hand preparation before CVC insertion and 12 mention “hand decontamination” without specifying any further. One document does not comment on hand hygiene at all. All but one guideline recommend using maximal sterile barrier (MSB) precautions for CVC insertion. The subclavian vein is recommended predominantly as the preferred site of CVC insertion in guidelines and always with the mention to weigh mechanical or other non-infectious risks for the decision. Three guidelines discourage CVC insertion in the femoral vein, unless there is no alternative. Two guidelines do not provide a specific recommendation on the insertion site. Interestingly, the Irish document features a priority list based on mechanical rather than infectious complications which changes the priority sequence starting with the jugular site, followed by the femoral, and ending with the subclavian site. All guidelines recommend skin preparation using an antiseptic solution. Four documents do not specify the product and merely mention a “skin antiseptic”. All other guidelines recommend alcohol alone (three documents) or alcohol-based products, either containing chlorhexidine or iodine. None of the guidelines recommends the routine use of impregnated CVCs. Such catheters are reserved for situations with high infection rates, despite preventive measures, and for patients at increased risk for CABSI. Some guidelines also allow such catheters when the expected dwell-time exceeds five days. Guidelines do not distinguish between antiseptics or antibiotics for catheter coating. Four guidelines mention the use of chlorhexidine-impregnated sponges, one favourably, two as an unresolved issue, and one mentioning its potential risk in children and pregnant women. Given the many documents issued in 2010 this is surprising, in particular after the publication of clear evidence of their efficacy in 2009.8 Eleven guidelines discourage the use of antibiotic ointments at the insertion site, with no mention in the remaining ones. Nine guidelines recommend a daily assessment of line necessity and routine replacement of the CVC is discouraged by most documents. Seventeen guidelines offer recommendations for an administration set change based on the different infusion solutions used, with an interval of 72 hours for aqueous solutions. The Dutch and US CDC recommendation go up to 96 hours. All guidelines comment on the use of two types of dressings: gauze and semi-permeable transparent dressings. Only six guidelines give preference to transparent dressings as these allow inspection of the insertion site. All guidelines advise a change of dressing when damp, soiled, or loosened. The regular interval of dressing change is seven days for transparent dressings and every second day for gauze dressings. Four guidelines recommend to change gauze dressings on a daily basis, especially if the patient is unconscious or uncooperative. Upon dressing change, the insertion site should be disinfected, preferably using an alcohol-based product.

Prevention of VAP
Twenty-one countries have national guidelines for VAP prevention (Belgium [2000], Bulgaria [2010], Denmark [2001], England [2008], France [2010], Germany [2000], Hungary [2009], Ireland [2010], Italy [2010], Lithuania [2010], Malta [2003], Netherlands [2008], Northern Ireland [2010], Norway [2009], Poland [2011], Portugal [2004], Scotland [2010], Slovenia [2009], Sweden [2006], Switzerland [1997], and Wales [2010]). England, Northern Ireland, and Scotland follow the same guideline.9 Nine guidelines (Bulgaria, England, Germany, Hungary, Netherlands, Portugal, Slovenia, Sweden, Switzerland) and the US CDC guideline (used on Malta) have implemented a grading system for the strength of recommendations. The analysis showed that the prevention of VAP is complex and the scope, detailing, terminology, and division of the guidelines vary widely.
Sixteen guidelines recommend keeping ventilated patients in an elevated backrest position. Of these, five specify that this should be only if no medical contraindications exist. Seven documents comment favourably on daily sedation vacation, often in combination with mentioning extubation as soon as possible. Another seven guidelines only recommend the latter. Nine guidelines comment on antiseptic oral care. Preferred solutions are chlorhexidine and iodine, but not all documents offer a specific product. Only a few guidelines provide time frames or define regular application intervals. Regular change of the ventilatory circuit is mentioned in 10 guidelines. However, the wording leaves room for interpretation and some guidelines state that a regular change is not necessary.

Prevention of SSI
Twenty-one countries have national guidelines for SSI prevention (Belgium [2000], Bulgaria [2010], Denmark [2011], England [2008], France [2010], Germany [2007], Hungary [2008], Ireland [2008], Italy [2010], Lithuania [2010], Malta [1999], The Netherlands [2006], Northern Ireland [2008], Norway [2009], Poland [2002], Portugal [2004], Scotland [2008], Slovenia [2009], Sweden [2006], Switzerland [1996], and Wales [2008]). Northern Ireland and Wales use the same documents as England. Eleven guidelines contain information about the rating system used to categorise the quality of evidence and the strength of recommendations. Malta and Portugal use the grading system of the US CDC guideline.
Fifteen guidelines explicitly do not recommend hair removal, and four do not comment on the issue. If performed, 17 guidelines recommend the use of a clipper (if possible immediately prior to surgery). Shaving is either discouraged or not mentioned as a means for hair removal. Depilation is recommended in some documents, mostly as an alternative to clipping. Only one guideline offers the choice between depilation and shaving without mentioning clipping. Fifteen guidelines comment on the issue of preoperative washing, but with varying recommendations: One recommends to do nothing, four recommend washing without further specification, five recommend to use an antiseptic soap without specifying the procedure, two recommend to use a soap without specifying the procedure, two recommend (antiseptic) soap for special procedures, and one recommends washing in the case of pre-operative hospital stay.
Methicillin-resistant Staphylococcus aureus (MRSA) screening is mentioned by five guidelines. Two specifically do not recommend it, two recommend screening for patients at risk, and one promotes “consideration” of screening without providing further details. Thirteen guidelines comment on the topic of routine nasal decontamination and all do not recommend the procedure. Eighteen guidelines comment on the product for skin antisepsis. Ten recommend to use an alcohol-based product for skin (four recommend aqueous solutions for mucous membranes), six recommend to use any authorised product, one guideline prefers a non-alcohol-based product, and one document does not specify a type of product. There are various recommendations about contact time, the technique, and whether the product must be left to dry.
Perioperative antibiotic prophylaxis is the most common means of SSI prevention. Seventeen guidelines comment on one or various aspects of antibiotic prophylaxis. Most documents specify the time of administration as between 30 to 60 minutes before incision. However, there are few guidelines that define the time of “induction of anaesthesia” or refer to the substance used. Only a few guidelines recommend a specific drug, but most discuss the choice based on the type of intervention and/or surgical classification. Eleven documents give advice about re-administration of antibiotics in the case of a prolonged intervention. Eleven guidelines give more or less explicit advice on the frequency of complete air changes per hour in the operating theatre. Recommendations range from 10 to 20 per hour. Eleven guidelines recommend positive pressure, two comment on the level of humidity, and one even defines a specific room temperature in the operating theatre. Nine guidelines define the role of laminar air flow in the operating theatre with varying scope and detail. Twelve also mention the benefit of “high efficiency particulate air filters”, mostly without further specifying their use. The use of ultraviolet irradiation is discouraged in four documents; the other guidelines do not comment on the issue. Finally, thirteen guidelines advise to limit the number of persons in the operating theatre.

Prevention of CAUTI
20 countries were included in the analysis (Belgium [2000], Bulgaria [2010], Denmark [2010], England [2007/2010], France [2010], Germany [1999], Hungary [2009], Ireland [2011], Italy [2010], Lithuania [2010], Malta [2009], Netherlands [2006], Northern Ireland [2007], Norway [2009], Portugal [2004], Scotland [2008], Slovenia [2009], Sweden [2006], Switzerland [2005], and Wales [2010]). There is general agreement on the issue of routine catheter replacement and the use of a closed and sterile system. Ten guidelines include information about the rating system used to categorise the quality of evidence from which the strength of recommendations are derived. Malta and Portugal use the grading system of the US CDC guideline. Sixteen guidelines advise not to replace a catheter routinely. Fifteen guidelines recommend routine hygiene of the periurethral area, usually once or twice daily with soap and water, but most documents discourage the application of antiseptic products. Latex catheters for short-term use are recommended in eight guidelines, although the definition of short-term varies. Silicone catheters for long-term use are mentioned in 10 guidelines. The use of antimicrobial/antiseptic-impregnated catheters is mentioned in five guidelines. One country recommends their use in general, and the others advise to “consider” their use, particularly if CAUTI rates are high. Most guidelines recommend the use of the smallest bore catheter that allows sufficient drainage. Half of the guidelines discourage routine screening for bacteriuria in catheterised patients.

Discussion
There is variation in the content of the guidelines, how they prioritize recommendations, and how (if at all) they rate the strength of recommendations. Although not formally analysed, the quality of the documents varies when comparing it with the evidence-base at the time of their issue. Some websites are not particularly user-friendly and the documents were difficult to retrieve, which may work against accessibility and their wide dissemination. However, this may be a problem for external users only (such as our group) and local stakeholders may be perfectly familiar with the websites and would not meet similar difficulties.



SURVEY OF POLICY AND PRACTICE FOR THE PREVENTION OF HEALTHCARE-ASSOCIATED INFECTIONS IN EUROPEAN HOSPITALS – WORK PACKAGE 3

Background
Effective IPC programmes contribute strongly to patient safety in hospitals. About 10-70% of HAIs are preventable, depending on the setting, study design, infection rates, and type of infection.10 IPC recommendations are not always adopted and implemented in European hospitals and, thus, IPC activities are likely to be variable.11-13 A comprehensive inventory of IPC practices in European hospitals has never been undertaken previously. An interdisciplinary survey was developed and conducted with the aim to assess and analyse IPC activities in European hospitals, and to identify barriers and facilitators to compliance with evidence-based recommendations.

Objectives
The objective of WP3 was to obtain a broad understanding of IPC activities in European hospitals with the following defined aims:
1. to create a European database on IPC structures and activities in a stratified random sample of hospitals in all European countries;
2. to investigate in the current status of HAI prevention among European hospitals and which guidelines and practice recommendations are adopted;
3. to collect data on ABHR consumption;
4. to collect data on patient-days, central line-days, and bloodstream infections (BSI) at the hospital and intensive care unit (ICU) levels;
5. to identify facilitators and barriers to compliance with good practice procedures;
6. to develop an assessment tool for IPC and organizational culture.

Methods
An interdisciplinary group within PROHIBIT members established an online questionnaire addressing the organisation and management of IPC, implementation of evidence-based practices, surveillance of process and outcome indicators, and shared ownership of IPC. The questionnaire was translated into Bulgarian, Finnish, German, Greek, Italian, Lithuanian, Portuguese, Slovak, Slovenian, and Spanish. ECDC NCPs and IPC experts of countries outside the EU were invited to organise national polls. NCPs invited national hospitals to participate between September 2011 and March 2012. All data were entered into a web- based survey tool (Limesurvey Version 1.92).
Data were checked for plausibility and analysed descriptively by summarising categorical parameters as proportions and continuous data were presented as the median and interquartile range (IQR). Countries were stratified according to the United Nations’ regional groups (Western, Northern, Southern, and Eastern European regions), and healthcare expenditure as a percentage of the general domestic product (GDP) (above or below the average European expenditure of 6.5% of the GDP). Multiple multivariable logistic regression analyses were performed with the outcome parameters of ABHR consumption and “best practice” and factors related to the ICU/ward, hospital, and national levels.

Based on the results and feedback from the hospitals, the PROHIBIT WP3 questionnaire was further developed by focusing on organisational and patient safety aspects with the aim to provide a tool for self-assessment and benchmarking with European hospitals. This final questionnaire was established in collaboration with CUB, ICL, and UniGE, and is one of the main two tools of PROHIBIT.

Results
Thirty-four European countries were invited to participate in the survey. Twenty-four (68%) submitted data from 308 hospitals representing 397 ICUs, 514 medical wards, and 515 surgical wards. One-third (32%) were university-affiliated hospitals. Participating hospitals, ICUs, and medical and surgical wards had a median [IQR] of 368 beds [200; 670], 9 beds [6; 12] (8 ventilator beds [5; 10]), 30 beds [24; 40], and 30 beds [25; 40], respectively. ICUs, and medical and surgical wards had a median of two single rooms. Most ICUs (68%) in the survey were mixed. Only 14% and 13%, respectively, were exclusively medical or surgical. Most ICUs (86%) had ABHR dispensers at the point of care in >75% of beds/bays. Personal pocket dispensers were available in 28%. Many medical wards (46%) in the survey were general. Most wards (65%) had ABHR dispensers at the point of care in >75% of beds/bays. Personal pocket dispensers were available for HCWs in 31% of wards.

IPC structure and organisation
Hospitals had a median (interquartile range [IQR]) of four IPC nurses [2; 6] and one IPC doctor [0; 3] per 1000 beds. Two-thirds (66%) of hospitals implemented a link nurse system. IPC was predominantly an independent department (39%), but affiliated to hospital administration in 20%, to microbiology in 14%, and to infectious diseases in 7%. Ninety-one per cent of hospitals have an interdisciplinary IPC committee. In addition to IPC professionals, the committee members included nursing directors (71%), medical directors (65%), the chief executive officer (CEO) (34%) and/or the deputy CEO (19%). Most hospitals (96%) defined IPC objectives, which mainly addressed hand hygiene (91%), HAI reduction (87%), and antibiotic stewardship (69%).

Training of HCWs
Regular IPC training sessions for HCWs were organised in 91% of ICUs, 90% of medical wards, and 88% of surgical wards. Most wards offered training in hand hygiene (95%), but other IPC measures were less often addressed. HAI prevention campaigns have been organised more often in ICUs (82%) than in non-ICUs (73%, medical wards; 71%, surgical wards).

Surveillance
Surveillance of any HAI was performed in 96% of survey hospitals. Most often, surveillance addressed CLABSI (86%), followed by SSI (85%), and CDI (68%). Interestingly, CLABSI surveillance was performed mostly hospital-wide (70%). The most frequent indicator operation for SSI surveillance was hip replacement (65%). Process indicators were monitored as ABHR consumption (80%), or direct observation of hand hygiene compliance (77%).

Prevention of CLABSI in the ICU
The majority of participating ICUs had local written policies for central line insertion (90%) and care (87%). A standardised approach for daily line assessment was established in 70%. The subclavian vein was the preferred insertion site in 54% (65% in Eastern and 63% in Southern European regions) and the jugular vein in 37% (59% in Northern and 49% in Western European regions). Skin antisepsis was performed mainly with alcohol-based chlorhexidine (37%), followed by ethanol/isopropyl alcohol alone (21%), aqueous iodine (17%), and alcohol-based iodine (14%). Most central lines were placed using MSB precautions (85%). Approximately one-third (31%) of the ICUs used impregnated catheters, although they were used in less than 50% of patients in half of the institutions. Sixteen percent use chlorhexidine-impregnated sponges. Single-use vials were discarded immediately after use in 58%. Administration sets for aqueous solutions were rarely (1%) left in place until 96 hours.

Prevention of VAP in the ICU
Most participating ICUs (72%) had local written policies for VAP prevention. Daily sedation vacations were performed in 67%. All ventilated patients were in elevated backrest position of 30°-45° unless contraindicated [IQR: 88%; 100%]. Oral care was predominantly performed with chlorhexidine (73%), followed by the use of other substances (24%) and selective oral decontamination (3%). Oral care was performed at fixed intervals in most ICUs (83%).
Prevention of CAUTI in medical wards
Most participating medical wards had local written policies for CAUTI prevention, 84% on urinary catheter insertion, and 72% on urinary catheter care. A system for daily assessment of catheter necessity was in place in 54%. Closed drainage systems are used in 83% (11% sometimes or rarely). Impregnated (antiseptic or antibiotic) urinary catheters are used only in 9%.
Prevention of CDI in medical wards
Most participating medical wards (65%) had local written policies for CDI prevention. Almost all wards (97%) use gloves for the care of symptomatic patients. Patients with CDI were isolated: 1) always in single rooms (44%); 2) always in single rooms if available (30%); or 3) in contact isolation in shared rooms (17%). Ten percent of medical wards do not isolate symptomatic CDI patients. Most wards (62%) change surface disinfectants to a sporicidal agent in the case of a symptomatic patient. Eighteen percent adhere also to this practice, but only in the presence of a cluster of symptomatic CDI patients. C. difficile testing is performed: 1) for any patient with loose stool (29%); 2) a patient with loose stool and concurrent antibiotic treatment (13%); and 3) a patient with loose stool within 72 hours upon admission (8%). Almost half of the wards (52%) perform laboratory testing of C. difficile only on upon the doctor’s request. This is observed most frequently in Southern regions (80%), but less frequently in Northern regions (21%).
Prevention of SSI in surgical wards
Participating surgical wards were general (43%) and orthopaedics/trauma (22%). Most wards (66%) had ABHR dispensers at the point of care in >75% of beds/bays. Personal pocket dispensers were available in 34%. The majority of participating wards (80%) had local written policies for SSI prevention. Most wards (41%) still remove hair by shaving on the day of surgery or the night before (15%), with particularly high proportions in the Eastern region (68% and 26%, respectively). Twenty-five percent use clipping (mostly in Southern and Western Europe [47% and 33%]). No preoperative hair removal is done in 15% (44% in Northern and 18% in Western regions). Surgical antimicrobial prophylaxis is applied: 1) as a single dose (71%); 2) with repeated administration after 4 hours in long procedures (51%); and 3) continuously after surgery (47%). The latter is less frequent in the Eastern region (40%), but more frequent in the Southern region (53%). In the case of continuous administration, 57% had a stop order in place (most frequently in Eastern and Northern regions, 66% and 67%, respectively). Responsibility for the correct application of antibiotic prophylaxis resides most often with the operating surgeon or the ward doctor (55% and 34%, respectively). Most wards (40%) use an alcohol-based combination for preoperative skin antisepsis (16% alcohol-based chlorhexidine, 23% alcohol-based iodine, 1% alcohol-based octenidine). Aqueous iodine was used by 25% of wards, and alcohol-only preparations by 20%.
Patient safety culture
In the ICUs and medical and surgical wards, senior professionals reminded frontline HCWs to perform hand hygiene if they forgot in 56%, 45%, and 43%, respectively. Frontline HCWs reminded senior professionals (e.g. head physicians) in 70%, 56%, and 62%. Rings were worn rarely in the different wards. Fifty-five percent of hospitals reported to have walk-rounds with the CEO or his/her deputy, although with large regional variations. Twenty-five percent of IPC staff reported that the implementation of evidence-based recommendations would be difficult in their hospital. This percentage was even higher in Southern and Western regions (31% and 43%, respectively).

Process and outcome indicators
ABHR consumption
Median [IQR] hospital-wide ABHR consumption per patient-day was 21 ml [9; 37]. ABHR consumption in the ICUs and the medical and surgical wards was 66 ml [38; 103], 12 ml [5; 25], and 14 ml [7; 25], respectively. For countries that spent more than the European average of 6.5% of the GDP for healthcare, the median [IQR] ABHR consumption in the ICUs and medical and surgical wards was 70 ml [41; 104], 18 ml [11; 30], and 19 ml [11; 29] ml per patient-day, respectively. Among those that spent less than the European average, the median [IQR] ABHR consumption in the ICUs and medical and surgical wards was 54 ml [26; 103], 6 ml [3; 10], and 8 ml [4; 16], respectively. The differences were statistically significant for all settings. Provision of ABHR at the point of care (>75%) depends significantly on healthcare expenditure: 74% vs. 96% in ICUs; 43% vs. 80% in medical wards; and 44% vs. 81% in surgical wards (p<0.001).

BSI
Sixty-eight ICUs reported data on healthcare-associated BSI. The median incidence density was 3.7 BSI per 1000 central-line days with significant variation [0.8; 22.7].

Guidelines and practice recommendations adopted and implemented in Europe
Local policies in ICUs and medical and surgical wards are based most often on national guidelines (82%, 85%, and 81%, respectively), followed by international guidelines (62%, 57%, and 54%, respectively), and the most recent evidence from the literature (51%, 45%, and 48%, respectively). Leading physicians (65%) and nurses (72%) were more often involved in the design and implementation of the local policies in the ICU compared to medical (38% and 53%, respectively) or surgical wards (50% and 53%, respectively). IPC staff was involved in such processes most of the time (74-87%). Most hospitals (81%) agreed that national guidelines in their language were available, although this proportion was lower in the Southern (76%) and Eastern (77%) regions. Most hospitals (80%) considered the available national guidelines as comprehensible and, consequently, most institutions (83%) use such documents in the implementation of IPC programmes.

Adoption and implementation of country-specific recommendations in national guidelines (matching of WP2 and WP3 data)
Matched analysis of WP2 and WP3 data revealed that less than 70% of wards implemented the specific national recommendations for a number of indicator procedures. Data for CLABSI prevention show that an equal proportion of countries with or without national guidelines follow best practice recommendations: 79% vs. 78% for MSB precautions; 91% vs. 92% for avoiding the femoral insertion site; and 63% vs. 74% for daily assessment of line necessity. However, 83% vs. 64% of hospitals in countries with or without such a recommendation used chlorhexidine for skin antisepsis.

Development of an IC assessment tool
The PROHIBIT WP3 questionnaire was further developed by focusing on: commitment (six questions); leadership (seven questions); supporting work environment (10 questions); shared ownership of IC (three questions); learning culture and culture of continuous improvement (three questions); open culture (seven questions); responsibility (seven questions); and accountability (five questions). All questions have a theoretical background that will allow scientific evaluation. The tool will be piloted in three to four countries in August/September 2014 and offered to European hospitals by the end of the year.

Discussion
Twenty-four countries provided data from 308 acute care hospitals with 397 ICUs and 1019 medical and surgical wards. Not unexpectedly, IPC practices vary across Europe as do HAIs and process indicators, such as ABHR consumption. For example, preoperative shaving is still common for hair removal in surgical patients, particularly in Eastern Europe, and antibiotic prophylaxis is unnecessarily prolonged, particularly in Southern Europe. Such differences may be due to variations in country-specific IPC recommendations. However, hospitals do not necessarily follow national guidelines, but this does not mean that they do not follow best practice. Although they claim to use internal guidelines, which are based on national guidelines, most hospitals follow best practice ⎼ even if this is not (yet) part of an available national document. Thus, hospitals are likely to follow partially best practice as defined in the literature. Although this is good news, hospitals do not always follow all recommendations for a specific prevention strategy, e.g. one-third of ICUs do not have a formal approach for the assessment of daily line necessity and only two-thirds offer daily sedation vacation for mechanically-ventilated patients. This raises concerns about the way hospitals decide to adopt or not adopt practices, but also about the availability of best practice guidelines in the local language and adapted to national circumstances. Together with the findings of WP2, this is a call for action to create a European body dedicated to guideline production. Staffing levels of IPC have slightly improved over the past 10 years. In 2001, the Antimicrobial Resistance Prevention and Control (ARPAC) study identified a rate of 2.3 IPC nurses per 1000 beds; the PROHIBIT survey revealed a rate of four nurses per 1000 beds.12 The number of IPC doctors (one per 1000 beds) has remained stable compared to the ARPAC data (0.94 IPC doctors per 1000 beds in 2001) and was lower than in the European point prevalence survey published in 2013, which identified 0.36 IPC doctors per 250 beds.17 Current staffing levels for IPC nurses are similar to the recommendations of the US SENIC study conducted 30 years ago, and it is debatable whether these recommendations are still valid.18 Expert opinion suggests that the IPC professional-per-bed-rate should be 0.8 -1/100 beds in acute care,19 and 1/150-250 in long-term care.20

PROHIBIT identified a hospital-wide ABHR consumption of 21 ml per patient-day, similar to the European point prevalence survey (18.7 ml/patient-day [IQR 10.3; 30.6]).17 Assuming that 2-3 ml are necessary for a single handrub action, ABHR use is largely insufficient and there is room for improvement in Europe. The correlation of ABHR consumption and the availability of handrub at the point of care with country healthcare expenditure highlight the role of sufficient resources for the prevention of HAI. However, ABHR consumption was associated also with reporting of hand hygiene compliance to an infection control committee, IPC as integrated in the hospital administration, and to less tangible cultural aspects.

HCWs do not remind each other appropriately in the case of low hand hygiene compliance. Reasons for such behaviour may be low priority of hand hygiene in an institution, or aspects of the organisational culture with an unfavourable power distance between senior and frontline staff. The newly-developed PROHIBIT assessment tool will address aspects of safety culture in European hospitals.


IN-DEPTH QUALITATIVE INVESTIGATION OF SUCCESS FACTORS FOR THE ADOPTION AND IMPLEMENTATION OF INFECTION PREVENTION PRACTICES – WORK PACKAGE 4

Background
The implementation of evidence-based infection control practices is challenging for healthcare institutions worldwide. Although acknowledged that implementation success differs according to contextual factors, little is known regarding the most critical specific conditions within the complex cultural milieu of varying economic, political, and healthcare systems. Given the increasing reliance on unified global schemes to improve patient safety and healthcare effectiveness, research on this topic is needed and timely.
Only recently has more attention been devoted to investigating why such variations exist in the abilities of hospitals to translate evidence into practice.11,12 Variations in implementation success depend not only on the quality of intervention programmes, but may be due also to differences in the organisational context, which Øvretveit defines as “all factors that are not part of a quality improvement intervention itself”.22 Contextual factors are often cited as playing a role in the varied success of best practice interventions. However, the existing literature on implementation strategies often lacks details related to specific contextual factors and provides limited information about the real barriers and facilitators to implementation.23,24 This is of particular importance in Europe given the present backdrop marked by a wide variety of cultures, healthcare organisations, and economies across countries.25

Objectives
WP4 of PROHIBIT (short title: InDepth) aims to assess barriers and facilitators to the successful implementation of CRBSI prevention in ICUs across several European countries. In particular, it seeks to fill the knowledge gap by answering the question: “By what processes are particular innovations in health service delivery and organisation implemented and sustained (or not) in particular contexts and settings, and can these processes be enhanced?”.26

Methods
InDepth used a qualitative case study approach in the ICUs of six purposefully selected acute care hospitals among the 14 participants in the PROHIBIT intervention study. We conducted interviews with hospital health providers at different organisational levels, performed ethnographic observations and artefact collection, and took photographs during two rounds of two-day on-site visits. The intitial visit took place one month before the intervention and the second after one year of implementation. These multiple data sources allowed for triangulation in an iterative coding process starting inductively according to the grounded theory in the first round and then using the themes that emerged for the second round. Implementation success was defined qualitatively (implementation fidelity, perceived benefit, adaptation) and quantitatively (CRBSI and performance indicators from the WP5 intervention).
The “InDepth” qualitative inquiry asked two questions:
• why are some hospitals/ICUs better at implementing CRBSI prevention than others?
• what are the main barriers and facilitators to successful implementation of best practice in CRBSI prevention?
Originally, the objectives of WP5 have been defined in the PROHIBIT study protocol as follows:
“The WP4 objectives are designed to deepen the knowledge obtained from the WP3 survey and bring practical realism to the quantitative results from WP5. Such knowledge is key for the formulation of policies and practice recommendations in WP6 because it takes into account the significant variability among healthcare settings in Europe.27,28

Our model of the PROHIBIT WP5 implementation scenario followed a simplified temporal scheme based on the ‘diffusion of innovation’ paradigm.26,29 According to this theory, adoption of an innovation (here the PROHIBIT targeted CRBSI prevention programme 30 and/or the WHO hand hygiene promotion strategy32,33) is followed by implementation, which leads to institutionalisation; the institutional characteristics and context and the characteristics of the ‘innovation’ determine diffusion success. The study group was sensitised by studying further models of implementation and practice change in healthcare, including the theories of sensemaking35 and new institutionalism.36 For our purposes, we interpret new institutionalism as a prolongation of sensemaking. According to this logic, individuals discovering something new will undergo a period during which they make sense of it (name, categorise, and relate it to concepts they are already familiar with) before it becomes part of the institution, aligning with social norms and expectations. The Consolidated Framework for Implementation Research (CFIR) served as a landscape constituted of numerous validated theories that impact the translation of research findings into practice.37 This framework incorporates also the extensive literature review and diffusion of innovation framework by Greenhalgh et al.26 We used many of the concepts, but were most influenced by the CFIR distinction of five implementation domains: intervention characteristics (with a core and an adaptable periphery); outer setting; inner setting; people involved; and implementation process. The InDepth research team included doctors, nurses, and qualitative researchers, thus forming an interdisciplinary group to enhance reflexivity throughout the inquiry and analysis.
In the face of non-linear causality and complex interactions of adoption and implementation,38 qualitative research methods are particularly suitable to put quantitative findings into perspective, in particular when conceived as mixed methods studies.39-42 The setting in this study was extremely complex given the European variations in culture, healthcare systems, political and economic background, and infection control policies. All of these contextual factors play an important role in the varied success of best practice interventions. We chose a case-study methodology with longitudinal follow-up to guide the qualitative investigation43 in order to acknowledge that the boundaries are not clearly defined between the context and the implementation itself. The hospital is the unit of randomisation considered for WP5 and, consequently, we define each hospital as a primary case with ICUs representing nested case studies.43
Six hospitals were deliberately selected among the 14 WP5 intervention study hospitals based on the scheme of extremes and intensity43,44 so as to include three hospitals predicted as being challenged and three unchallenged by the implementation task. Selection was performed following an online questionnaire and telephone interviews with the OSI to identify information-rich cases. Implementation success was defined and assessed on several levels: 1) quantitatively by CRBSI rate (outcome measure) and adherence to CVC insertion and hand hygiene protocols (process measures) produced under the WP5 study protocol; and 2) qualitatively, defined as the satisfaction of our interviewees with the intervention programme, the implementation, and outcome; the extent of intervention fidelity (adherence to core content); and successful resolution of intervention-context fit issues (adaptation, local re-engineering, and innovation).
Data collection took place during on-site visits one month before and one year after the intervention launch. These include 10 to 15 semi-structured interviews per site with key professionals (e.g. hospital CEO, head doctor, head nurse; ICU head nurse and doctor; infection control team head doctor and nurse; frontline nurse and doctor; PROHIBIT study nurse and OSI) who were identified in advance or ad hoc, ethnographic observations in the ICUs focusing on ergonomic principles, work attitude, hand hygiene behaviour, catheter handling, and collection of written protocols and photos. Three months into the implementation, additional telephone interviews with the OSI were conducted to follow up on the implementation process and for member checking, i.e. to verify our interim conclusions with study participants.43
For the first round visits, we employed a grounded approach to data analysis.45,46 This research process was iterative and emergent such that interim findings may influence the ongoing data collection. For the second round analysis we applied a more deductive approach, based on the analysis of the first round visits. Themes established through cross-case analysis during the first round visits were established as a priori codes used to inform the interview guide for the second round and anticipated topics regarding the implementation of CRBSI prevention and hand hygiene promotion procedures. The first round served to reveal the contextual factors and readiness for implementation of the institution; the second round visits served to relate these contextual factors to the implementation of the PROHIBIT programme.
Finally, we analysed the material in a three-dimensional matrix using a stacking technique, whereby cases were grouped according to common traits; differences and similarities in themes or phenomena are then investigated across stacked cases.47 Interviews were conducted primarily in English, when necessary with the aid of translators. Language barriers were overcome by the researchers’ language skills (German, Italian) or translators. The interviewees and study institutions gave written consent based on a confidentiality agreement. An ethics committee in each study institution approved the study protocol.

Results
The sample of InDepth study sites was established to represent a balanced mix of different study intervention arms, geographical locations, and predicted implementation success. These were located in the following European regions: Mediterranean basin, Baltic states, British Isles, Alpine arc, and the Visegrad group (an alliance of the Czech Republic, Hungary, Poland, and Slovakia).
The PROHIBIT intervention
The PROHIBIT intervention programme was not a set of hard-edged prescriptions, but rather designed as a train-the-trainer approach. This means that each study site had to identify an OSI and a study nurse (whose salary was covered at 50% for data collection) who then attended the several workshops held by the investigators and other faculty. Justification for the study design was based on the insight of the PROHIBIT expert advisory board who collectively brought a great deal of implementation and campaigning expertise. In addition, it was informed by knowledge gained through the WHO hand hygiene campaign and more general, recent implementation literature.26,37
The three workshops in Groningen, the Netherlands, and one in Geneva, Switzerland intentionally included a team-building dimension. This was reported by our interviewees as important. Social cooking events were organized following some workshops. Study site representatives who attended the workshops maintained an informal and self-initiated network among themselves. We discovered later that this contributed to a context of social norm among study centres.
IPC nurse – “Well, there’s always the thing with being involved with other people in a multicentre study, obviously, and particularly one like this that is multicountry as well.”
Themes
Among the full list of 21 items, the following six themes emerged as most central to implementation success: resources; staff issues; influential individuals; attitude of excellence; networks and communication; and previous implementation experience. These are discussed here in the light of barriers and facilitators to the successful implementation of PROHIBIT.
Resources
The theme Resources was omnipresent and almost always referred to as a barrier to implementation of the PROHBIT programme and for patient safety in general. Resource problems were more prominent in the Eastern European study sites (see also Staff issues) where it impacted on an organisational level and interfered with decisions concerning purchasing materials for the programme, such as ABHR, drapes, and CVC insertion sets. On an individual level it had an impact on salaries, which in turn influenced staffing and staff motivation (see also Staff issues).
ICU nurse - “The main barrier... [we] have struggled very, very hard to get these single-use units together. [We] have worked with the companies in order to make these sets from all the materials necessary but, unfortunately, the hospital is not able to approve this type of set so [we] actually feel that [we] have gone back to 20 years ago by using equipment that has to be sterilized”

In Western European study centres, financial crises in the EU led to budget cuts that influenced nurse-to-patient ratios. It is worth noting that in some instances, particularly in high-performing institutions, this was seen as a challenge, which motivated hospitals to become more efficient and innovative in order to maintain high quality of care.
Staff issues
Staff issues was a relevant theme at the hospital and ICU levels, but also at the IPC level and, more specifically, at the PROHIBIT programme level. Linked to resource constraints in Eastern European study sites, many HCWs reported either having to hold down two or more jobs in parallel or to leave the country and migrate to Western Europe where salaries are better. In some cases, salaries for ICU nurses were lower than that of a cashier of newly-opened supermarket chains in town. As a consequence, brain drain (out of the institution and the country) affects also nurses as well as doctors.
IPC doctor - “Well, for example, take the hospital ICU where the brightest doctors have left - the brightest! Because they are very much in demand, particularly from the ICU and anaesthesiology, and some other specialties, like pathology general surgery, but not all. But in some fields, we have seen an exodus. And nurses, lots of nurses go too […]”

Moreover, because of insufficient pensions, HCWs stay beyond retirement age in some cases if the system allows them to do so. High turnover of HCWs in the ICU across most sites is usually considered as a barrier because re-training is necessary and not always possible. However, it can be a facilitator when young teams emerge who are more open to change and innovation, and where the culture has a lower power distance. Staffing shortage was common in IPC teams and indeed was also a frequent motivation of hospitals to apply to participate in PROHIBIT so as to benefit from a 0.5 full-time equivalent of a dedicated IPC nurse (see also Perceived value of the intervention). Absence or turnover of the local PROHIBIT representatives (change agent) was a barrier for implementation.
Previous implementation experience
This theme was especially important in the first round site visits where we learned more about the history of the hospital with prior experiences of implementing IPC- or patient safety-related programmes. We found proof of implementation fitness in sites that:
- strategically had used international research programmes for funding of infection control;
- benefited by learning from negative experience, such as outbreaks, to establish IPC in the hospital;
- demonstrated a culture of elaborating good practice protocols as interdisciplinary projects to obtain frontline buy-in;
- sustained an integrated culture of performance indicators;
- showed innovative capacities by using social marketing strategies to overcome barriers.
We learnt that the model of Diffusion of innovation29 is only partially applicable to programmes such as PROHIBIT, since most study hospitals already knew about international good practice guidelines for CRBSI prevention and hand hygiene promotion ⎼ if they had not already implemented such protocols and practices, at least in theory (see also Perceived value of the intervention and Discussion).
ICU doctor - “The protocol is quite clear and quite simple. It is only a change in mentality, as far as I’ve seen, yes...”
Attitude of excellence, work attitude, safety culture
This theme belongs to the context of the outer and inner settings. Some hospitals had achieved the status of a magnet hospital where HCWs were eager to work because of good working conditions and image; some were even a reference for other hospitals in the region. This results in being able to select better trained and motivated staff. In two cases, implementation was very successful in command-and-control cultures with a larger power-distance when a formal leader decided to engage. This removed many barriers instantly in more administrative-led institutions.
ICU nurse - “Well, we have a chief here who is really good at persuading people, so for our doctors it wasn’t a problem. She just said: that’s the way we are going to do it now and everyone should do this. So that was the way that we persuaded everyone to follow the procedure.”

However, sustainability is questionable in such situations. In one case, we observed the retirement of an influential leader, which clearly jeopardized the position of IPC in the institution. In several sites, nurses showed a work attitude of obedience, but in all countries education now moves towards academic training – although the proportion of nurses with a bachelor or master vary substantially. Safety culture understood as an attitude of learning and speaking up to maximise safety and performance is a clear facilitator for implementation. We observed a positive association between attitude of excellence, safety culture, and low baseline CRBSI rates.
Influential individuals
People make a difference – particularly in a challenging context. We established this theme to include not only assigned leaders, but also champions at any professional level and to link individual personalities directly to their influence in the implementation process. We noted the importance of who has the power in an organisation in both situations – when a gifted leader came on board and when good leaders left. The latter case demonstrated also the main potential problem with this facilitator, i.e. the limited sustainability, especially in a context of a command-and-control culture. Examples of influential individuals facilitating PROHIBIT were: a medical director and head of IPC/ICU who managed to obtain funds to finalise an ICU that had been left unfinished for 20 years; a head of ICU who was particularly skilled in motivating staff and who established an open culture for innovation; and an IPC nurse who changed the attitude of ICU staff towards IPC by her personality and ongoing presence in the ICU. Influential individuals, particularly in the ICU, were able to overcome organisational financial barriers by networking with assigned leaders to make CRBSI prevention an institutional priority. Influential individuals succeeded to become central to implementation success mostly through continuity, their presence in the units, their street credit with frontline HCWs, their boundary spanning networking between IPC and ICU, and their personality and social talent. Interestingly, we observed greater levels of nurse investment and nurses acting as influential individuals in hospitals randomised to the hand hygiene intervention, whereas greater doctor involvement was observed for hospitals randomised to the CVC bundle.
Networks and communication
Separation of nurses and doctors, where nurses are seen as subordinate, IPC perceived as police, and complicated administrative procedures were cited as barriers. Conversely, friendship among staff across hierarchies and professional categories, a system of link nurses and doctors (or liaisons), leaders who are readily accessible, and open communication and full integration of IPC nurses in the ICU were judged as facilitators of PROHIBIT implementation.
ICU doctor - “In my experience, my leaders, I have a lot of persons above me that decide, but the persons that I consider my real leaders/chiefs are directly connected with me and are friends of mine. So it’s quite easy for me to work with them. We go to the restaurant.”

Interestingly, in one study site where friendship was perceived as a facilitator, the institution had established also a culture of regular formal meetings across hierarchical levels. As a project linking IPC and ICU and because of a dedicated person to this programme, PROHIBIT improved networking and collaboration in almost all study sites.
ICU head nurse - “I think it allowed the infection control department to be more present within the ICU … and [the IC nurse] would always have a very good link with our staff but now she’s got an extra reason to be around and to communicate and to come and work with the staff...”
Perceived value of the intervention
PROHIBIT helped networking between IPC and ICU staff in many hospitals and had a facilitating effect for IPC image and projects beyond the declared scope of the intervention.
IPC doctor - “I think that that it has had an effect, not just on hand hygiene, but in being able to suggest other bits and pieces around the whole infection prevention and control area.”

Standardisation of protocols and training perceived as evidence-based and supported by an internationally renowned or respected source (PROHIBIT) helped to alleviate ambiguity (“being sure to do the right thing”) and increased hopes for sustainability. Several hospitals also viewed their participation in PROHIBIT as providing the momentum to establish CRBSI surveillance in their facility. The international character of the programme and its affiliations offered the promise of international collaboration, thus creating a halo effect that may have been influential to project participation decisions, particularly in under-resourced settings. In hospitals that already had well-established international collaborations (and also higher resources), the PROHIBIT “brand” was not necessarily retained and implementation activities were perceived rather as local actions. Hospital senior management was not always aware of the programme. In general, the IPC and ICU had to fight to get the necessary resources for programme implementation.

Discussion
Six themes emerged as the most relevant facilitators and barriers to implementation of PROHIBIT. However, we observed a considerable variety in the implementation narratives and who were the pivotal drivers. Combinations of factors led to implementation success rather than any one single facilitator. People always made a difference with their social talent being important as well as their position in the institution, in particular if they succeeded in being boundary spanners. Such individuals often played a pivotal role in overcoming organisational barriers, particularly in countries with lower resources.
Of note, disruptive success occurred in organisational and cultural environments that appeared the most challenging prior to the intervention. The context of an international, evidence-based initiative, but still leaving room for local adaptation was a major driving/enacting force. Providing hospitals with 50% of a nurse’s salary helped to generate the necessary focus and, together with social bonding among change agents, generated a social commitment to succeed. While the halo effect of PROHIBIT was instrumental in under-resourced settings, the provided resources were integrated in pre-existing, locally-branded programmes in institutions that viewed themselves as leaders in the field. Knowledge of international good practice guidelines and even local protocols for CRBSI prevention had been present in all institutions prior to PROHIBIT. Thus, we believe that a theoretical framework of social norm activation within and across institutional boundaries applied as a model in the majority of cases, rather than diffusion of innovation. Establishing a link between the qualitative analysis and the quantitative results remained challenging due to the small number of cases, the already low baseline rate in some hospitals, and the relatively short follow-up time. However, an important message comes from the fact that the highest baseline rate reduction was most prominent in the most challenging pre-implementation contexts.

Further research
Two follow-up research programmes should be considered. First, because PROHIBIT provided resources and a socially-enacted goal, it seems crucial to follow up on the evolution after the end of the programme. This would provide knowledge to a future funder of a similar programme, particularly if rolled out on a larger scale. Second, we observed that the very qualitative approach of InDepth created experiential knowledge in the research team about facilitators and barriers to successful implementation and the process of inquiry itself made a difference in hospitals. For these reasons, it seems beneficial to further develop InDepth into an implementation toolset.

Based on the results of WP4, the following facilitators emerge as recommendations to the EC to enhance implementation of good practice guidelines in healthcare institutions.
1. Make the programme attractive
o endorse by well-respected international organisations;
o make it attractive for evidence-driven healthcare workers by being scientifically sound and evidence-based;
o establish buy-in by involving local, front-end stakeholders in the train-the-trainer approach:
o create and sustain a social platform for those involved in the programme;
o consider a train-the-trainer approach for internal change agents rather than just guidelines;
o provide opportunity for local adaptation, but maintain unambiguous core messages;
o grant salaries for under-resourced institutions for a specified start-up time.
2. Help to establish institutional focus
o provide guidance and support for standardised monitoring and feedback;
o consider visiting centres using a qualitative inquiry approach, such as InDepth, with a supportive intention;
o ensure that national policies are compatible with the programme.
3. Help attributing necessary means
o dedicate programme agents who are talented and have street credit with nurses and doctors;
o ensure coverage for material declared essential to the programme;
o provide adequate staffing with well-educated personnel in general;
o encourage the establishment of dedicated IPC budgets at the hospital level.
4. Safety culture
o cultivate a local safety culture through the establishment of error-reporting mechanisms, together with root-cause analysis to encourage learning from past events.
5. Networking
o create and sustain a social platform for those involved in the programme;
o encourage the establishment of link-nurses and doctors (or liaisons) at the hospital level to span institutional boundaries and facilitate multidisciplinary collaboration.


Table 2. Facilitators and barriers to implementation success

Facilitators Barriers
Outer setting - Sufficiently resourced healthcare system
- Culture of performance (indicators) in healthcare - Disruption in HCW education systems
- Inexistent or changing national healthcare politics
- Brain drain of well-trained HCWs
- Inflexible national staffing requirements’ regulations
Inner setting - Previous implementation success
- Culture of performance indicator use
- IC well networked/
- recognised
- Attitude of excellence culture - Senior management not interested in patient safety
- Competing priorities, such as institutional reorganisation
- Understaffing at bedside
People involved - Person dedicated to the programme
- Personality, presence, and street credit of involved person - Changes in personnel dedicated to the programme
Intervention - Branded as international evidence-based programme
- Potential for local adaptation without losing a clearly-defined core
- Monitoring and feedback of process and outcome indicators - Complexity in data entry
Implementation process - Success in creating focus and positive image of the programme
- Support by person in a power position - Too many programmes in parallel
- Lost to follow-up
- Successful implementation by people in power positions might threaten sustainability (to be verified by a follow-up research project)

RANDOMISED EFFECTIVENESS TRIAL OF TWO INTERVENTIONS TO REDUCE CATHETER-RELATED BLOODSTREAM INFECTIONS – WORK PACKAGE 5

Background
Several studies have shown the effectiveness of multimodal and bundle strategies in CRBSI prevention. All published studies were either performed in a single centre or in multiple centres that were quite similar. Similarly, improvement of hand hygiene was associated with the reduction of healthcare-associated BSI. There is no information about the efficacy of bundle strategies in the European multicultural context and the contribution of such initiatives to the positive outcome was often ill defined.

Objectives
The objective of the CRBSI prevention strategy was to demonstrate the implementation and effectiveness of two interventions in European countries: 1) the WHO hand hygiene promotion strategy, and 2) a CVC bundle. Specifically, the objectives were:
• to build competence among OSIs of participating hospitals in outcome surveillance and performance feedback;
• to empower OSIs to implement defined intervention packages;
• to demonstrate effectiveness of interventions by encouraging local stakeholders to adopt, adapt, and conduct best practice recommendations regarding hand hygiene and CVC insertion.

Methods
Call for tender and selection of hospitals
Hospitals were invited by the local HAI-Net contact points, or if they participated in the European Antimicrobial Resistance Surveillance System. Of 18 interested hospitals, 15 were selected. Among these, three dropped out and were replaced by two other hospitals during the study period. The final number of participating hospitals to be analysed included 14 sites in 11 countries (Figure 3).

Study protocol and online database
The detailed study protocol was established in collaboration with the University of Geneva (UniGE) and included the organisation of three workshops and the prospective surveillance of hand hygiene performance by direct observation and CVC insertion by using a checklist.

Study design
The study followed a cluster randomised, stepped-wedge design. After a baseline period of six months, the hospitals were randomised in batches of three and intervals of three months to one of three arms, a hand hygiene promotion campaign, a CRBSI prevention programme (bundle), or both. Randomisation was done centrally using a computer-generated algorithm. Hospitals were informed about their randomisation between three and six months ahead.

Interventions
The implementation of the two interventions, i.e. the WHO hand hygiene promotion strategy, and an extensive CVC bundle, developed and successfully implemented at UniGE, was facilitated by the organisation of three workshops where local stakeholders (usually an intensive care doctor and/or a study nurse, depending on the intervention) were trained centrally about best practice in CRBSI prevention and hand hygiene. The study included prospective monitoring of compliance with hand hygiene and CVC insertion. Performance feedback was provided from the start of the intervention. Hand hygiene observations and intervention were based on material developed and made available by WHO (http://www.who.int/gpsc/information_centre/en/). Observation of CVC insertion and the CRBSI prevention strategy was based on a successful CRBSI prevention programme established at UniGE. For training purposes, an e-learning programme from the Geneva project was adapted and made publicly available by PROHIBIT (www.carepractice.net). The workshops followed the train-the-trainer concept, which allowed adoption and adaptation of the programme by local stakeholders. The individual implementation strategies were recorded in a logbook. Additionally, it was assessed by the qualitative InDepth study (WP4) in six of the 14 hospitals.

Timeline
- End of September 2010: all participants received the study protocol.
- Kick-off workshop for OSIs and study nurses (15/16 November 2010).
- CRBSI surveillance started in January 2011.
- Training for hand hygiene and insertion bundle compliance (24/25 January 2011).
- Hospitals started to collect data on hand hygiene and CVC insertion compliance in March 2011.
- Intervention training (May 2011, November 2011, and May 2012).
- Wrap-up meeting with presentation of the preliminary results of the intervention study, as well as the qualitative InDepth study (11/12 November 2013).

Figure 3. Centres participating in the CRBSI prevention study


Results
Settings
Fourteen centres participated up to study completion; seven were university-affiliated. Five hospitals had over 50,000 admissions a year, four between 30,000 and 50,000, and the remaining five admitted less than 30,000 patients a year. The number of ICU beds per hospital varied between 10 and 185. Two hospitals with a high number of ICU beds did not participate with all their ICU wards. The median (range) nurse-to-patient ratio was 0.42 (0.25-1) with an unfavourable trend over time in three hospitals.
Process indicators
Hand hygiene compliance
A total of 59,122 hand hygiene opportunities were observed during 6749 sessions. Baseline hand hygiene compliance for all 14 hospitals was 48.6%. Compliance improved significantly in all but two hospitals.

Table 3. Adjusted and unadjusted average hand hygiene compliance between baseline and intervention for the three intervention groups

Baseline
(%) Intervention
(%) Unadjusted change
(%) Adjusted
change
(%) 95% CI*
HH intervention 35.9 58.1 22.2 18.3 15.4 – 21.1
CVC intervention 51.3 62.2 10.9 1.9 -1.2 – 5.0
Both interventions 53.5 62.7 9.2 7.5 4.8 – 10.3
* CI, confidence interval

The overall trend of hand hygiene compliance was negative in the baseline period. A linear regression analysis including baseline and intervention trends are shown in Table 4. During the intervention phase, the slightly negative baseline compliance trend became significantly positive in all intervention groups.

Table 4. Adjusted quarterly trends for hand hygiene compliance for the three intervention groups

Adjusted change (%) 95% CI
Baseline trend (per quarter) -1.4 -2.3 – -0.6
Intervention trend hand hygiene (per quarter) 2.3 1.4 – 3.3
Intervention trend CVC (per quarter) 2.9 1.9 – 3.9
Intervention trend for both (per quarter) 2.2 1.3 – 3.3

All professional categories in hospitals addressing hand hygiene, either alone or in combination with the CRBSI prevention bundle, improved hand hygiene compliance. Improvement was highest in hospitals allocated to the hand hygiene intervention alone. In the CVC hospitals, hand hygiene compliance of nurses improved modestly, but significantly, whereas doctors showed no improvement. On average, doctors (and medical students) had lower hand hygiene compliance than nurses in all centres at baseline and following intervention.


Table 5. Average hand hygiene compliance per professional category during baseline and the intervention phase for all hospitals and per intervention group.

Professional category Baseline
(%) Intervention
(%) Unadjusted
change
(%) 95% CI
Hand hygiene intervention
Nurse/student nurse 37.1 58.2 21.1 18.9 – 23.2
Auxiliaries 30.5 59.0 28.5 23.3 – 33.7
Medical doctors/students 31.9 51.3 19.4 15.5 – 23.2
Other HCWs 45.6 74.9 29.3 21.2 – 37.4
CVC intervention
Nurse/student nurse 51.8 64.0 12.2 10.8 – 13.6
Auxiliaries Not reported because of low numbers
Medical doctors/students 45.1 46.3 1.3 -4.4 – 6.9
Other HCWs 38.7 48.5 9.8 -4.0 – 23.5
Both interventions
Nurse/student nurse 55.9 64.2 8.3 6.4 – 10.1
Auxiliaries 49.5 61.5 12.0 8.3 – 15.7
Medical doctors/students 50.8 57.9 7.1 3.9 – 10.3
Other HCWs 27.8 47.2 19.4 0.8 – 38.1

CVC bundle compliance
A total of 3784 CVC insertions were observed. In a few hospitals, the acquisition of large drapes or alcohol-based chlorhexidine for skin antisepsis was not possible until the end of the study (mostly due to budget restrictions), despite efforts of the onsite investigators. Therefore, the overall bundle compliance was assessed without these two items. The median (range) bundle compliance in hospitals allocated to the CVC bundle alone improved from 6.3% (range 0.0–11.4) at baseline to 63.7% (0.0–73.9) in the intervention. In hospitals allocated to both interventions, compliance improved from 0.8% (range 0.0–91.7) at baseline to 88.9% (1.7– 100.0). In hospitals allocated to hand hygiene alone, bundle compliance remained low (1.5%; range 0.0–12.5) at baseline and 3.3% (0.0–16.9) in the intervention. Bundle improvement correlated with the formal start of the bundle for hospitals allocated to the CVC bundle intervention. Most variations with compliance were observed for patient preparation, the technique of skin antisepsis, and adherence to MSB precautions. Work organisation, catheter fixation, and correct application of the dressing had already a high compliance from the start of the intervention.

CRBSI
In total, the database included 35,831 catheters from 25,348 patients, 262,377 CVC days, and 382 CRBSIs. CRBSI incidence densities (range) at baseline and in the intervention period were 2.4/1000 catheter-days (0.0–10.2) and 0.9/1000 (0.0–3.8) respectively. This reduction was significant (p<0.0001). The individual quarterly rates are shown in Figure 4. Six hospitals had an average baseline incidence density of less than 1.5/1000 CVC days.




Figure 4. Quarterly CRBSI incidence densities per hospital*


* The dots indicate the start of the intervention.

Competing events analysis
By accounting for death or discharge as possible competing events, the effects of the interventions were comparable with that of the Cox proportional hazards.

Cox proportional hazards model
In a multiple regression analysis adjusted for a possible independent effect associated with calendar time (quarter), sex, and age category, including all covariates associated with CRBSI at a p-value level <0.2 in the univariate analysis (type of ICU, duration of ICU stay until insertion, bacteraemia at insertion, insertion department, insertion vein, use of CVC for antibiotics, blood products, dialysis), the hand hygiene only and hand hygiene/CVC bundle interventions were significantly associated with CRBSI reduction (Table 6). The effect of implementing both strategies simultaneously interventions was approximately 50%. The effect in the CVC intervention-only group was lower. Adjusting for patient and CVC characteristics hardly affected the hazard ratios of the interventions.

Table 6. Adjusted hazard ratios of the three intervention groups

Intervention
Hazard ratio 95% CI
Hand hygiene alone 0.49 0.26-0.92
CVC bundle alone 0.63 0.38-1.06
Both interventions 0.49 0.30-0.81


Discussion
This study showed the feasibility and effectiveness of a multimodal intervention in various cultural and economic backgrounds in Europe. CRBSI rates overall decreased while compliance with the two process indicators, hand hygiene and the CVC bundle, increased. “Multimodality” in our study addressed promotion of evidence-based best practice, use of a train-the-trainer approach, leaving room for local adaptation of the implementation strategy, but creating hard edges by prospective observation of compliance with hand hygiene and good CVC insertion practices.
The average baseline CRBSI rate of 2.4/1000 catheter-days is about what has become a standard in high-income countries. However, there was a range from 0 to 10.2/1000 with high rates in some hospitals (but not all) in Latvia, Hungary, and Ireland. Such a variation is in line with the published evidence available at the time of grant submission.13,48,49,51 The fact that hospitals in Latvia, Hungary, and Ireland are among those with high CRBSI baseline rates raises concerns about the role of public healthcare expenditure in quality of care, which was the lowest of all 14 centres in these three countries. The association between healthcare expenditure and ABHR consumption identified in WP3 points in the same direction. Six hospitals had an average baseline rate below 1.5/1000 catheter-days. These hospitals further reduced their rates or remained low.
Implementing the hand hygiene programme alone or in combination with the CVC bundle improved hand hygiene compliance, which is in line with other studies evaluating multimodal hand hygiene promotion strategies.52 Our average baseline compliance of 48.6% was similar to the 52% identified in the Mastering hospital antimicrobial resistance and its spread into the community (MOSAR) project, a European study performed in 13 ICUs.53,54 Compared to the combined intervention and the CVC bundle alone, the group addressing hand hygiene alone had the largest effect in hand hygiene improvement but also the lowest level at baseline. The average increase of hand hygiene compliance in hospitals allocated to both interventions or the CVC bundle alone was moderate.
CVC bundle implementation, alone or in combination with hand hygiene promotion, improved bundle compliance. Some hospitals were not capable of purchasing either alcohol-based chlorhexidine for skin preparation or the large drape due to budget constraints. The hospital that did not purchase the large drape covered the insertion site with three smaller drapes to compensate. One hospital acquired permission to buy large drapes only at the very end of the study. Although some institutions in the group of hospitals allocated to hand hygiene alone improved CVC bundle compliance, the numbers remained lower compared to the two other groups. The slight improvement is most likely due to the observation of CVC insertions. Our study is one of a few initiatives to report compliance with the promoted insertion bundle as a process indicator. The vast majority of quality improvement studies only offer CLABSI or CRBSI outcome data.30,55-65 Previous studies that measured also bundle compliance achieved adherence in the range of 20-37%,66 slightly over 90%,67 and (almost) 100%.68,69 Similar to our findings, the bundle compliance in an Australian multicentre study showed significant variation ranging between 0% and 100%.67
Analysis of the three groups revealed the surprising result that CRBSI rates were significantly reduced only in the group of hospitals allocated to hand hygiene only or both interventions. In the CVC bundle only group, the reduction was close to significance, which is possible linked with the fact that the baseline rates in these hospitals were lower (mean, 1.4 CRSBI/1000 catheter-days) than those anticipated in the power calculation. One of the inclusion criteria was an expected or confirmed baseline rate of two CRBSI/1000 catheter-days or more. However, although all hospitals either proved or assumed such a rate, the measured baseline rate in the study was lower in seven hospitals. In the one hospital allocated to the CRBSI prevention bundle alone and with a baseline incidence >1.5/1000 catheter-days (2.4/1000 catheter-days), CRBSI reduction was significant.
In summary, the interventions were successful in lowering overall CRBSI rates. An average low baseline rate of 2.4/1000 catheter-days could be further reduced in a number of European hospitals with different sociocultural backgrounds. CRBSI reduction was statistically significant in all ICUs with baseline rates above average. ICUs with low baseline rates still further reduced CRBSI or they remained low. The association of hand hygiene with CRBSI reduction points towards the role of this intervention in HAI prevention and demonstrates that CRBSI prevention not only relies on CVC insertion, but also on catheter care. Hand hygiene as a means of HAI prevention was taken up as an intervention, even in the group of hospitals not formally allocated to such an intervention. Compliance with the CRBSI prevention bundle improved only in the groups allocated to a CVC intervention (either alone or both), which assumes that a specific intervention is necessary for such a technical improvement.

Potential Impact:
The overarching objective of PROHIBIT was to provide the vision for policy-makers, managers, and healthcare workers (HCWs) to prevent HAI by improving the understanding of European guidelines and hospital policies and practices, by overcoming common obstacles for implementing evidence-based best practices, and by testing the effectiveness of two interventions to prevent CRBSI. As a consequence and based on the findings of PROHIBIT, we suggest a number of actions to the European Commission (EC) as to how IPC can be made more relevant in Europe, as well as tangible recommendations to improve patient safety.
- Establish a European IPC strategy.
- Establish a European-wide surveillance for outcome and/or process indicators (a prospective initiative or repeated point prevalence surveys as done at present).
- Create and sustain an IPC platform that serves as a European IPC network.
- Consider train-the-trainer workshops for hospital staff on a European level.
- Consider providing dedicated salaries for a limited time period to get IPC programmes off the ground
- Establish an authoritative body for state-of-the-art guideline production.
- Encourage the adoption of evidence-based IPC strategies in European hospitals through authoritative guidelines and prevention projects.
- Improve hand hygiene in hospitals through a European initiative.
- Raise awareness of the importance of IPC among European HCWs, particularly doctors.
- Foster a positive organisational culture and teamwork in European hospitals.
- Encourage (or mandate) hospitals to establish IPC standards of staffing and organisation/structure (IPC goals, indicators, committee).
- Encourage the establishment of link nurses and doctors (or liaisons) at the hospital level to span institutional boundaries and facilitate multidisciplinary collaboration.
- Encourage countries to provide a budget for IPC activities; in particular, ensure that IPC budgets in hospitals are separate.
- Fund high-quality effectiveness research in the field of IPC emphasizing implementation and behaviour change

List of Websites:
www.prohibit.unige.ch