The implementation strategy encompasses a multifaceted program tailor made for this specific hospital and ready for use. We hypothesize that implementation of this program will result in fewer CA-BSIs and improved adherence to CVC bundles [5, 22, 24, 37]. Tailored implementation strategies – i.e. based on content analysis of barriers and facilitators – seem to be more effective than non-tailored strategies [38, 39]. The Pronovost-model effectively helped to design strategies from the first to the last phase.
Various strategies will be employed. First, education by means of a mini-symposium, workshops, presentations, and so on. Second, improving staff intrinsic motivation, e.g. by audit, feedback, and reminders. Third, organisational change: preparing on single protocol for all patients. Fourth, use of ICT: e.g. a timer on the computer screen at the bedside to clock the 30 seconds air drying time, online access to guidelines and the CVC infection prevention manual, use of screen savers, and daily reminder informing whether the CVC is still needed. Fifth, feedback: quarterly reporting of the CA-BSI incidence and the rates of adherence to the insertion and maintenance bundles. Sixth, awareness: time out procedure and daily goal sheet. Seventh, engagement of the managing staff: showing their commitment to the aim.
The Pronovost-model recommends enlisting all local stakeholders involved in patient care and discussing with them potential barriers and facilitators to adherence to the developed protocol. The Pronovost-model does not provide for eliciting support from senior management of the hospital. Nevertheless, they will be asked to demonstrate their commitment and assume ownership of the general aim to reduce CA-BSIs hospital-wide . We will explicitly designate senior management as ambassadors of the goals  by publishing their one-liners and portrait pictures. These ambassadors should convince all healthcare workers that safer CVC care is an important goal and make clear they support the campaign. The senior management is officially accountable for patient safety and even may act as role models.
Use of opinion leaders will be added to the implementation strategy. Clinical and senior management need to show their vision and clearly dissimilate this particular aspect of safety culture. Senior management of effective infection prevention programs dissimilated their success as improved clinical excellence and inspired their staff . Furthermore, senior management will help resolve organizational and financial barriers and practically support initiatives [25, 26].
The Pronovost-model describes the implementation process in broad terms; development of a fitting protocol or work-instruction is not included . We will therefore add a comprehensive description of how to develop a work-instruction. Evidence from recently published studies should support the newly developed protocol and providing evidence will perfectly fit into phase 1 of the Pronovost-model; building evidence in favour for the chosen intervention.
Feedback is a widely used, powerful measure to increase adherence to infection prevention measures such as good hand hygiene [24, 41]. However, once feedback is stopped any unwanted behaviour could come up again [24, 42]. On the other hand, providing continuous feedback is time consuming and therefore not realistic. Feedback should be used firstly to alter initial unwanted behaviour and this should ideally move into desired behaviour as an intrinsic driven and well-conditioned behaviour.
Education is often used to support behavioural change, especially if flaws in knowledge are observed [5, 40, 43]. A complicating factor is that level of knowledge varies among healthcare workers categories and within categories. This should be borne in mind when developing a hospital-wide education program. Physicians may tend to appreciate knowledge more than do nurses, and evidence-based education could be very useful to promote physicians’ desired behaviour . On the other hand, a wash-out effect is often observed [45, 46]. This means that knowledge previously received may recede to the background. Ideas on what is effective in infection prevention are developing over time, so regular updates are essential. In addition, repetition of education programs is necessary in a teaching hospital like ours; many healthcare workers are in training and leave the hospital after having completed their education program.
The Pronovost-model is merely medical oriented. The main point of departure from the model is knowledge transfer, which fits into physicians’ learning style . Goossens et al. found that strong scientific evidence was the strongest determinant of physicians’ behaviour . However, regarding our goal, a clear healthcare team angle seems to be more appropriate for a broad dissimilation of improved hygienic behaviour among all members of the multidisciplinary team. Nurses have a more active learning style, and the strongest determinant of knowledge acquisition was found the be the fact whether the subject was 'interesting […] or not’ . This phenomenon affects whether a new procedure is potentially beneficial for patients or gets embedded in the daily care.
We postulate that the proposed study has methodological strength because it is guided by a validated implementation model that has been translated into a hands-on program and is described in detail for implementation hospital-wide. Furthermore, regularly reporting the outcomes is in line with the ORION statement promoting transparent reporting on intervention studies aimed to reduce nosocomial infections. ORION recommended interrupted time series as preferable method for analysis showing the change in results over time.
We anticipate several challenges in this study. (1) Effectiveness of interventions is preferably evaluated by randomisation into an intervention and control group. However, this method is inappropriate for our aim to implement an intervention hospital-wide, and the interrupted time series design is the second best solution and in line with the ORION guidelines (2). The CVC bundle is adapted and tailor made for our hospital. Although this results in less generalizability, this naturalistic approach could help develop practical implementation strategies for other hospitals or other interventions. (3) The effective ingredients of the bundle are still unclear. However, as some interventions have been shown to be effective it would be unethical to test all separate interventions. (4) To control for confounders is a challenge due to the different safety climates in the different departments. By establishing a clear leadership we will try to show the benefits of a univocal approach towards infants’ CVC care.
Nevertheless, this detailed implementation strategy of the CVC bundle has a potential to effectively modify healthcare workers behaviour and reduce the number of CA-BSIs hospital-wide.