Implementation of central line-associated bloodstream infection prevention bundles in a surgical intensive care unit using peer tutoring

Central line-associated bloodstream infection (CLABSI) is one of serious healthcare-associated infections that cause increased medical costs, morbidity and mortality; however, CLABSIs have been prevented in many developed and developing countries using multifaceted approaches [1–5]. Several guidelines for the prevention of CLABSIs are available, but the core contents of the evidence-based recommendations are shared in common [6, 7]. Although the objectives of the CLABSI prevention guidelines are evident and simple, the implementation of these guidelines in clinical practices requires many factors to be well-coordinated. Heterogeneity in compliance or performance with the guidelines exists worldwide, and interventions have not always been successful [8]. The importance of infection control in healthcare settings for patient safety and quality of care cannot be emphasized enough, but the available resources including expert personnel, reimbursement systems and managerial support are not always sufficient to deal with many active issues in most healthcare facilities. Different strategies for different regional or institutional situations are needed for the successful implementation of CLABSI prevention guidelines.

The education of and feedback from healthcare workers are core components of implementing an intervention program. The education component should be organized in a manner that allows the healthcare workers to collaborate, learn from, and support each other. We used ‘learning by teaching method’-based education to implement CLABSI prevention bundles in a surgical intensive care unit (SICU) with a high CLABSI rate. This peer tutoring approach was intended to motivate the healthcare workers to actively participate in their own workplace problems and to develop a safety culture in the unit through the sharing of a common understanding.

The rate of CLABSIs per 1000 catheter-days decreased from 6.9 infections in the pre-intervention period to 2.4 and 1.8 in the intervention (6 m; P = 0.102) and post-intervention (9 m; P = 0.036) periods, respectively (Table 1 and Fig. 2). The regression model showed a significantly decreasing trend in the infection rate from the pre-intervention period (P < 0.001) with incidence-rate ratios of 0.348 (95% confidence interval [CI], 0.98–1.23) in the intervention period and 0.257 (95% CI, 0.07–0.91) in the post-intervention period. However, after the 9 months post-intervention period, the yearly CLABSI rates reverted to 3.0–5.4 per 1000 catheter-days (Table 1).

Period	Central line-days	Patient-days	Central line utilization ratio	aCLABSI events	bCLABSI rate
Pre-intervention, 9 m	1734	3273	0.53	12	6.9
Intervention, 6 m	1245	2147	0.58	3	2.4
After intervention,
9 m	1684	3207	0.53	3	1.8
1st year (12 m)	1994	3920	0.51	9	4.5
2nd year (12 m)	2030	4137	0.49	6	3.0
3rd year (12 m)	2222	4657	0.48	12	5.4

^a CLABSI central line-associated bloodstream infection

^b CLABSI rate CLABSI events per 1000 central line-days

Adherence to each component in the insertion bundle reached 100% in the 5th month. Adherence to each component in the maintenance bundle reached 100% in the 2nd month. The absolute amount of central line use increased gradually but central line utilization ratio taking patient-days into account has been steadily decreasing since the initial rise during intervention period from 0.58 to 0.48 (Table 1). The places of central line insertion during intervention (6 m) and post-intervention (9 m) periods were operating room (58.6%), interventional radiology (20.8%), SICU (16.4%), emergency room (3.5%) and general ward (0.7%). The awareness of core knowledge about CLABSI prevention practices during the interventional period of 6 months increased, with scores of 15.8 (1st month), 17.1 (3rd month), and 18.9 (6th month) points out of a total of 20 points for the 20-question assessment. The performance of hand hygiene which was monitored weekly were 93.4% (range, 92–96), 89.7% (range, 79–97) and 90.9% (range, 83–96) during the pre-intervention, intervention and post-intervention periods, respectively.

Our intervention to reduce CLABSIs by implementing CLABSI prevention bundles using peer tutoring was effective during the interventional period of 6 months and a post-interventional period of approximately 1 year. However, without continuous active interactions and dominating internal governance, the virtuous reduction of CLABSIs was not sustained. Our educational method had several advantages. Each healthcare worker had the opportunity to voice her/his own work-related problems, to hear from other colleagues, to understand the principles behind their routine activities and to receive responses to their problems. These factors minimized resistance to the introduction of a new job pattern. The purpose of the educational sessions was to motivate the healthcare workers, to provide them with expert knowledge for practical use and to improve knowledge retention over longer time periods. Determining the most effective educational method has been an area of utmost interest for a long time. Generally, active participatory education is more effective than passive learning. ‘Learning by teaching’ or peer tutoring is one of the active participatory educational methods in which the core idea is to have a pair or group of students teaching the majority of topics to their classmates in a way that encourages their classmates’ active participation and communication in the best possible way [14]. ‘Learning Pyramids’ have shown that learners can retain approximately 90% of the content of a subject matter when they teach the material to someone else.

The commitment of the infection control staffs to coordinate the overall program and the presence of governing leadership in the unit to maintain a level of practice quality were basic components of the intervention. Regardless of the strong positive aspects of our educational approach, the intervention failed to sustain a reduced CLABSI rate beyond 9 months post-intervention. Several complex factors seem to be responsible for this lack of sustainability. High job turnover in the unit coupled with the lack of a continuing education system for new nurses and doctors after the intervention period likely weakened the effectiveness of the prevention protocols. The turnover rates of nurses in the SICU were 24.9% and 59.5% due to resignation, leave of absence and rotation during intervention (6 m) and post-intervention (9 m) periods. Thereafter, the rate was 39.1% - 46.8% annually for 3 years. Interns or residents had a rotational program every 1 or 2 months among 3 affiliated family hospitals. The patient-to-nurse ratio of 1:3 was much higher than the previously reported rate of 1:2 which was a risk factor for CLABSIs [15]. Moreover, there was no dedicated intensivist in the SICU who could have led and maintained the program at the hospital.

Successful large-scale interventions with sustained reductions in CLABSIs have been reported worldwide [2, 3, 5, 16, 17]. In South Korea, there have been a few single center or small-scale trials to reduce CLABSIs [18–21]. The only recent multicenter interventional study involving 58 ICUs in 26 hospitals funded by the Korea Centers for Disease Control and Prevention resulted in a CLABSI rate of 2.23 infections per 1000 catheter-days, which did not significantly improve the pre-interventional rate of 2.09 infections per 1000 catheter-days [22]. There have been no reports about sustaining a reduced CLABSI rate in South Korea. The distribution of CLABSI pathogens in the pre-interventional period was similar to recent Korean National Healthcare-associated Infections Surveillance System (KONIS) data from 166 ICUs, which showed A. baumannii (14.6%) as second most common pathogen of CLABSI in 2013 [23]. The high number of A. baumannii infections in our study might reflect high A. baumannii infection/colonization in the SICU. Regarding relatively higher frequency of candida infection during the intervention and the immediate post-intervention periods, there were no changes of antibiotics prophylaxis or strategies of treatment of surgical site infection. The only systematic change was a use of 2% chlorhexidine tincture for skin disinfection before catheter insertion and catheter site dressing instead of 10% povidone-iodine as a CLABSI bundle component. Whether the chlorhexidine had weaker effect on candida that on other bacterial pathogens requires further data due to the small absolute number of candida infections in our study, but previous studies about CLABSI have not supported the role of chlorhexidine in candida infections.

There are diverse institutional situations and cultures that deal with infection control programs and patient safety issues. Although standard CLABSI prevention guidelines are well-known, their translation into clinical practice needs to be individualized according to regional or institutional feasibility. Our educational method was efficient in the initial implementation of the program and effectively reduced the CLABSI rate for a short time period of approximately one year. To maintain the reduced rate, a multifaceted integrative approach must be continued [15]. Intra-departmental or intra-unit effort hardly maintained the effects of the program over a longer time period. Our approach may not be effective in certain situations, but it deserves to be considered by institutions implementing a CLABSI prevention program for the first time, especially in resource-limited settings.

Our study has some limitations. First, as this was a single-center study, the effectiveness of our approach may not be generalizable. However, one institutional success may provide a good model for generalization, as has been shown previously [1, 24]. Second, we did not perform a comparative study to prove the effectiveness of the peer tutoring method. However, from a practical point of view, it was a useful tool to guide the program as it received the cooperation of the participants and improved the understanding between healthcare workers. Third, the indirect educational approach implemented for doctors in the SICU might have weakened the effect of the intervention. As residents from each clinical department who were mainly involved in the insertion procedure in the SICU rotated at 1 or 2 months interval among 3 family hospitals, integrative successive group education and follow-up assessment were impractical. And maintenance bundles were mainly related to the nurses. So, we decided to educate indirectly the doctors through preceptors in each department who had regular meetings with infection control office. The doctors had individual feedbacks on the spot in the SICU for their performance. Fourth, the performance of insertion bundle might be incomplete. Operating room was the most frequent place of central line insertion, but fully adopted the insertion bundle only after intervention period. They had been using a similar protocol except for skin disinfection with 10% povidone-iodine and a small sized drape. Emergency room (3.5%) and general ward (0.7%) had no insertion bundle though the number was small.

A peer tutoring educational method was useful for the implementation of CLABSI prevention bundles in a SICU and effectively reduced infection rates for a short time period of approximately 1 year. This approach is applicable for hospitals with limited resources that are trying to initiate prevention bundles. However, to maintain the reduced CLABSI rates, resources support and multifaceted cooperative approaches may be essential.