Patients with end-stage renal disease (ESRD) receiving hemodialysis are at high risk for complications from infection, especially bloodstream infections. In 2008, an estimated 37,000 cases of central line–associated bloodstream infection occurred in patients receiving outpatient dialysis in the United States. 1 , Reference Patel, Kallen and Arduino 2 Bloodstream infections are major causes of hospitalization and death in chronic dialysis patients. 3 Thus, infection prevention practices in dialysis facilities are critical to improving quality of life and survival in this high-risk population as well as reducing use of costly resources.
Despite multiple guidelines outlining infection prevention practices for ESRD patients, 4 – 8 many of the practices remain incompletely embraced by the dialysis community. The high use of central venous catheters (CVCs) for dialysis remains a huge concern in this context, with most patients presenting for initiation of chronic hemodialysis with a catheter as their initial vascular access. Moreover, practices related to preparation and maintenance of vascular accesses (eg, exit site care, scrubbing access hubs) vary considerably.
The National Opportunity to Improve Infection Control in ESRD (NOTICE) initiative was designed to assess recommended infection control practices at US dialysis facilities with the aim to identify variation in practices that would inform future quality improvement initiatives. The aim was accomplished through the creation of a tool that enabled unit staff, facility leadership, and external evaluators to assess the extent to which required and best practices in infection control were being followed.
METHODS
The NOTICE project partners included individuals from the Agency for Healthcare Research and Quality, the Health Research and Educational Trust, University of Michigan Kidney Epidemiology and Cost Center, Renal Network of the Upper Midwest (Network 11), TB Spencer Consulting, and a dialysis facility surveyor.
Infection Control Worksheet Development
Using information gathered from an extensive literature review of best and required infection control practices, the NOTICE team created an infection control worksheet (ICWS). The ICWS contained 73 distinct infection control facility practices in 8 separate area-specific checklists: treatment initiation with CVC access; CVC exit site care; treatment initiation with arteriovenous fistula or graft access; parenteral medication preparation and administration; treatment termination with CVC access; treatment termination with arteriovenous fistula or graft access; cleaning and disinfection of the dialysis station; and supply management and contamination prevention. The ICWS contained 2 versions of the checklists: one intended for use by facility supervisory staff and auditors when observing staff practices and the other intended for use by direct care staff at the dialysis station. The ICWS was vetted by a technical expert panel composed of experts in dialysis and infection control, as well as representatives from the Centers for Disease Control and Prevention and the Centers for Medicare and Medicaid Services survey and certification.
Dialysis Facilities
There were approximately 6,000 dialysis facilities in the continental United States as reported in the 2009 Dialysis Facility Reports (www.dialysisreports.org). From those, we identified 3,078 facilities with more than 29 hemodialysis patients and having vascular access–related infection rates higher than 12.5%. From this subset, facilities were randomly selected from ESRD Networks 6, 11, 15, and 17 to participate in the NOTICE study. Facility selection was stratified into 12 strata on the basis of infection rate, dialysis organization affiliation, and socioeconomic status (median family income in zip code according to the 2000 Census). The stratification categories were defined as follows.
-
∙ Dialysis organization affiliation: defined as large dialysis organizations (LDO; 66.8%), small dialysis organizations (21.6%), and independent (11.6%). The LDO group is composed of the 2 largest LDOs in the country.
-
∙ Vascular access–related infection rate: higher or lower than 21.35% infection rate (overall median after removing the lowest 25%).
-
∙ Median income: higher or lower than $37,283 (overall median for facilities after exclusions) as a measure of patient socioeconomic status. Data were obtained from the 2000 census by facility zip code.
Eighty-six randomly selected facilities were invited to participate, with a target enrollment of 40 facilities. Thirty-four facilities were enrolled, 8 from Network 6 (GA, NC, SC), 9 from Network 11 (MI, MN, ND, SD, WI), 9 from Network 15 (AZ, CO, NV, NM, UT, WY) and 8 facilities in Northern California, from Network 17 (Northern CA, HI, Pacific Islands).
Infection Control Evaluator Observations
Study participation included an on-site evaluation by an infection control evaluator (ICE) using the 8 ICWS checklists, from October 1, 2011, through January 31, 2012. Six evaluators with prior dialysis and/or surveyor experience received prospective training in use of the checklists and performed all observations. ICEs observed infection control practices using each checklist twice at the participating facilities. The observations were conducted during dialysis treatment for 2 to 4 patients dialyzed through a CVC, and 2 to 4 patients dialyzed through an arteriovenous fistula or graft at each facility. Medication preparation and administration practices were also observed. Each facility received 2 observations for each of the 73 items on the ICWS. For each item and for each patient observed, the ICE recorded whether or not the particular infection control practice was appropriately conducted; in addition, extensive notes regarding observations were recorded.
Statistical Analysis
Results from the 73 individual items on the ICWS were combined to represent composite measures of adherence. Individual items related to hand hygiene (HH) were also combined into an overall measure of HH, a composite measure of HH items from each checklist occurring after set-up but before contact with the patient (HH during), and a similar measure of HH after contact with the patient (HH after). Analyses assessing ICWS results and relationships between ICWS information were performed using SAS, version 9.2 (SAS Institute).
RESULTS
There was considerable variation in infection control practices across enrolled facilities. Facilities were categorized into 1 of 3 groups: independent, LDO, and small dialysis organization. Table 1 illustrates the variation in performance of selected infection prevention items by dialysis organization affiliation. Individual items shown were selected from the 73 possible items for having less than 50% overall adherence to evaluated practices. There were few individual items that were related to dialysis organization and, in particular, performance of HH did not vary across dialysis organization. Nonetheless, there were interesting observed differences in practice among these subgroups. Use of antimicrobial ointment was infrequent, 17% overall, but ranged from 0% in LDO facilities to 43% in small dialysis organization facilities. Likewise, chlorhexidine was used for exit site care 19% of the time, but varied from 35% in independent facilities to 0% in LDO facilities.
TABLE 1 Percentage Performance of Selected Infection Prevention Practices at National Opportunity to Improve Infection Control in ESRD (End-Stage Renal Disease) (NOTICE) Facilities by Dialysis Organization Affiliation
Note. CVC, central venous catheter; ESRD, end-stage renal disease; ICWS, infection control worksheet; LDO, large dialysis organization; SDO, small dialysis organization.
Adherence to HH (mean, 58%) was lower during a procedure (after potential contamination of hands and before initiation of critical steps within the procedure), compared with either before or after a procedure (Figure 1). HH was least likely to occur (35%) during CVC exit site care, after removing contaminated dressing and before placement of clean dressing. Adherence to HH after completion of a procedure was 79% on average overall. Overall adherence to infection control practices was not consistently different in analyses individually stratified by facility size, location (urban/rural), or median income.
FIGURE 1 Adherence to Hand Hygiene Opportunities, Before, During and After Selected Procedures in NOTICE Facilities (n=68)
Figure 2 illustrates the variation in perfect performance on each of the ICWS checklists. For each observation, perfect performance is characterized by correctly completing proper care practices for every item on the checklist. The 8 checklists varied in the frequency of perfect performance from 0% for meeting every item on the checklist for disinfection practices to 22% on the arteriovenous access practices at treatment initiation.
FIGURE 2 Overall Percentage of Perfect Performance of Checklist Items by Checklist in NOTICE Facilities (n=68)
DISCUSSION
Despite advances in dialysis-related technology and documented improvements in many processes of care (eg, small solute clearance, anemia management) over the past decade, the US chronic dialysis population continues to have poor outcomes compared with chronic dialysis patients in other industrialized nations, even after adjustment for age and comorbidity.Reference Pisoni, Arrington and Albert 9 In addition, US chronic dialysis patients are hospitalized at alarming rates. In 2007, 12% of US dialysis patients were hospitalized owing to septicemia and 16% of all deaths in dialysis patients were attributable to an infectious cause. 10 This unacceptably high infection-related morbidity and mortality has been attributed to multiple factors, including underlying patient comorbidity, an immunocompromised state related to inflammation, malnutrition, the invariable presence of fluid overload, the risk of exposure to infectious pathogens during dialysis and the administration of parenteral medications, and, for hemodialysis patients in particular, the type of vascular access. Goals for phase I of the NOTICE project presented here included development of procedural and audit checklists that summarized consensus infection prevention practices, focused especially on vascular access–related infection, in the chronic dialysis care setting.
Several noteworthy findings emerged through this study. First, there is substantial room for improvement in infection prevention practices in hemodialysis facilities. Performance of individual infection prevention practices, as observed by the ICE, occurred less than half of the time in NOTICE facilities. Second, there was considerable variation for some infection prevention practices, such as use of chlorhexidine for CVC exit site care or use of antimicrobial ointment with CVC dressing change; this was especially the case for facilities affiliated with LDOs. A potential explanation might be that facilities may follow organization-specific policies that are not consistent with all of the consensus best practices examined by the ICE. Finally, although HH was consistently performed before and after procedures, HH required in the middle of a procedure, for example, after supplies are gathered but before critical steps in dialysis procedures, was completed much less frequently and was as low as 35% during CVC exit site care. These results could overestimate the general level of adherence in US dialysis facilities as a group because of our decision to exclude the quartile of dialysis facilities with the lowest historical infection rates from the selection process.
However, previous studies support our findings of inadequate adherence to HH in hemodialysis centers. In a survey of staff members from a representative sample of 45 US dialysis facilities, 36% of staff reported always following recommended HH and glove use practices.Reference Shimokura, Weber, Miller, Wurtzel and Alter 11 Another study of 53 Medicare-approved US dialysis facilities (2,933 patients) used protocol on-site surveys that included direct observation, patient interview, and medical record review and blood testing for hepatitis C. HH adherence was found to be less than in our study with a frequency of less than 33% in more than 75% of patients. An independent correlation was found between observed infection control practices and relative risk of hepatitis C.Reference Shimokura, Chai and Weber 12
Use of infection prevention “bundles” and participation in collaboratives have resulted in improved outcomes for infection events in other healthcare settings.Reference Pronovost, Needham and Berenholtz 13 , Reference Fakih, Watson and Greene 14 The National Healthcare Safety Network has provided ESRD facilities with tools to analyze their own data so that they can monitor trends, evaluate needs for prevention, and measure the impact of their prevention efforts.Reference Klevens, Edwards and Andrus 15 In addition, collaboratives and bundles have shown significant initial success in decreasing central line–associated bloodstream infections in ESRD facilities. 16 , Reference Patel, Yi and Booth 17 Findings from the initial phase of the NOTICE study are well suited for incorporation into a complementary collaborative initiative for reduction of infections in hemodialysis centers.
There were several limitations to this study. First, this is an observational study; resource constraints precluded a controlled study design. Second, the ICE observations were limited to only 2 observations of each of the 8 ICWS checklists at each facility. Thus, there is a clear possibility that infection control practices at the facility level may not be fully reflected in our observations. Finally, the ICE observers were not masked and facility staff were aware of their presence, so provider behaviors may have been altered during ICE visits. However, it would be expected that the presence of observers would result in improved adherence to facility policies and procedures, so our findings may well represent a best-case scenario.
On the basis of our findings, HH, scrubbing the CVC hub, use of antibiotic ointment, and use of chlorhexidine site disinfection showed substantial opportunities for improvement at ESRD facilities. We recommend further investigation into the effectiveness of consistent adherence to these variables for the prevention of bloodstream infections and vascular access–related infections in ESRD patients. Phase II of the NOTICE project will use knowledge gained in this study to inform quality improvement interventions for prevention of infections in chronic outpatient dialysis facilities.
Acknowledgments
Financial support. The Agency for Healthcare Research and Quality and the Health Research Educational Trust (contract HHSA290200600022I, “Improving Infection Control Practices in End-Stage Renal Disease [ESRD] Facilities”).
Potential conflicts of interest. All authors report no conflicts of interest relevant to this article.
Disclaimer. The opinions expressed are those of the authors and do not reflect the official position of the Agency for Healthcare Research and Quality or of the US Department of Health and Human Services.
