- Email: [email protected]
Impact of a program to prevent central line-associated bloodstream infection in the zero tolerance era
Impact of a program to prevent central line-associated bloodstream infection in the zero tolerance era Alexandre R. Marra, MD,a Ruy Guilherme Rodrigues Cal, MD,a Marcelino Souza Dura˜o, MD,a,b Luci Correa, MD,c Luciana Reis Guastelli, RN,a Denis Faria Moura Jr, RN,a Michael B. Edmond, MD, MPH, MPA,d and Oscar Fernando Pava˜o dos Santos, MDa,b Sa˜o Paulo, Brazil, and Richmond, Virginia
Background: Central line-associated bloodstream infection (CLABSI) is one of the most important health care-associated infections in the critical care setting. Methods: A quasiexperimental study involving multiple interventions to reduce the incidence of CLABSI was conducted in a medical-surgical intensive care unit (ICU) and in 2 step-down units (SDUs). From March 2005 to March 2007 (phase 1 [P1]), some Centers for Disease Control and Prevention evidence-based practices were implemented. From April 2007 to April 2009 (P2), we intervened in these processes at the same time that performance monitoring was occurring at the bedside, and we implemented the Institute for Healthcare Improvement central line bundle for all ICU and SDU patients requiring central venous lines. Results: The mean incidence density of CLABSI per 1000 catheter-days in the ICU was 6.4 in phase 1 and 3.2 in phase 2, P , .001. The mean incidence density of CLABSI per 1000 catheter-days in the SDUs was 4.1 in phase 1 and 1.6 in phase 2, P 5 .005. Conclusion: These results suggest that reducing CLABSI rates in an ICU setting is a complex process that involves multiple performance measures and interventions that can also be applied to SDU settings. Key Words: Bloodstream infection; prevention; intensive care unit; step-down unit; zero tolerance; central venous catheter; epidemiology; central line bundle. Copyright ª 2010 by the Association for Professionals in Infection Control and Epidemiology, Inc. Published by Elsevier Inc. All rights reserved. (Am J Infect Control 2010;38:434-9.)
Central line-associated bloodstream infection (CLABSI) is one of the most important health careassociated infections in the critical care setting.1 These infections are associated with high mortality rates in adult intensive care patients2 and are the leading cause of death among health care-associated infections.3 Thus, CLABSI is an important target for interventions. The Centers for Disease Control and Prevention (CDC) and the Society for Healthcare Epidemiology of America, Inc (SHEA)/Infectious Disease Society of America guidelines describe many different
From the Intensive Care Unit, Hospital Israelita Albert Einstein, Sa˜o Paulo, Brazila; Division of Nephrology, Universidade Federal de Sa˜o Paulo, Sa˜o Paulo, Brazil (UNIFESP-EPM)b; Infection Control Unit, Hospital Israelita Albert Einstein, Sa˜o Paulo, Brazilc; and Department of Internal Medicine, Virginia Commonwealth University School of Medicine, Richmond, VA.d Address correspondence to Alexandre R. Marra, MD, Av Albert Einstein, 627/701-58 Andar-Bloco B, Sa˜o Paulo, Brazil, CEP 05651-901. E-mail: [email protected].
Conflicts of interest: None to report. 0196-6553/$36.00 Copyright ª 2010 by the Association for Professionals in Infection Control and Epidemiology, Inc. Published by Elsevier Inc. All rights reserved. doi:10.1016/j.ajic.2009.11.012
434
interventions for preventing CLABSI in intensive care units (ICUs).3,4 Each of these recommendations is categorized on the basis of existing scientific evidence, theoretical rationale, applicability, and potential economic impact. As part of the 5 Million Lives campaign, endorsed by leading US agencies and professional societies, The Institute for Healthcare Improvement (IHI) (http://www.ihi.org/ihi) recommends that all ICUs implement a central line bundle to attempt to reduce the incidence of CLABSI to zero.5 However, in the ICU, central venous access might be needed for extended periods of time; patients can be colonized with hospital-acquired organisms; and the catheter may be manipulated multiple times per day for the administration of fluids, drugs, and blood products. Moreover, some catheters must be inserted in urgent situations when optimal attention to aseptic technique may not be feasible. Certain catheters (eg, pulmonary artery catheters and hemodialysis catheters) can be accessed multiple times per day for hemodynamic measurements or for hemodialysis, increasing the potential for contamination and subsequent clinical infection.3,4 Recent data suggest that non-ICU medical wards have device utilization rates that are considerably lower than those of medical ICUs, but CLABSI rates were similar to those in medical ICUs.6
www.ajicjournal.org Vol. 38 No. 6
Unfortunately, few data exist about the prevention of CLABSI in step-down units (SDUs). Types of organisms that most commonly cause hospital-acquired bloodstream infections (BSIs) change over time, but gram-positive organisms are responsible for the great majority of BSI cases, principally coagulase-negative staphylococci.7 The purpose of this prospective, quality improvement study was to examine the effect of a series of interventions, implemented in an ICU and SDUs to reduce the incidence of CLABSI and to analyze the differences in CLABSIs and microorganisms in the 2 study phases.
METHODS Setting and study design A quasiexperimental study was conducted in a 38bed medical-surgical ICU and in two 20-bed SDUs with the same physical layout in a tertiary care, private hospital in Sa˜o Paulo, Brazil. The ICU has an open staffing model, and approximately 2200 patients are admitted annually. In the SDUs, all the rooms are single bed rooms. The SDU patients are transferred from the medical-surgical ICU of this hospital, from wards, and from the emergency department. Because this study was considered a quality improvement project, it was not submitted to our Institutional Review Board. The study was carried out in 2 phases: phase 1 (March 2005 to March 2007) and phase 2 (April 2007 to April 2009). In phase 1, attending physicians (eg, surgery doctors), ICU doctors, or ICU medical residents inserted the catheters through a new venipuncture, into the subclavian, jugular, or femoral vein using full sterilebarrier precautions and 2% chlorhexidine preparation for skin antisepsis per CDC guidelines.3 The decision to remove the catheter was made solely by the patient’s physician, who kept the catheter in place until it was no longer needed or until an adverse event necessitated its removal. Catheters were not routinely replaced. Each year, a convenience sample of patients was chosen for whom catheter insertion and catheter dressings were directly observed by assigned nurses. We then provided feedback via e-mail on compliance with these processes for the ICU team (doctors and nurses). In phase 2, after the hospital’s chief executive officer articulated zero tolerance for CLABSI, we continued intervening in these processes as in phase 1; however, these process measures were audited once monthly at random intervals in a small sample of patients undergoing central line insertion. In April 2007, we implemented the IHI central line bundle8 (creation of a central catheter insertion cart; hand hygiene; maximal barrier precautions for insertion; chlorhexidine skin antisepsis; optimal catheter site selection, with
Marra et al.
435
avoidance of the femoral vein for central venous access in adult patients; daily review of line necessity with prompt removal of unnecessary lines) for all ICU and SDU patients requiring central venous lines. This central line bundle was monitored everyday by ICU and SDU nurses and doctors. The ICU and SDU nurses intervened in this process at the same time that performance monitoring was occurring at the bedside if noncompliance with an element of the bundle was detected (eg, hand hygiene was not performed). Prior to the start of phase 2, we delivered a brief presentation to the ICU staff on BSI prevention, reviewed the study protocol, and encouraged participation in our ‘‘central line bundle—getting to zero’’ program. Every month, we provided feedback on compliance with the bundle components via e-mail for the ICU and SDU team (doctors and nurses). In addition, we provided posters in the ICU and SDU with bar graphs displaying compliance with process of care measures. In these posters, we also provided the CLABSI rate determined by surveillance conducted by the infection control and hospital epidemiology program. We implemented an ICU doctor’s group to remove the unnecessary catheters everyday. Once a day, an ICU doctor (not on clinical duty) checked all the ICU patients with central lines and asked the ICU doctors (on duty) if each central line was necessary. The same strategy was performed in the SDUs, except in these units there were nurses who questioned SDU doctors (on duty) about central line necessity. In phase 2, the same infection control practices were performed. We did not use impregnated catheters or chlorhexidine dressings for central lines in ICU and SDU patients. Our CLABSI rates did not include central and peripheral arterial catheter infections. Compliance with all process measures while inserting the central venous catheter (CVC) was evaluated for the CVCs that were placed in the ICU and in the SDUs. All percutaneous CVCs for intravenous fluid, medication, dialysis, or administration of total parenteral nutrition were included. We have used peripherally inserted central catheter lines in these units since 2003. The peripherally inserted central catheters are placed by a hospital nurse team that is specialized in this procedure.
Definitions CLABSI surveillance was performed by trained infection control professionals using the CDC definition9 for the 2 phases of the study for laboratory-confirmed central venous BSI. A CLABSI was considered to be related to a specific unit if detected at least 48 hours after admission to or less than 48 hours after discharge from the unit.3 The device utilization ratio was defined as
436
Marra et al.
American Journal of Infection Control August 2010
Table 1. Characteristics of both phases in the ICU and in the SDUs during the study periods ICU
SDUs
P1 Patient-days CVC-days CVC utilization ratio* CVC check-list, n (%) Compliance with all process measures while inserting the CVC, n (%)z Hand hygiene Full barrier precaution Chlorhexidine skin preparation Sterile field (at least 2) Not necessary to stop the procedure Sterile dressing Compliance with CVC site care, n (%) Site of catheter insertion Adequate dressing CVC type inserted, n (%) Double lumen Hemodialysis catheter PICC Central venous oximetry catheter Swan-Ganz BSI rate per 1000 catheter-days, mean (SD)
P2
P1
20,046 15,650 0.77 -
20,924 14,111 0.67 1833/2392 (76.6) 1723/1833 (94.0)
29,324 11,252 0.38 -
-
1820/1833 (99.3) 1815/1833 (99.0) 1810/1833 (98.7) 1786/1833 (97.4) 1814/1833 (99.0) 1799/1833 (98.1)
-
123/139 (88.4) 127/139 (91.4) 6.4 (2.9)
-
41/43 (95.3) 33/43 (76.7) 4.1 (3.3)
612/697 (87.8)y 2392 (100) 1580 (66.0) 423 (17.7) 172 (7.2) 141 (5.9) 76 (3.2) 3.2 (2.5)
P2 28,871 10,369 0.35 176/232 (75.9)y 161/176 (91.5)y 156/176 (88.6)y 155/176 (88.1)y 159/176 (90.3)y 158/176 (89.8)y 176/176 (100)y 176/176 (100)y 1034/1085 (95.3)y NA NA NA NA 1.6 (2.4)
BSI, bloodstream infection; CVC, central venous catheter; NA, not available; P1, phase 1: March 2005 to March 2007; P2, phase 2: April 2007 to April 2009; SD, standard deviation. *Defined as the number of CVC-days divided by the number of patient-days. y From January 2008 to April 2009. z Compliance is defined as the percentage of time the required component was performed without intervention by the bedside nurse.
the number of catheter-days divided by the number of patient-days.
Microbiologic methods All isolates were identified by manual or automated methods and confirmed using the Vitek 2 system (bioMerieux Vitek, Inc., Hazelwood, MO).
Statistical analysis Comparisons between phase 1 versus phase 2 for compliance with process measures were performed using the x2 test. Continuous variables were compared using the Student t test for normally distributed variables and the Mann-Whitney U test for non-normally distributed variables. All tests of statistical significance were 2-sided with a significance level set at .05. All the data analyses were performed using SPSS for Windows 14.0 (SPSS Inc., Chicago, IL).
RESULTS Study sample, compliance, and incidence density of CLABSI in each phase In ICU phase 1, there were 20,046 patient-days and 15,650 CVC-days, and the CVC utilization ratio was 0.77
(Table 1). In ICU phase 2, there were 20,924 patientdays and 14,111 CVC-days and a 0.67 CVC utilization ratio. In SDU phase 1, there were 29,324 patient-days and 11,252 CVC-days and a 0.38 CVC utilization ratio and, in SDU phase 2, 28,871 patient-days and 10,369 patientdays and a 0.35 CVC utilization ratio. Compliance with process measures for catheter insertion and site care (site of catheter insertion and adequate dressing) is shown in Table 1. The mean incidence density of CVC-BSI per 1000 catheter-days in the ICU was 6.4 in phase 1 and 3.2 in phase 2, P , .001. The mean incidence density of CVC-BSI per 1000 catheter-days in the SDUs was 4.1 in phase 1 and 1.6 in phase 2, P 5 .005. The CVC type inserted is also shown in Table 1.
Microbiologic features As seen in Table 2, 42.1% of all the microorganisms identified in ICU phase 1 were gram positive, followed by gram-negative organisms (38.9%) and fungi (19.0%). In ICU phase 2, 42.6% of microorganisms identified were gram positive, followed by gramnegative organisms (38.1%) and fungi (19.1%). In SDU phase 1, the same proportion (43.6%) of microorganisms was gram positive and gram negative, and a small proportion were fungi (12.8%). In SDU phase 2,
www.ajicjournal.org Vol. 38 No. 6
Marra et al.
437
Table 2. Characteristics of pathogens causing BSI associated to CVC during the 2 study phases in ICU and SDU patients ICU
SDUs
P1
P2
P1
P2
Pathogen
n
%
n
%
n
%
n
%
Gram positive CNS Enterococcus faecalis Enterococcus faecium Lactobacillus gasseri Staphylococcus aureus Gram negative Acinetobacter baumannii Acinetobacter calcoaceticus Burkholderia cepacia Enterobacter cloacae Escherichia coli Klebsiella pneumoniae Morganella morganii Proteus mirabilis Pseudomonas aeruginosa Pseudomonas putida Serrratia marcescens Stenotrophomonas maltophilia Fungi Candida albicans Candida glabrata Candida krusei Candida parapsilosis Candida tropicalis Saccharomyces cerevisiae Scytalidium dimidiatum Trichosporon asahii Total identified agents
40 26 5 1 8 37 12 2 4 2 10 1 3 1 2 18 5 2 4 6 1 95
100 65.0 12.5 2.5 20.0 100 32.4 5.4 10.8 5.4 27.0 2.7 8.1 2.7 5.4 100 27.8 11.1 22.2 33.3 5.6 100
20 13 3 1 3 18 2 2 1 3 2 1 2 2 1 2 9 3 1 1 2 1 1 47
100 65.0 15.0 5.0 15.0 100 11.1 11.1 5.6 16.7 11.1 5.6 11.1 11.1 5.6 11.1 100 33.3 11.1 11.1 22.1 11.1 11.1 100
17 12 3 1 1 17 2 2 1 7 1 2 2 5 3 1 1 39
100 70.6 17.6 5.9 5.9 100 11.8 11.8 5.8 41.2 5.8 11.8 11.8 100 60.0 20.0 20.0 100
10 6 2 2 5 2 3 2 1 1 17
100 60.0 20.0 20.0 100 40.0 60.0 100 50.0 50.0 100
CNS, coagulase-negative staphylococci; P1, phase 1: March 2005 to March 2007; P2, phase 2: April 2007 to April 2009.
58.8% of microorganisms identified were gram positive, followed by gram-negative organisms (29.4%) and fungi (11.8%). Coagulase-negative staphylococci accounted for over 60% of the gram-positive pathogens in the 2 phases in the ICU and in the SDUs. Candida non-albicans was more prevalent in ICU and Candida albicans in SDUs. The most prevalent gram-negative pathogens were Acinetobacter baumannii and Klebsiella pneumoniae.
DISCUSSION Our hospital is engaged in a patient safety program that is a resource from the IHI. This was one of our reasons for adopting the ‘‘CLABSI—getting to zero’’ goal as a quality indicator for patient safety in our ICU and SDUs. However, eliminating health care-associated infection includes difficulties with the case definition and risk adjustment. Moreover, Edmond has pointed out that a ‘‘getting to zero’’ approach can be associated with adverse unintended consequences.10 Health care professionals should recognize the difference between
surveillance definitions and clinical definitions. The surveillance definitions for catheter-associated BSI includes all BSIs that occur in patients with CVCs when other sites of infection have been excluded. That is, the surveillance definition overestimates the true incidence of catheter-related BSI because not all BSIs originate from a catheter. A more rigorous definition might include only those BSIs for which other sources were excluded by careful examination of the patient record and where a culture of the catheter tip demonstrated substantial colonies of an organism identical to those found in the bloodstream.3 NHSN methodology overestimates CLABSI rates.11 Although each central line represents a separate risk for infection, the NHSN CLABSI rate is expressed as the number of CLABSI per thousand catheter-days; for any given day, only 1 catheter can be counted per patient regardless of the number of central lines in situ.3 In our study, we used the denominator for the CLABSI rate as per the NHSN methodology; however, it must be emphasized that a sizable proportion of ICU patients have more than 1 central line.
438
Marra et al.
American Journal of Infection Control August 2010 Incidence density of CLABSI in ICU (per 1000 catheter-days)
14.0 12.0 10.0 8.0 6.0 4.0 2.0 0.0
m 2005
m
j
s
n
m
j 2006
m
j
s
n
j 2007
m
m
Rate
j
s
n
j 2008
m
m
j
s
n
j 2009
m
6-month rolling average
Phase 1
Phase 2
CLABSI=Central line associated bloodsream infection
Fig 1. Incidence density of bloodstream infections in ICU per 1000 catheter-days.
Incidence density of CLABSI in SDUs (per 1,000 catheter-days) 12.0 10.0 8.0 6.0 4.0 2.0 0.0 m 2005
m
j
s
n
j 2006
m
m
j
s
n
j 2007
Rate
m
m
j
s
n
j 2008
m
m
j
s
n
j 2009
m
6-month rolling average
Phase 1
Phase 2
CLABSI=Central line associated bloodsream infection
Fig 2. Incidence density of bloodstream infections in SDU per 1000 catheter-days.
It must be pointed out that eliminating health careassociated infections has been identified as an important priority in many hospitals,5,8,12 especially because there was a decision in the United States by Medicare to no longer provide increased payments for CLABSIs. In Brazilian hospitals, this has not yet happened, although our hospital’s chief executive officer and senior management are responsible for ensuring that the health care system supports an infection prevention and control program to effectively prevent CLABSI.4 Hospitals have increased the size and the numbers of their ICUs,13 adding SDUs to provide appropriate care for patients whose acuity of illness falls between that of ICU patients and that of ward patients. Weber et al demonstrated that the infection rate provides a better delineation of the impact of health careassociated infections across different types of adult units (ICU, SDU, or ward) and that the infection rate in the SDU is more similar to the infection rate in the
ward setting.13 We decided to apply the central line bundle in our SDUs; however, this infection control practice has good evidence only for intensive care patients.5,8 It is important to note that our CLABSI surveillance was performed by trained infection control professionals using the US CDC definition9 for the 2 phases of the study in the ICU and in the SDUs. Early studies of chlorhexidine/silver sulfadiazine- and minocycline/rifampin-coated catheters showed reduced rates of both catheter colonization and catheter-related BSIs when compared with controls.14,15 Although more expensive, a cost-effective analysis showed cost savings with the minocycline/rifampin catheter, considering less money spent on treatment and catheter exchange.16 Interestingly, Berenholtz et al showed a dramatic reduction in BSIs in ICU patients, without the use of coated catheters simply by standardizing the CVC insertion process and asking providers daily whether catheters could be removed.8 We do not use
www.ajicjournal.org Vol. 38 No. 6
impregnated CVCs in the ICU and in the SDU because we previously demonstrated no statistically significant difference in colonization rates in our ICU.17 To our knowledge, our study is the first to evaluate preventive measures for SDU patients that have been applied to ICU patients for controlling CLABSI. There are several limitations to this study. First, this is not a randomized trial but a quasiexperimental, interrupted time series study. Quasiexperimental study designs are frequently used when it is not logistically feasible to conduct a controlled trial. Thus, other unmeasured factors might have occurred coincident with the interventions that occurred since April 2007 (implementation of the central line bundle), resulting in a decrease in BSI rates in our ICU. However, this seems unlikely because there had been no decrease in BSI rates previously over the several last years (Figs 1 and 2). Second, we did not collect data on preintervention process measure compliance, and this lack of data decreased our ability to measure the full impact of the bundles; however, a significant difference was demonstrated in CLABSI rates in the ICU and in the SDUs, and we have achieved a zero infection rate by applying the CLABSI prevention measures that are recommended in the literature.4,5,8 In the second phase, in the SDUs, the central line bundle was applied, but we have compliance data only from January 2008 to April 2009. Finally, because this intervention was performed at a single medical center, these results may not be generalizable to other hospitals. In conclusion, the process and measures for CLABSI presented here are derived from published guidelines and other relevant literature, showing that implementing these interventions must be a priority not only for ICU settings but also for all health care facilities including SDU settings.
Marra et al.
4.
5.
6.
7.
8.
9.
10. 11.
12.
13.
14.
15.
References 1. Richard MJ, Edwards JR, Culver DH, Gaynes RP. Nosocomial infections in combined medical-surgical intensive care units in the United States. Infect Control Hosp Epidemiol 2000;21:510-5. 2. Garrouste Orgeas M, Timsit JF, Soufir L, Tafflet M, Adrie C, Philippart F, et al. Impact of adverse events on outcomes in intensive care unit patients. Crit Care Med 2008;36:2041-7. 3. O’Grady NP, Alexander M, Dellinger EP, Gerberding JL, Heard SO, Maki DG, et al. Guidelines for the prevention of intravascular
16.
17.
439
catheter-related infections. Centers for Disease Control and Prevention. MMWR Recomm Rep 2002;51:1-29. Marschall J, Mermel LA, Classen D, Arias KM, Podgorny K, Anderson DJ, et al. Strategies to prevent central line-associated bloodstream infections in acute care hospitals. Infect Control Hosp Epidemiol 2008;29(Suppl 1):S22-30. Pronovost P, Needham D, Berenholtz S, Sinopoli D, Chu H, Cosgrove S, et al. An intervention to decrease catheter-related bloodstream infections in the ICU. N Engl J Med 2006;355:2725-32. Marschall J, Leone C, Jones M, Nihill D, Fraser VJ, Warren DK. Catheter-associated bloodstream infections in general medical patients outside the intensive care unit: a surveillance study. Infect Control Hosp Epidemiol 2007;28:905-9. Wisplinghoff H, Bischoff T, Tallent SM, Seifert H, Wenzel RP, Edmond MB. Nosocomial bloodstream infections in US hospitals: analysis of 24,179 cases from a prospective nationwide surveillance study. Clin Infect Dis 2004;39:309-17. Berenholtz SM, Pronovost PJ, Lipsett PA, Hobson D, Earsing K, Farley JE, et al. Eliminating catheter-related bloodstream infections in the intensive care unit. Crit Care Med 2004;32:2014-20. Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance definition of health care-associated infection and criteria for specific types of infections in the acute care setting. Am J Infect Control 2008;36: 309-32. Edmond MB. Getting to zero: is it safe? Infect Control Hosp Epidemiol 2009;30:74-6. Nalepa M, Sanogo K, Edmond MB. NHSN methodology overestimates central line-associated bloodstream infection rates. Society for Healthcare Epidemiology of America Meeting, San Diego, CA; 2009. Jeffries HE, Mason W, Brewer M, Oakes KL, Mun˜oz EI, Gornick W, et al. Prevention of central venous catheter-associated bloodstream infections in pediatric intensive care units: a performance improvement collaborative. Infect Control Hosp Epidemiol 2009;30:645-51. Weber DJ, Sickbert-Bennett EE, Brown V, Rutala WA. Comparison of hospitalwide surveillance and targeted intensive care unit surveillance of healthcare-associated infections. Infect Control Hosp Epidemiol 2007;28:1361-6. Raad I, Darouiche R, Dupuis J, Abi-Said D, Gabrielli A, Hachem R, et al. Central venous catheters coated with minocycline and rifampin for the prevention of catheter-related colonization and bloodstream infections: a randomized, double-blind trial. The Texas Medical Center Catheter Study Group. Ann Intern Med 1997;127:267-74. Maki DG, Stolz SM, Wheeler S, Mermel LA. Prevention of central venous catheter-related bloodstream infection by use of an antisepticimpregnated catheter: a randomized, controlled trial. Ann Intern Med 1997;127:257-66. Shorr AF, Humphreys CW, Helman DL. New choices for central venous catheters: potential financial implications. Chest 2003;124: 275-84. Camargo LF, Marra AR, Bu¨chele GL, Sogayar AM, Cal RG, de Sousa JM, et al. Double-lumen central venous catheters impregnated with chlorhexidine and silver sulfadiazine to prevent catheter colonisation in the intensive care unit setting: a prospective randomised study. J Hosp Infect 2009;72:227-33.
Background: Central line-associated bloodstream infection (CLABSI) is one of the most important health care-associated infections in the critical care setting. Methods: A quasiexperimental study involving multiple interventions to reduce the incidence of CLABSI was conducted in a medical-surgical intensive care unit (ICU) and in 2 step-down units (SDUs). From March 2005 to March 2007 (phase 1 [P1]), some Centers for Disease Control and Prevention evidence-based practices were implemented. From April 2007 to April 2009 (P2), we intervened in these processes at the same time that performance monitoring was occurring at the bedside, and we implemented the Institute for Healthcare Improvement central line bundle for all ICU and SDU patients requiring central venous lines. Results: The mean incidence density of CLABSI per 1000 catheter-days in the ICU was 6.4 in phase 1 and 3.2 in phase 2, P , .001. The mean incidence density of CLABSI per 1000 catheter-days in the SDUs was 4.1 in phase 1 and 1.6 in phase 2, P 5 .005. Conclusion: These results suggest that reducing CLABSI rates in an ICU setting is a complex process that involves multiple performance measures and interventions that can also be applied to SDU settings. Key Words: Bloodstream infection; prevention; intensive care unit; step-down unit; zero tolerance; central venous catheter; epidemiology; central line bundle. Copyright ª 2010 by the Association for Professionals in Infection Control and Epidemiology, Inc. Published by Elsevier Inc. All rights reserved. (Am J Infect Control 2010;38:434-9.)
Central line-associated bloodstream infection (CLABSI) is one of the most important health careassociated infections in the critical care setting.1 These infections are associated with high mortality rates in adult intensive care patients2 and are the leading cause of death among health care-associated infections.3 Thus, CLABSI is an important target for interventions. The Centers for Disease Control and Prevention (CDC) and the Society for Healthcare Epidemiology of America, Inc (SHEA)/Infectious Disease Society of America guidelines describe many different
From the Intensive Care Unit, Hospital Israelita Albert Einstein, Sa˜o Paulo, Brazila; Division of Nephrology, Universidade Federal de Sa˜o Paulo, Sa˜o Paulo, Brazil (UNIFESP-EPM)b; Infection Control Unit, Hospital Israelita Albert Einstein, Sa˜o Paulo, Brazilc; and Department of Internal Medicine, Virginia Commonwealth University School of Medicine, Richmond, VA.d Address correspondence to Alexandre R. Marra, MD, Av Albert Einstein, 627/701-58 Andar-Bloco B, Sa˜o Paulo, Brazil, CEP 05651-901. E-mail: [email protected].
Conflicts of interest: None to report. 0196-6553/$36.00 Copyright ª 2010 by the Association for Professionals in Infection Control and Epidemiology, Inc. Published by Elsevier Inc. All rights reserved. doi:10.1016/j.ajic.2009.11.012
434
interventions for preventing CLABSI in intensive care units (ICUs).3,4 Each of these recommendations is categorized on the basis of existing scientific evidence, theoretical rationale, applicability, and potential economic impact. As part of the 5 Million Lives campaign, endorsed by leading US agencies and professional societies, The Institute for Healthcare Improvement (IHI) (http://www.ihi.org/ihi) recommends that all ICUs implement a central line bundle to attempt to reduce the incidence of CLABSI to zero.5 However, in the ICU, central venous access might be needed for extended periods of time; patients can be colonized with hospital-acquired organisms; and the catheter may be manipulated multiple times per day for the administration of fluids, drugs, and blood products. Moreover, some catheters must be inserted in urgent situations when optimal attention to aseptic technique may not be feasible. Certain catheters (eg, pulmonary artery catheters and hemodialysis catheters) can be accessed multiple times per day for hemodynamic measurements or for hemodialysis, increasing the potential for contamination and subsequent clinical infection.3,4 Recent data suggest that non-ICU medical wards have device utilization rates that are considerably lower than those of medical ICUs, but CLABSI rates were similar to those in medical ICUs.6
www.ajicjournal.org Vol. 38 No. 6
Unfortunately, few data exist about the prevention of CLABSI in step-down units (SDUs). Types of organisms that most commonly cause hospital-acquired bloodstream infections (BSIs) change over time, but gram-positive organisms are responsible for the great majority of BSI cases, principally coagulase-negative staphylococci.7 The purpose of this prospective, quality improvement study was to examine the effect of a series of interventions, implemented in an ICU and SDUs to reduce the incidence of CLABSI and to analyze the differences in CLABSIs and microorganisms in the 2 study phases.
METHODS Setting and study design A quasiexperimental study was conducted in a 38bed medical-surgical ICU and in two 20-bed SDUs with the same physical layout in a tertiary care, private hospital in Sa˜o Paulo, Brazil. The ICU has an open staffing model, and approximately 2200 patients are admitted annually. In the SDUs, all the rooms are single bed rooms. The SDU patients are transferred from the medical-surgical ICU of this hospital, from wards, and from the emergency department. Because this study was considered a quality improvement project, it was not submitted to our Institutional Review Board. The study was carried out in 2 phases: phase 1 (March 2005 to March 2007) and phase 2 (April 2007 to April 2009). In phase 1, attending physicians (eg, surgery doctors), ICU doctors, or ICU medical residents inserted the catheters through a new venipuncture, into the subclavian, jugular, or femoral vein using full sterilebarrier precautions and 2% chlorhexidine preparation for skin antisepsis per CDC guidelines.3 The decision to remove the catheter was made solely by the patient’s physician, who kept the catheter in place until it was no longer needed or until an adverse event necessitated its removal. Catheters were not routinely replaced. Each year, a convenience sample of patients was chosen for whom catheter insertion and catheter dressings were directly observed by assigned nurses. We then provided feedback via e-mail on compliance with these processes for the ICU team (doctors and nurses). In phase 2, after the hospital’s chief executive officer articulated zero tolerance for CLABSI, we continued intervening in these processes as in phase 1; however, these process measures were audited once monthly at random intervals in a small sample of patients undergoing central line insertion. In April 2007, we implemented the IHI central line bundle8 (creation of a central catheter insertion cart; hand hygiene; maximal barrier precautions for insertion; chlorhexidine skin antisepsis; optimal catheter site selection, with
Marra et al.
435
avoidance of the femoral vein for central venous access in adult patients; daily review of line necessity with prompt removal of unnecessary lines) for all ICU and SDU patients requiring central venous lines. This central line bundle was monitored everyday by ICU and SDU nurses and doctors. The ICU and SDU nurses intervened in this process at the same time that performance monitoring was occurring at the bedside if noncompliance with an element of the bundle was detected (eg, hand hygiene was not performed). Prior to the start of phase 2, we delivered a brief presentation to the ICU staff on BSI prevention, reviewed the study protocol, and encouraged participation in our ‘‘central line bundle—getting to zero’’ program. Every month, we provided feedback on compliance with the bundle components via e-mail for the ICU and SDU team (doctors and nurses). In addition, we provided posters in the ICU and SDU with bar graphs displaying compliance with process of care measures. In these posters, we also provided the CLABSI rate determined by surveillance conducted by the infection control and hospital epidemiology program. We implemented an ICU doctor’s group to remove the unnecessary catheters everyday. Once a day, an ICU doctor (not on clinical duty) checked all the ICU patients with central lines and asked the ICU doctors (on duty) if each central line was necessary. The same strategy was performed in the SDUs, except in these units there were nurses who questioned SDU doctors (on duty) about central line necessity. In phase 2, the same infection control practices were performed. We did not use impregnated catheters or chlorhexidine dressings for central lines in ICU and SDU patients. Our CLABSI rates did not include central and peripheral arterial catheter infections. Compliance with all process measures while inserting the central venous catheter (CVC) was evaluated for the CVCs that were placed in the ICU and in the SDUs. All percutaneous CVCs for intravenous fluid, medication, dialysis, or administration of total parenteral nutrition were included. We have used peripherally inserted central catheter lines in these units since 2003. The peripherally inserted central catheters are placed by a hospital nurse team that is specialized in this procedure.
Definitions CLABSI surveillance was performed by trained infection control professionals using the CDC definition9 for the 2 phases of the study for laboratory-confirmed central venous BSI. A CLABSI was considered to be related to a specific unit if detected at least 48 hours after admission to or less than 48 hours after discharge from the unit.3 The device utilization ratio was defined as
436
Marra et al.
American Journal of Infection Control August 2010
Table 1. Characteristics of both phases in the ICU and in the SDUs during the study periods ICU
SDUs
P1 Patient-days CVC-days CVC utilization ratio* CVC check-list, n (%) Compliance with all process measures while inserting the CVC, n (%)z Hand hygiene Full barrier precaution Chlorhexidine skin preparation Sterile field (at least 2) Not necessary to stop the procedure Sterile dressing Compliance with CVC site care, n (%) Site of catheter insertion Adequate dressing CVC type inserted, n (%) Double lumen Hemodialysis catheter PICC Central venous oximetry catheter Swan-Ganz BSI rate per 1000 catheter-days, mean (SD)
P2
P1
20,046 15,650 0.77 -
20,924 14,111 0.67 1833/2392 (76.6) 1723/1833 (94.0)
29,324 11,252 0.38 -
-
1820/1833 (99.3) 1815/1833 (99.0) 1810/1833 (98.7) 1786/1833 (97.4) 1814/1833 (99.0) 1799/1833 (98.1)
-
123/139 (88.4) 127/139 (91.4) 6.4 (2.9)
-
41/43 (95.3) 33/43 (76.7) 4.1 (3.3)
612/697 (87.8)y 2392 (100) 1580 (66.0) 423 (17.7) 172 (7.2) 141 (5.9) 76 (3.2) 3.2 (2.5)
P2 28,871 10,369 0.35 176/232 (75.9)y 161/176 (91.5)y 156/176 (88.6)y 155/176 (88.1)y 159/176 (90.3)y 158/176 (89.8)y 176/176 (100)y 176/176 (100)y 1034/1085 (95.3)y NA NA NA NA 1.6 (2.4)
BSI, bloodstream infection; CVC, central venous catheter; NA, not available; P1, phase 1: March 2005 to March 2007; P2, phase 2: April 2007 to April 2009; SD, standard deviation. *Defined as the number of CVC-days divided by the number of patient-days. y From January 2008 to April 2009. z Compliance is defined as the percentage of time the required component was performed without intervention by the bedside nurse.
the number of catheter-days divided by the number of patient-days.
Microbiologic methods All isolates were identified by manual or automated methods and confirmed using the Vitek 2 system (bioMerieux Vitek, Inc., Hazelwood, MO).
Statistical analysis Comparisons between phase 1 versus phase 2 for compliance with process measures were performed using the x2 test. Continuous variables were compared using the Student t test for normally distributed variables and the Mann-Whitney U test for non-normally distributed variables. All tests of statistical significance were 2-sided with a significance level set at .05. All the data analyses were performed using SPSS for Windows 14.0 (SPSS Inc., Chicago, IL).
RESULTS Study sample, compliance, and incidence density of CLABSI in each phase In ICU phase 1, there were 20,046 patient-days and 15,650 CVC-days, and the CVC utilization ratio was 0.77
(Table 1). In ICU phase 2, there were 20,924 patientdays and 14,111 CVC-days and a 0.67 CVC utilization ratio. In SDU phase 1, there were 29,324 patient-days and 11,252 CVC-days and a 0.38 CVC utilization ratio and, in SDU phase 2, 28,871 patient-days and 10,369 patientdays and a 0.35 CVC utilization ratio. Compliance with process measures for catheter insertion and site care (site of catheter insertion and adequate dressing) is shown in Table 1. The mean incidence density of CVC-BSI per 1000 catheter-days in the ICU was 6.4 in phase 1 and 3.2 in phase 2, P , .001. The mean incidence density of CVC-BSI per 1000 catheter-days in the SDUs was 4.1 in phase 1 and 1.6 in phase 2, P 5 .005. The CVC type inserted is also shown in Table 1.
Microbiologic features As seen in Table 2, 42.1% of all the microorganisms identified in ICU phase 1 were gram positive, followed by gram-negative organisms (38.9%) and fungi (19.0%). In ICU phase 2, 42.6% of microorganisms identified were gram positive, followed by gramnegative organisms (38.1%) and fungi (19.1%). In SDU phase 1, the same proportion (43.6%) of microorganisms was gram positive and gram negative, and a small proportion were fungi (12.8%). In SDU phase 2,
www.ajicjournal.org Vol. 38 No. 6
Marra et al.
437
Table 2. Characteristics of pathogens causing BSI associated to CVC during the 2 study phases in ICU and SDU patients ICU
SDUs
P1
P2
P1
P2
Pathogen
n
%
n
%
n
%
n
%
Gram positive CNS Enterococcus faecalis Enterococcus faecium Lactobacillus gasseri Staphylococcus aureus Gram negative Acinetobacter baumannii Acinetobacter calcoaceticus Burkholderia cepacia Enterobacter cloacae Escherichia coli Klebsiella pneumoniae Morganella morganii Proteus mirabilis Pseudomonas aeruginosa Pseudomonas putida Serrratia marcescens Stenotrophomonas maltophilia Fungi Candida albicans Candida glabrata Candida krusei Candida parapsilosis Candida tropicalis Saccharomyces cerevisiae Scytalidium dimidiatum Trichosporon asahii Total identified agents
40 26 5 1 8 37 12 2 4 2 10 1 3 1 2 18 5 2 4 6 1 95
100 65.0 12.5 2.5 20.0 100 32.4 5.4 10.8 5.4 27.0 2.7 8.1 2.7 5.4 100 27.8 11.1 22.2 33.3 5.6 100
20 13 3 1 3 18 2 2 1 3 2 1 2 2 1 2 9 3 1 1 2 1 1 47
100 65.0 15.0 5.0 15.0 100 11.1 11.1 5.6 16.7 11.1 5.6 11.1 11.1 5.6 11.1 100 33.3 11.1 11.1 22.1 11.1 11.1 100
17 12 3 1 1 17 2 2 1 7 1 2 2 5 3 1 1 39
100 70.6 17.6 5.9 5.9 100 11.8 11.8 5.8 41.2 5.8 11.8 11.8 100 60.0 20.0 20.0 100
10 6 2 2 5 2 3 2 1 1 17
100 60.0 20.0 20.0 100 40.0 60.0 100 50.0 50.0 100
CNS, coagulase-negative staphylococci; P1, phase 1: March 2005 to March 2007; P2, phase 2: April 2007 to April 2009.
58.8% of microorganisms identified were gram positive, followed by gram-negative organisms (29.4%) and fungi (11.8%). Coagulase-negative staphylococci accounted for over 60% of the gram-positive pathogens in the 2 phases in the ICU and in the SDUs. Candida non-albicans was more prevalent in ICU and Candida albicans in SDUs. The most prevalent gram-negative pathogens were Acinetobacter baumannii and Klebsiella pneumoniae.
DISCUSSION Our hospital is engaged in a patient safety program that is a resource from the IHI. This was one of our reasons for adopting the ‘‘CLABSI—getting to zero’’ goal as a quality indicator for patient safety in our ICU and SDUs. However, eliminating health care-associated infection includes difficulties with the case definition and risk adjustment. Moreover, Edmond has pointed out that a ‘‘getting to zero’’ approach can be associated with adverse unintended consequences.10 Health care professionals should recognize the difference between
surveillance definitions and clinical definitions. The surveillance definitions for catheter-associated BSI includes all BSIs that occur in patients with CVCs when other sites of infection have been excluded. That is, the surveillance definition overestimates the true incidence of catheter-related BSI because not all BSIs originate from a catheter. A more rigorous definition might include only those BSIs for which other sources were excluded by careful examination of the patient record and where a culture of the catheter tip demonstrated substantial colonies of an organism identical to those found in the bloodstream.3 NHSN methodology overestimates CLABSI rates.11 Although each central line represents a separate risk for infection, the NHSN CLABSI rate is expressed as the number of CLABSI per thousand catheter-days; for any given day, only 1 catheter can be counted per patient regardless of the number of central lines in situ.3 In our study, we used the denominator for the CLABSI rate as per the NHSN methodology; however, it must be emphasized that a sizable proportion of ICU patients have more than 1 central line.
438
Marra et al.
American Journal of Infection Control August 2010 Incidence density of CLABSI in ICU (per 1000 catheter-days)
14.0 12.0 10.0 8.0 6.0 4.0 2.0 0.0
m 2005
m
j
s
n
m
j 2006
m
j
s
n
j 2007
m
m
Rate
j
s
n
j 2008
m
m
j
s
n
j 2009
m
6-month rolling average
Phase 1
Phase 2
CLABSI=Central line associated bloodsream infection
Fig 1. Incidence density of bloodstream infections in ICU per 1000 catheter-days.
Incidence density of CLABSI in SDUs (per 1,000 catheter-days) 12.0 10.0 8.0 6.0 4.0 2.0 0.0 m 2005
m
j
s
n
j 2006
m
m
j
s
n
j 2007
Rate
m
m
j
s
n
j 2008
m
m
j
s
n
j 2009
m
6-month rolling average
Phase 1
Phase 2
CLABSI=Central line associated bloodsream infection
Fig 2. Incidence density of bloodstream infections in SDU per 1000 catheter-days.
It must be pointed out that eliminating health careassociated infections has been identified as an important priority in many hospitals,5,8,12 especially because there was a decision in the United States by Medicare to no longer provide increased payments for CLABSIs. In Brazilian hospitals, this has not yet happened, although our hospital’s chief executive officer and senior management are responsible for ensuring that the health care system supports an infection prevention and control program to effectively prevent CLABSI.4 Hospitals have increased the size and the numbers of their ICUs,13 adding SDUs to provide appropriate care for patients whose acuity of illness falls between that of ICU patients and that of ward patients. Weber et al demonstrated that the infection rate provides a better delineation of the impact of health careassociated infections across different types of adult units (ICU, SDU, or ward) and that the infection rate in the SDU is more similar to the infection rate in the
ward setting.13 We decided to apply the central line bundle in our SDUs; however, this infection control practice has good evidence only for intensive care patients.5,8 It is important to note that our CLABSI surveillance was performed by trained infection control professionals using the US CDC definition9 for the 2 phases of the study in the ICU and in the SDUs. Early studies of chlorhexidine/silver sulfadiazine- and minocycline/rifampin-coated catheters showed reduced rates of both catheter colonization and catheter-related BSIs when compared with controls.14,15 Although more expensive, a cost-effective analysis showed cost savings with the minocycline/rifampin catheter, considering less money spent on treatment and catheter exchange.16 Interestingly, Berenholtz et al showed a dramatic reduction in BSIs in ICU patients, without the use of coated catheters simply by standardizing the CVC insertion process and asking providers daily whether catheters could be removed.8 We do not use
www.ajicjournal.org Vol. 38 No. 6
impregnated CVCs in the ICU and in the SDU because we previously demonstrated no statistically significant difference in colonization rates in our ICU.17 To our knowledge, our study is the first to evaluate preventive measures for SDU patients that have been applied to ICU patients for controlling CLABSI. There are several limitations to this study. First, this is not a randomized trial but a quasiexperimental, interrupted time series study. Quasiexperimental study designs are frequently used when it is not logistically feasible to conduct a controlled trial. Thus, other unmeasured factors might have occurred coincident with the interventions that occurred since April 2007 (implementation of the central line bundle), resulting in a decrease in BSI rates in our ICU. However, this seems unlikely because there had been no decrease in BSI rates previously over the several last years (Figs 1 and 2). Second, we did not collect data on preintervention process measure compliance, and this lack of data decreased our ability to measure the full impact of the bundles; however, a significant difference was demonstrated in CLABSI rates in the ICU and in the SDUs, and we have achieved a zero infection rate by applying the CLABSI prevention measures that are recommended in the literature.4,5,8 In the second phase, in the SDUs, the central line bundle was applied, but we have compliance data only from January 2008 to April 2009. Finally, because this intervention was performed at a single medical center, these results may not be generalizable to other hospitals. In conclusion, the process and measures for CLABSI presented here are derived from published guidelines and other relevant literature, showing that implementing these interventions must be a priority not only for ICU settings but also for all health care facilities including SDU settings.
Marra et al.
4.
5.
6.
7.
8.
9.
10. 11.
12.
13.
14.
15.
References 1. Richard MJ, Edwards JR, Culver DH, Gaynes RP. Nosocomial infections in combined medical-surgical intensive care units in the United States. Infect Control Hosp Epidemiol 2000;21:510-5. 2. Garrouste Orgeas M, Timsit JF, Soufir L, Tafflet M, Adrie C, Philippart F, et al. Impact of adverse events on outcomes in intensive care unit patients. Crit Care Med 2008;36:2041-7. 3. O’Grady NP, Alexander M, Dellinger EP, Gerberding JL, Heard SO, Maki DG, et al. Guidelines for the prevention of intravascular
16.
17.
439
catheter-related infections. Centers for Disease Control and Prevention. MMWR Recomm Rep 2002;51:1-29. Marschall J, Mermel LA, Classen D, Arias KM, Podgorny K, Anderson DJ, et al. Strategies to prevent central line-associated bloodstream infections in acute care hospitals. Infect Control Hosp Epidemiol 2008;29(Suppl 1):S22-30. Pronovost P, Needham D, Berenholtz S, Sinopoli D, Chu H, Cosgrove S, et al. An intervention to decrease catheter-related bloodstream infections in the ICU. N Engl J Med 2006;355:2725-32. Marschall J, Leone C, Jones M, Nihill D, Fraser VJ, Warren DK. Catheter-associated bloodstream infections in general medical patients outside the intensive care unit: a surveillance study. Infect Control Hosp Epidemiol 2007;28:905-9. Wisplinghoff H, Bischoff T, Tallent SM, Seifert H, Wenzel RP, Edmond MB. Nosocomial bloodstream infections in US hospitals: analysis of 24,179 cases from a prospective nationwide surveillance study. Clin Infect Dis 2004;39:309-17. Berenholtz SM, Pronovost PJ, Lipsett PA, Hobson D, Earsing K, Farley JE, et al. Eliminating catheter-related bloodstream infections in the intensive care unit. Crit Care Med 2004;32:2014-20. Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance definition of health care-associated infection and criteria for specific types of infections in the acute care setting. Am J Infect Control 2008;36: 309-32. Edmond MB. Getting to zero: is it safe? Infect Control Hosp Epidemiol 2009;30:74-6. Nalepa M, Sanogo K, Edmond MB. NHSN methodology overestimates central line-associated bloodstream infection rates. Society for Healthcare Epidemiology of America Meeting, San Diego, CA; 2009. Jeffries HE, Mason W, Brewer M, Oakes KL, Mun˜oz EI, Gornick W, et al. Prevention of central venous catheter-associated bloodstream infections in pediatric intensive care units: a performance improvement collaborative. Infect Control Hosp Epidemiol 2009;30:645-51. Weber DJ, Sickbert-Bennett EE, Brown V, Rutala WA. Comparison of hospitalwide surveillance and targeted intensive care unit surveillance of healthcare-associated infections. Infect Control Hosp Epidemiol 2007;28:1361-6. Raad I, Darouiche R, Dupuis J, Abi-Said D, Gabrielli A, Hachem R, et al. Central venous catheters coated with minocycline and rifampin for the prevention of catheter-related colonization and bloodstream infections: a randomized, double-blind trial. The Texas Medical Center Catheter Study Group. Ann Intern Med 1997;127:267-74. Maki DG, Stolz SM, Wheeler S, Mermel LA. Prevention of central venous catheter-related bloodstream infection by use of an antisepticimpregnated catheter: a randomized, controlled trial. Ann Intern Med 1997;127:257-66. Shorr AF, Humphreys CW, Helman DL. New choices for central venous catheters: potential financial implications. Chest 2003;124: 275-84. Camargo LF, Marra AR, Bu¨chele GL, Sogayar AM, Cal RG, de Sousa JM, et al. Double-lumen central venous catheters impregnated with chlorhexidine and silver sulfadiazine to prevent catheter colonisation in the intensive care unit setting: a prospective randomised study. J Hosp Infect 2009;72:227-33.