Pa Patient Saf Advis 2013 Sep;10(3):107-9. 
Calculation of Outcome Rates That Diagnose Bedside Performance: Central-Line-Associated Bloodstream Infection
Cardiology; Critical Care; Infectious Diseases; Nursing; Pulmonary Medicine
Expand All Collapse All
Authors
James Davis, MSN, RN, CCRN, CIC
Sr. Infection Prevention Analyst

Edward Finley, BS
Data Analyst
Pennsylvania Patient Safety Authority

Corresponding Author
James Davis

Abstract

Historically, central-line-associated bloodstream infection (CLABSI) has been presented using one rate: infections per 1,000 central-line-days. While this calculation is useful for looking at overall central-line infection prevention, the calculation fails to provide information specifically related to central-line insertion or maintenance problems. Pennsylvania Patient Safety Authority analysts queried the National Healthcare Safety Network database to determine the dates of CLABSI infection events for calendar years 2010 through 2012 in Pennsylvania hospitals, along with the accompanying dates of insertion for central venous catheters (CVCs). The analysis shows that both the combined CLABSI rate and the CVC insertion infection rate trend lines are trending upward and that the CVC maintenance infection rate trend line is essentially flat. This example of trending over three years shows the limitation of using the traditional aggregate CLABSI rate to identify the CVC infection phase causing the increase. Authority analysts have shown that splitting CLABSI infection rates in a manner that correlates to the specific phase of CVC life enables clinicians to track insertion and maintenance performance. Insertion and maintenance of CVCs are separate processes; hence, there is a need for separate measurements to better target resources and improvement efforts.

Introduction

Rates of central-line-associated bloodstream infection (CLABSI) have historically been presented as the number of infections divided by the total number of central-line-days, multiplied by 1,000.1 The Pennsylvania Department of Health publishes an overall hospital CLABSI rate as part of its healthcare-associated infection (HAI) annual reports.2 While this traditional rate calculation is useful for evaluating overall central-line infection prevention performance, it does not provide information about the different components of care for patients with central lines. The use of the standardized infection ratio (SIR) also holds no promise for use by the bedside clinician. “In HAI data analysis, the SIR compares the actual number of HAIs reported with the baseline U.S. experience,” writes the Centers for Disease Control and Prevention. The SIR is designed to be “a summary measure used to track HAIs at a national, state, or local level over time.”3

A September 2011 Pennsylvania Patient Safety Advisory article, “Central-Line-Associated Bloodstream Infection: Comprehensive, Data-Driven Prevention,” delineated the insertion and maintenance phases of the central venous catheter’s (CVC) life in terms of the frequency and timing of infection.4 In another study, Ryder et al. notes that the internal lumen can be the primary source of bacteremia in short-term catheters as early as day five postinsertion.5 In the Advisory article, analysts noted that 71.7% of hospitals reported that CLABSI occurred more than five days after insertion.4 Observations by Ryder et al. targeting the source of CLABSI at or after day five correlated with the findings published by the Pennsylvania Patient Safety Authority.

The correlation between patient-level infection surveillance data4 and known pathogenesis related to intraluminal biofilm formation5,6 made it possible to differentiate between insertion-related bacteremia and maintenance-related bacteremia. If a patient experiences a CLABSI between days one and five, it is likely due to practices related to CVC insertion. If a patient experiences a CLABSI on day five or later, it is likely due to CVC maintenance practices. To produce data that can be used by clinicians, facility infection preventionists can split CLABSI infection rates in a manner that correlates to the specific phase of CVC life, thereby enabling clinicians to track insertion and maintenance performance and directly target clinical practice improvement efforts.

Methods

Using fields readily available in the data analytics function of the Centers for Disease Control and Prevention’s National Healthcare Safety Network (NHSN), Authority analysts queried the NHSN database to determine the dates of infection events occurring from 2010 through 2012 in Pennsylvania hospitals. Analysts also extracted the accompanying date of insertion for CVCs, when documented. Approximately 51% of Pennsylvania hospital CLABSI events within the selected time period had a documented insertion date. Date of CVC insertion and date of infection event were the two fields chosen to isolate data related to the determination of early versus late-onset CLABSI, yielding numerator data. Denominator data was determined as the overall number of central-line-days reported. Denominator data for CLABSI is entered into NHSN as one complete data set.

Currently, there is no mechanism within NHSN that would enable splitting the denominator data into specific patient-level data sets matching insertion phase or maintenance phase numerators. Therefore, the same denominator data set was used for each of the two numerator data sets. It is important to note that this calculation is a best possible fit designed to produce actionable data for performance tracking within each CVC phase given the current limitations of the NHSN database. The Figure shows the insertion and maintenance rates as compared with the traditional CLABSI rate per month for those facilities that reported insertion dates.

Figure. CLABSI Rates: Maintenance Phase versus Insertion Phase, by Month, 2010 through 2012

Figure. CLABSI Rates: Maintenance Phase versus Insertion Phase, by Month, 2010 through 2012 Note: Data is from the Centers for Disease Control and Prevention's National Healthcare Safety Network database.

Results

The Figure shows that the combined CLABSI rate for the Pennsylvania facilities that enter insertion dates in NHSN is trending upward. The maintenance CLABSI rate trend line is essentially flat, and the CLABSI insertion rate trend line is elevating faster than the combined trend over time. This example of trending over three years shows the limitation of the traditional combined CLABSI rate for identifying the CVC infection phase (insertion) causing the increase. The combined rate alone provides no data as to which CVC phase is actually influencing the rate increase, but as illustrated in the Figure, it appears that the maintenance phase infections have less influence over the increase in the combined rate.

This method of data presentation shows the importance of knowing which phases influence the increase in the overall CLABSI rate. This observation will have increased value on a monthly basis at the facility level, as clinicians need to react as quickly as possible to determine why infections occurred and to implement strategies to prevent future infections.

Discussion

Data Collection at the Facility Level

The Authority has demonstrated herein an approach utilizing the same denominator data set used for each of the two numerator data sets due to limitations related to the NHSN data. However, when a facility is able to produce denominator data that matches numerator data for insertion and maintenance phase infections, the data has the ability to become more meaningful. For example, the facility is able to link the number of CLABSIs that occur at less than five days to a matching denominator representative of central-line-days for those patients with CVCs for less than five days. Likewise, rate calculation for the maintenance phase would be similar, substituting the numerator data for the number of CLABSIs occurring at five or more days and the denominator data representative of central- line-days for those patients with CVCs for five or more days.

It is possible to utilize the electronic health record (EHR) in order to collect the numerator and denominator data needed to calculate separate insertion and maintenance rates. A hypothetical example of utilizing the EHR incorporates the workflow of the inserting clinicians and the staff performing maintenance. The insertion date could be tagged within the EHR for all patients with a CVC through a procedural note or upon first entry in a progress note. The EHR could then begin the algorithm for calculation of central-line-days on a patient-specific basis. The endpoint for insertion rate calculation within the system would be development of infection at less than five days. Patient-specific central-line-days data for those patients with and those patients without infection at less than five days would be combined to construct the total insertion-related denominator for the month. The algorithm for maintenance central-line-day calculation would then be the patient-specific central-line-day data for those patients with and those patients without infection at or after day five. Numerator data would then be the number of CLABSIs. Once a facility is able to generate data as described above, it is possible to plot the data over time and add confidence limits enabling the facility to establish limits of stability for the data.

Conclusion

In order to prevent CLABSI, the processes that protect the patient need to be stable. For low infection rates to be sustained, facilities must actively design specific data collection related to process performance in order to create stable, well-performing systems of CVC care. According to Wiemken, “Adequate data collection and critical analysis of control charts, [will allow] the infection preventionist [to] detect aberrant data early, which allows for prompt intervention and mitigation of any poor outcomes.”7

When the traditional CLABSI rate is plotted over time, it offers little data about performance regarding the two separate CVC phases of care that together form a combined rate. When the CLABSI rate is divided into insertion and maintenance performance calculations, targeted improvement strategies can be implemented, potentially saving time and money. For example, in reference to insertion, is the bundle of best practices being followed? Depending on the answer, clinicians performing insertion could be surveyed for potential causes of a rate that falls outside the limit. Likewise, if the maintenance rate signals an increase, clinicians who perform maintenance could be surveyed for possible causes. For example, is daily review for CVC necessity being performed, are the appropriate dressings being used, and are the dressings intact? Understanding CVC care variation through trending CLABSI insertion and maintenance rates could help healthcare facilities target their improvement efforts and prevent further CLABSI events.

Notes

  1. Dudeck MA, Horan TC, Peterson KD, et al. National Healthcare Safety Network (NHSN) report, data summary for 2010, device-associated module. Am J Infect Control 2011 Dec;39(10):798-816.
  2. Pennsylvania Department of Health. Healthcare-associated infection (HAI) annual reports [online]. [cited 2013 Jul 12]. http://www.portal.state.pa.us/portal/server.pt/community/healthcare_associated_infections/14234/hai_annual_reports/1403644
  3. Centers for Disease Control and Prevention. NHSN e-news: SIRs special edition [online]. 2010 Dec 10 [cited 2013 Jul 12]. http://www.cdc.gov/nhsn/PDFs/Newsletters/NHSN_NL_OCT_2010SE_final.pdf
  4. Davis J. Central-line-associated bloodstream infection: comprehensive, data-driven prevention. Pa Patient Saf Advis [online] 2011 Sep [cited 2013 Jul 12]. http://patientsafety.pa.gov/ADVISORIES/Pages/201109_100.aspx
  5. Ryder M, Gunther RA, Breznock EM, et al. The effect of chlorhexidine antimicrobial coating on the reduction of intraluminal biofilm formation in a clinically simulated ovine model (pilot study). Abstract at: SHEA 2011 Annual Scientific Meeting; 2011 April; Hilton Anatole, Dallas.
  6. Ryder MA. Catheter-related infections: it’s all about biofilm [online]. Top Adv Pract Nurs 2005 Aug 18 [cited 2013 Jul 12]. http://www.medscape.com/viewarticle/508109
  7. Wiemken T. Statistical process control. Chapter 6. In: APIC text of infection control and epidemiology. 3rd ed. Washington (DC): Association for Professionals in Infection Control and Epidemiology, Inc.; 2009.
PSA LOGO

The Pennsylvania Patient Safety Advisory may be reprinted and distributed without restriction, provided it is printed or distributed in its entirety and without alteration. Individual articles may be reprinted in their entirety and without alteration, provided the source is clearly attributed.

Current and previous issues are available online at http://patientsafety.pa.gov.

©2018 Pennsylvania Patient Safety Authority