Eliminating CLABSI: A National Patient Safety Imperative
State Participation. Recruitment to the project is an ongoing process that began in fall 2008 and is continuing. Lead organizations in States were encouraged to recruit as many teams of participants as they could. The ongoing success of this program, awareness of impending CMS public reporting of CLABSI rates, and the 2011 requirement that hospitals submit CLABSI data into NHSN are encouraging additional hospitals to enroll. Once States agreed to participate, they were placed into a project group or "cohort" along with other States beginning the project at the same time. At present, five cohorts of States are participating in the project, and a sixth is scheduled to begin in January 2011. Many States are continuing to recruit additional hospital teams to participate. As a result, some lead organizations are supporting hospital participants in one or more of the project cohorts. Figure 1 illustrates the current status of State recruitment efforts.
Forty-four States plus the District of Columbia and Puerto Rico have committed to participating in one of the six cohorts. Although Michigan is not listed as a formal project participant, Michigan hospitals continue to work with the MHA Keystone Center on sustaining the exceptionally low CLABSI rates they achieved in the initial Keystone Project.3,4 The State of Washington has participated in some project activities but has not submitted any CLABSI data.
Levels of participation within States vary substantially. Table 1 provides a breakdown of the number of participating hospitals and teams by State. This table is limited to the first four project cohorts; cohorts 5 and 6 are still actively recruiting hospital participants. Because some States have a higher percentage of very small hospitals that do not have an ICU or insert central lines, some variation in the percentage of hospitals in each State that could benefit from project participation is to be expected. Alabama, Delaware, Hawaii, and Maryland all have more than 50 percent of their hospitals participating in the project.
Hospital Participation. Because CLABSIs can occur in all sizes and types of hospitals in which central lines are used, this project has encouraged each State lead organization to enlist the participation of all hospitals that use central lines when recruiting. Figure 2 summarizes the participation of hospitals in this project based on their size (defined by the number of hospital beds). Percentages were based on the number of hospitals with an adult ICU of each size participating in the project divided by the total number of hospitals of each size with an adult medical or surgical ICU (based on data in the 2008 AHA Annual Survey).
Thus far, project recruitment has been strongest among hospitals with more than 400 beds, and weakest among hospitals with fewer than 100 beds. Although small hospitals insert central lines in their ICUs, the number of insertions is often very low, which perhaps has made CLABSI prevention a lower priority for these hospitals.
Figure 3 summarizes the recruitment levels for hospitals with a range of characteristics. More than 24 percent of teaching hospitals with ICUs are participating in the project. The involvement of rural referral hospitals, hospitals from the 100 largest cities, and hospitals that are part of hospital systems is slightly lower. Efforts to bolster recruitment in groups where involvement is lower are ongoing.
Appendix A provides a full list of all hospitals participating in the project as of January 2011. Ohio and Washington have not provided the names of participating hospitals, so those are not included. Because the number of participating hospitals continues to grow, the list of participating hospitals maintained on the project Web site at www.onthecuspstophai.org provides the best source of current information on which hospitals have chosen to participate. While the list in Appendix A and on the Web site includes every hospital that has been enrolled in the national project database, some hospitals on this list may not be continuing to submit CLABSI data or participate in scheduled project activities on a regular basis.
Unit Team Participation. Because the vast majority of central lines are placed in patients in an ICU, recruiting ICU teams has been the project's primary focus. However, some central lines are placed and maintained in non-ICU units. Thus, teams representing these units are also participating. At present, some pediatric or neonatal ICUs are participating in the project; HRET is working with other organizations that specialize in pediatric care to provide focused guidance to hospitals seeking to prevent CLABSIs in pediatric and neonatal ICUs. Figure 4 illustrates the range of units participating in the first four cohorts of the On the CUSP: Stop BSI initiative.
More than 75 percent of units participating in the project thus far are ICUs, with a substantial majority of the ICUs consisting of adult medical/surgical ICUs.
While participation of hospitals is critical, the ultimate measure of success is the reduction of CLABSI rates.8 Table 2 summarizes CLABSI rates at baseline and in each of the four measurement periods following the inception of the project.
For all adult ICUs participating in the first two project cohorts, rates have dropped from an average of 1.8 infections per 1,000 central line days to 1.17 infections per 1,000 central line days, an overall relative reduction of 35 percent. Because mean CLABSI rates can be distorted by one or more units with very high rates, it is useful to examine both mean and median CLABSI rates. Figure 5 illustrates the changes in mean and median CLABSI rates across the data reporting periods. More than half of all participating units reported zero CLABSIs in each reporting period since they began participating in the project. Working to assist units that have not achieved this rate remains a top project priority.
Table 2 and Figure 5 include data from all units that reported data in each of the measurement periods. It is possible that the worst performing units at baseline simply failed to report data during later time periods, leading to an artificial drop in the rates. To rule out this possibility, the same analysis was performed using only units that reported data at baseline and in each subsequent reporting period. Table 3 presents these results, which are very similar.
Overall, rates decreased in the units with complete data by 31 percent, from 1.76 to 1.21 CLABSIs per 1,000 central line days, indicating that the CLABSI rate reductions observed in project participants cannot be attributed to missing data. One of the more surprising findings in this data is the relatively low baseline CLABSI rates. Alternative explanations for these low rates are being examined. However, we do not believe that only hospitals with good rates are participating in the project.
The other data concern is the accuracy of self-reported CLABSI rates. Data validation is not included in the scope of the project at this point. However, data submitted directly into our national project database is subject to more screens for accuracy than is currently the case for data submitted into NHSN. The national project team is continuing to work with AHRQ and CDC to explore the issue of validation.
3 Pronovost P, Needham D, Berenholtz S, et al. An intervention to decrease catheter-related bloodstream infections in the ICU. N Engl J Med 2006 Dec 28;355(26):2725-2732.
4 Pronovost PJ, Goeschel CA, Colantuoni E, et al. Sustaining reductions in catheter-related bloodstream infections in Michigan intensive care units: an observational study. BMJ 2010 Feb 4;340:c309.
8 A second stated project goal is the improvement in safety culture in participating units.�Significant and sustained improvement in a clinical outcome such as CLABSI requires a culture where all staff understand and can be held accountable for ensuring the safety of patients.�To assess progress in culture change, participating units are asked to complete the AHRQ Safety Culture survey at the start and end of the project. Results of these administrations of the survey will be reported when they become available.