The Ability of Intensive Care Units to Maintain Zero Central Line–Associated Bloodstream Infections FREE

Central line–associated bloodstream infections (CLABSIs) are common, costly, and largely preventable and are a target of many recent national initiatives.^1- 3 One common method of reporting performance is the duration of time without infection,⁴ a simple technique used for motivating sustained improvement. Despite evidence demonstrating reduced infection rates from evidence-based interventions for CLABSIs, the field has not yet defined the attainable duration of time hospitals can go without a CLABSI. Without understanding “best achievable” time without any infection, hospitals may anchor themselves, and public expectations, to suboptimal performance. Such data are required to establish performance benchmarks that represent best practices, create realistic public expectations, motivate internal improvement efforts, and design fair pay-for-performance policies.

The objective of this study was to explore and quantify the ability of intensive care units (ICUs) to sustain zero CLABSIs over time. This study examined data from ICUs that participated in the Michigan Keystone ICU Project (an initiative including the Johns Hopkins University Quality and Safety Research Group's Comprehensive Unit-Based Safety Program and CLABSI Program), which was associated with statewide reduction in CLABSI for up to 36 months.^5- 6

Data for CLABSIs were obtained from 80 ICUs in 57 hospitals located predominantly in the state of Michigan. Intensive care units were eligible for this study if they reported 36 continuous months or more of data during and after implementation of the CLABSI intervention (23 of the 103 ICUs included in the original study were excluded because they did not meet this requirement).^5- 6 Up to 4 years of data (March 2004–February 2008) were available at the time of analysis (median, 45 months; interquartile range [IQR]: 42-48 months). Data for the number of central line days and CLABSIs, as defined by the National Nosocomial Infections Surveillance (NNIS) system,¹ were analyzed on a quarterly basis. However, for ease of interpretation, duration of time without an infection is reported in months, using a conservative assumption that any CLABSI occurred in the first month of a quarter.

The primary outcome was the greatest number of consecutive months with zero CLABSIs reported during the study period for each ICU (the unit of analysis). Follow-up data were available for 36 months. Secondary outcomes were binary indicators for sustaining zero CLABSIs for 12 and 24 consecutive months for each ICU. Two hospital-level factors obtained from the American Hospital Association were included: teaching status and bed size (<200; 200-299; 300-399; and >399 beds).^5- 6

Proportions of ICUs sustaining zero infections for 12 and 24 months were calculated for the secondary outcomes. Outcome measures were stratified by teaching status and within each teaching status by hospital bed size. We used SAS software, version 9.1 (SAS Institute Inc, Cary, North Carolina) for statistical analysis. The Johns Hopkins Medical School institutional review board approved this study.

Among all 80 ICUs, the mean (SD) and median (IQR) consecutive number of months with zero CLABSIs was 16.5 (10.8) and 15 (6.0-25.5), respectively. Sixty percent of ICUs sustained zero CLABSIs for 12 months or more and 26% for 24 months or more (Table).

Within both teaching and nonteaching hospital categories, these outcome measures decreased as hospital bed size increased. Among teaching hospitals with more than 399 beds, 32% of ICUs went 1 year or more without a CLABSI. In general, nonteaching hospitals had a higher number of consecutive months with zero CLABSIs and a greater proportion of ICUs sustaining zero CLABSIs for 12 and 24 months compared with teaching hospitals (Table).

Overall, the majority of ICUs achieved 12 months or more without a reported CLABSI, and more than quarter achieved 24 months or more. These findings suggest that in the setting of a comprehensive initiative focused on reducing CLABSIs, extended periods without infections are possible; even among large teaching hospitals, one-third achieved at least 1 year without a CLABSI. As such, incentives for hospitals to implement and sustain evidence-based infection prevention initiatives appear warranted.

While self-reported infection rates were not externally validated, best practices to collect data were used (eg, collection by hospital-based infection-control practitioners, ensuring ICU anonymity, and instruction to use the NNIS definition for CLABSI⁷). The data are from a single state that participated in a focused CLABSI program, and a similar program may be necessary for others to achieve these results. However, the intervention is publicly available (http://www.onthecuspstophai.org/) and being implemented across 44 other states.⁸

To our knowledge, this is the first large-scale study to demonstrate the potential of ICUs to prevent CLABSIs for a sustained period. Hospitals should strive to emulate these results and implement similar interventions.⁶ Although this study offers hope that CLABSI is largely preventable, health care lacks the scientific foundation to measure and improve other types of infections. Given the lack of progress in improving patient safety, efforts to measure and prevent health care–associated infections and other types of preventable harm should be a research and policy priority.⁹

Correspondence: Dr Lipitz-Snyderman, Quality & Safety Research Group, Johns Hopkins University, 1909 Thames St, Second Floor, Baltimore, MD 21231 (alipitz@jhsph.edu).

Author Contributions: Dr Lipitz-Snyderman had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Lipitz-Snyderman, Goeschel, Marsteller, Thompson, Berenholtz, and Pronovost. Acquisition of data: Berenholtz, Lubomski, Watson, and Pronovost. Analysis and interpretation of data: Lipitz-Snyderman, Needham, Colantuoni, Marsteller, and Pronovost. Drafting of the manuscript: Lipitz-Snyderman and Pronovost. Critical revision of the manuscript for important intellectual content: Needham, Colantuoni, Goeschel, Marsteller, Thompson, Berenholtz, Lubomski, Watson, and Pronovost. Statistical analysis: Lipitz-Snyderman, Colantuoni, and Pronovost. Obtained funding: Goeschel and Pronovost. Administrative, technical, and material support: Goeschel, Thompson, Lubomski, Watson, and Pronovost. Study supervision: Needham, Lubomski, and Pronovost.

Financial Disclosure: Dr Pronovost reports receiving grants from the Agency for Healthcare Research and Quality (AHRQ) and National Patient Safety Agency to reduce CLABSIs, speaking honoraria from hospitals and the Leigh Bureau, and royalties from the sale of his book on patient safety. Dr Goeschel reports receiving honoraria for speaking on quality and patient safety nationally and internationally; current contractual support through the AHRQ ACTION network; and grant support through the Robert Wood Johnson Foundation (RWJF). Previous support included grant support from the AHRQ, RWJF, and Society for Cardiovascular Anesthesia Foundation and contracts from the UK National Patient Safety Agency and World Health Organization Patient Safety Programme. Dr Berenholtz has received support from the Michigan Health & Hospital Association, AHRQ, National Institutes of Health, World Health Organization, and RWJF for work on other studies and has received honoraria and travel expenses from various hospitals and hospital associations for consulting work related to patient safety and quality improvement. Dr Berenholtz has equity ownership in Docusys Inc that is unrelated to this study.

Funding/Support: This study was supported in part by grant 1UC1HS14246 from the AHRQ.

Additional Contributions: The Michigan Health & Hospital Association's Keystone Center and Michigan hospitals partnered in this effort to improve patient safety.