Validation: A Resource Effective Method for Central-Line Associated Blood Stream Infections among Intensive Care Units in Tennessee

Background:  Reporting of central-line associated blood stream infections (CLABSIs) to the National Healthcare Safety Network (NHSN) is the foundation for surveillance of these infections in many states.  Ensuring validity of these data is essential in an era of public reporting and future value-based purchasing.

Objective: To describe a resource effective method for conducting validation of CLABSI data reported to NHSN.

Methods: The goal was to conduct validation on a selection of facilities (at least one-third of the 78 acute-care facilities with adult/pediatric intensive care units (ICUs) and half of the 26 facilities with neonatal ICUs (NICUs) who reported CLABSI data to NHSN in 2009.  Ten facilities with an adult/pediatric ICU which reported a CLABSI rate ≥90th percentile (based on NHSN 2006-8 reference data) were chosen.  Additionally, 38 facilities reported a CLABSI rate of 0 in at least one adult/pediatric ICU.  Of these facilities, 3 were selected because the standardized infection ratio (SIR) for their whole facility was significantly less than 1 in the prior state CLABSI report (July 2008–June 2009 data), and 7 others were selected at random.  Of the 30 remaining facilities 6 were randomly chosen.  In NICUs, umbilical catheter-associated BSI (UCABSI) and CLABSI rates were calculated for 5 birth weight categories resulting in up to 10 BSI rates per NICU.  The 6 NICUs with at least 1 BSI rate ≥90th percentile were selected.  Seven NICUs with ≥7 rates of 0 were also chosen.  One facility with 6 rates of 0, which had a dramatic reduction in NICU CLABSI rates from 2008, was also selected.  A list of all positive blood cultures from all of the facilities’ ICUs was obtained; this provided a sampling frame within each facility. 

Results: Through this method 26/78 adult/pediatric ICUs and 14/26 NICUs were selected for validation.  At each facility, the aim is to review at least 16 charts.  For adult/pediatric ICUs, charts are chosen both from the ICU selected and from other ICUs in the same facility.  Charts targeted include those with coagulase negative Staphylococcus and Candida species.  The remaining charts will be randomly selected.  Reviewers are blinded to any determinations reported to NHSN.  Validation of denominator data is also performed by speaking directly with the person(s) collecting these data.

Conclusions: Targeting facilities with high and low CLABSI rates along with focusing on reviewing cases with higher potential for inaccurate reporting provides a resource efficient manner to detect reporting errors.   Problem areas can then be communicated to all facilities who can review their records to correct similar mistakes.  This creates a sustainable method for providing education and maximizing the validity of data reported to NHSN.