Statewide Validation of Hospital-Reported Central Line–Associated Bloodstream Infections: Oregon, 2009

Authors:
  • JTJennifer Tujo
    Jennifer Tujo
Publisher website
Google Scholar
Abstract
Background.: Mandatory reporting of healthcare-associated infections is common, but underreporting by hospitals limits meaningful interpretation.Objective.: To validate mandatory intensive care unit (ICU) central line–associated bloodstream infection (CLABSI) reporting by Oregon hospitals.Design.: Blinded comparison of ICU CLABSI determination by hospitals and health department–based external reviewers with group adjudication.Setting.: Forty-four Oregon hospitals required by state law to report ICU CLABSIs.Participants.: Seventy-six patients with ICU CLABSIs and a systematic sample of 741 other patients with ICU-related bacteremia episodes.Methods.: External reviewers examined medical records and determined CLABSI status. All cases with CLABSI determinations discordant from hospital reporting were adjudicated through formal discussion with hospital staff, a process novel to validation of CLABSI reporting.Results.: Hospital representatives and external reviewers agreed on CLABSI status in 782 (96%) of 817 bacteremia episodes (k = 0.77 [95% confidence interval (CI), 0.70-0.84]). Among the 27 episodes identified as CLABSIs by external reviewers but not reported by hospitals, the final status was CLABSI in 16 (59%). The measured sensitivities of hospital ICU CLABSI reporting were 72% (95% CI, 62%-81%) with adjudicated CLABSI determination as the reference standard and 60% (95% CI, 51%-69%) with external review alone as the reference standard (P = .07). Validation increased the statewide ICU CLABSI rate from 1.21 (95% CI, 0.95-1.51) to 1.54 (95% CI, 1.25-1.88) CLABSIs/1,000 central line–days; ICU CLABSI rates increased by more than 1.00 CLABSI/1,000 central line–days in 6 (14%) hospitals.Conclusions.: Validating hospital CLABSI reporting improves accuracy of hospital-based CLABSI surveillance. Discussing discordant findings improves the quality of validation.

This publication has 15 references indexed in Scilit: