Identification of device-associated infections utilizing administrative data

Am J Infect Control. 2013 Dec;41(12):1195-9. doi: 10.1016/j.ajic.2013.03.295. Epub 2013 Jun 13.


Background: Health care-associated infections are a cause of significant morbidity and mortality in US hospitals. Recent changes have broadened the scope of health care-associated infections surveillance data to use in public reporting and of administrative data for determining Medicare reimbursement adjustments for hospital-acquired conditions.

Methods: Infection surveillance results for catheter-associated urinary tract infections (CAUTI), central line-associated bloodstream infections (CLABSI), and ventilator-associated pneumonia were compared with infections identified by hospital administrative data. The sensitivity and specificity of administrative data were calculated, with surveillance data considered the gold standard.

Results: The sensitivity of administrative data diagnosis codes for CAUTI, CLABSI, and ventilator-associated pneumonia were 0%, 21%, and 25%, respectively. The incorporation of additional diagnosis codes in definitions increased the sensitivity of administrative data somewhat with little decrease in specificity. Positive predictive values for definitions corresponding to Centers for Medicare and Medicaid services-defined hospital-acquired conditions were 0% for CAUTI and 41% for CLABSI.

Conclusions: Although infection surveillance methods and administrative data are widely used as tools to identify health care-associated infections, in our study administrative data failed to identify the same infections that were detected by surveillance. Hospitals, already incentivized by the use of performance measures to improve the quality of patient care, should also recognize the need for ongoing scrutiny of appropriate quality measures.

Keywords: Health care-associated infections; Positive predictive value; Sensitivity; Specificity; Surveillance.

MeSH terms

  • Cohort Studies
  • Cross Infection / diagnosis*
  • Epidemiologic Methods*
  • Equipment and Supplies*
  • Humans
  • Predictive Value of Tests
  • Retrospective Studies
  • Sensitivity and Specificity
  • United States