Prevalence and test characteristics of national health safety network ventilator-associated events

Crit Care Med. 2014 Sep;42(9):2019-28. doi: 10.1097/CCM.0000000000000396.

Abstract

Objectives: The primary aim of the study was to measure the test characteristics of the National Health Safety Network ventilator-associated event/ventilator-associated condition constructs for detecting ventilator-associated pneumonia. Its secondary aims were to report the clinical features of patients with National Health Safety Network ventilator-associated event/ventilator-associated condition, measure costs of surveillance, and its susceptibility to manipulation.

Design: Prospective cohort study.

Setting: Two inpatient campuses of an academic medical center.

Patients: Eight thousand four hundred eight mechanically ventilated adults discharged from an ICU.

Interventions: None.

Measurements and main results: The National Health Safety Network ventilator-associated event/ventilator-associated condition constructs detected less than a third of ventilator-associated pneumonia cases with a sensitivity of 0.325 and a positive predictive value of 0.07. Most National Health Safety Network ventilator-associated event/ventilator-associated condition cases (93%) did not have ventilator-associated pneumonia or other hospital-acquired complications; 71% met the definition for acute respiratory distress syndrome. Similarly, most patients with National Health Safety Network probable ventilator-associated pneumonia did not have ventilator-associated pneumonia because radiographic criteria were not met. National Health Safety Network ventilator-associated event/ventilator-associated condition rates were reduced 93% by an unsophisticated manipulation of ventilator management protocols.

Conclusions: The National Health Safety Network ventilator-associated event/ventilator-associated condition constructs failed to detect many patients who had ventilator-associated pneumonia, detected many cases that did not have a hospital complication, and were susceptible to manipulation. National Health Safety Network ventilator-associated event/ventilator-associated condition surveillance did not perform as well as ventilator-associated pneumonia surveillance and had several undesirable characteristics.

MeSH terms

  • APACHE
  • Academic Medical Centers / statistics & numerical data
  • Female
  • Hospital Mortality
  • Humans
  • Intensive Care Units / statistics & numerical data*
  • Length of Stay
  • Male
  • Middle Aged
  • Patient Safety
  • Pneumonia, Ventilator-Associated / epidemiology*
  • Pneumonia, Ventilator-Associated / mortality
  • Prevalence
  • Prospective Studies
  • Public Health Surveillance / methods*
  • Quality Indicators, Health Care
  • Risk Factors