Ventilator-Associated Pneumonia: A riddle, wrapped in pleura, inside an enigma

Author(s):
Walter Ramsey; Christopher O’Neil; Rebecca Saberi; Gareth Gilna; Gary Danton; Joyce I Kaufman; Howard Lieberman; Edward Lineen; Jonathan P Meizoso; Louis R Pizano; Kenneth G Proctor; Shevonne S Satahoo; Nicholas Namias

Background:

Difficulty in defining ventilator-associated pneumonia (VAP) has implications in quality reporting. The Trauma Quality Improvement Program (TQIP) defines VAP using laboratory findings, pathophysiologic signs/symptoms, and imaging criteria. Many critically ill trauma patients meet both laboratory and sign/symptom thresholds for VAP; therefore, the TQIP designation of VAP depends heavily upon imaging evidence. Reliance on imaging to identify VAP is problematic, as the radiology report verbiage reviewed by registrars is not specific to infectious etiologies and likely contributes to variability in reporting.

Hypothesis:

Physicians do not always agree with radiology reported chest radiograph findings significant for VAP.

Methods:

The TQIP Spring 2021 Benchmark Report was used to identify patients diagnosed with VAP at an academic Level 1 Trauma Center (VAP group). A control group consisted of trauma patients who spent at least four days intubated in the ICU without VAP, according to the TQIP report. For each patient, four successive chest X-rays (images only, no associated reports) were compiled and arranged sequentially. In the VAP group, the selected images overlapped with the date of VAP diagnosis. All patient identifiers were removed, and images from the VAP and control groups were randomly arranged in an electronic presentation. Twenty-two physicians (trauma/critical care and radiology attendings) were asked to identify patients with VAP based solely on imaging evidence, as defined by TQIP.

The presentation included 14 VAP-positive patients from the TQIP report, and 11 VAP-negative controls. Respondents’ answers were compared to their peers to assess the likelihood of internal agreement.

Results:

Internal agreement among physicians was 70% for trauma/critical care and 81% for radiologists. Physicians agreed with the TQIP designation of VAP on only 60% of reported VAPs (p<0.001). Trauma/critical care attendings agreed with TQIP on 52% of reported VAPs, while radiologists agreed with TQIP on 72%, both p <0.001.

Conclusions:

Compared to physicians, registrar interpretation of imaging resulted in 40% overreporting of VAP to TQIP. The current definition of VAP ignores the biologic reality that there are multiple causes for opacities on chest X-ray. The subjectivity of imaging interpretation among physicians and the protean physiologic findings for VAP in trauma patients should preclude VAP from being used as a quality improvement metric in TQIP.