Physicians overestimate diagnostic probability of ventilator-associated pneumonia
Click Here to Manage Email Alerts
Key takeaways:
- The diagnostic probability of ventilator-associated pneumonia is overestimated.
- Attending and infectious disease physicians were more accurate in their LR estimates than trainees.
ICU physicians regularly overestimate the diagnostic probability of ventilator-associated pneumonia, according to a study published in Infection Control & Hospital Epidemiology.
“This study was inspired by my own clinical experiences with the challenging nature of accurately diagnosing ventilator-associated pneumonia (VAP), as well as personal interests in antimicrobial and diagnostic stewardship,” Nathaniel S. Soper, MD, a fellow in the division of infectious diseases at the University of Michigan Department of Internal Medicine, told Healio.
“I've seen numerous patient cases where different providers will look at the same set of lab results and clinical situation and come to widely different conclusions as to the likelihood of VAP. I wanted to know more about what factors drive [a] physician's concern for VAP and how that aligns with the evidence we have,” Soper said.
Soper and colleagues performed a clinical survey of ICU physicians to evaluate provider estimates of VAP diagnostic probability before and after isolated cardinal VAP clinical changes and VAP diagnostic test results at Michigan Medicine University Hospital.
In total, the researchers collected 133 survey responses, 30% of which were from attending physicians and 70% from residents or fellows and 66% of whom reported being “moderately” to “extremely confident” in their ability to accurately diagnose VAP.
The study showed that provider estimates of VAP diagnostic probability were consistently higher than evidence-based diagnostic probabilities, with data showing that survey respondents overestimated the baseline probability of VAP with a median estimated probability of 20% relative to evidence-based baseline probability of 16%.
These rates were similar following presented isolated cardinal VAP clinical changes and VAP diagnostic test results with the median provider-estimated pretest probability of VAP following patient development of isolated purulent endotracheal secretions being 34% relative to an evidence-based pretest probability of 12.6%.
The median provider-estimated post-test VAP probability then rose to 80% relative to an evidence-based post-test probability of 16.5% after receipt of a positive bronchoalveolar lavage culture.
The study also showed that imputed likelihood ratios (LRs) from provider-estimated diagnostic probabilities were consistently higher than evidence-based LRs — with the differences being most notable for positive bronchoalveolar lavage culture (provider-estimated LR = 5.7 vs. evidence-based LR = 1.4; P < .01), chest radiograph with air bronchogram (6 vs. 3.6; P < .01), and isolated purulent endotracheal secretions (1.6 vs. 0.8; P < .01).
Finally, the study showed that attending physicians and infectious disease physicians were more accurate in their LR estimates than trainees (P = .04) and non-ID physicians (P = .03).
“VAP is very challenging to accurately diagnose, and we tend to over diagnose and overtreat. It's important to ensure that the diagnostic probabilities we implicitly or explicitly use in clinical practice are consistent with available evidence,” Soper said. “Beyond this, it's likely that additional diagnostic stewardship interventions may be needed to improve diagnostic accuracy and reduce overtreatment for challenging diagnoses like VAP.”
Collapse
Read more about
Click Here to Manage Email Alerts