Background: The concept of health-care-associated pneumonia (HCAP) exists to identify patients infected with highly resistant pathogens. It is unclear how precise this concept is and how well it performs as a screening tool for resistance.
Methods: We retrospectively identified patients presenting to the hospital with pneumonia complicated by respiratory failure. We examined the microbiology of these infections based on pneumonia type and determined the sensitivity and specificity of HCAP as a screen for resistance. Through logistic regression and modeling, we created a scoring tool for determining who may be infected with resistant pathogens.
Results: The cohort included 190 subjects (37% with ARDS) and we noted resistant pathogens in 33%. Resistance was more common in HCAP (78% vs 44%, P = .001). HCAP alone performed poorly as a screening test (sensitivity and specificity 78.3% and 56.2%, respectively). Variables independently associated with a resistant organism included immunosuppression (adjusted odds ratio [AOR] 4.85, P < .001), long-term care admission (AOR 2.36, P = .029), and prior antibiotics (AOR 2.12, P = .099). A decision rule based only on these factors performed moderately well at identifying resistant infections. The presence of HCAP itself, based on meeting defined criteria, was not independently associated with resistance using logistic regression to control for covariates.
Conclusions: HCAP is common in patients presenting to the hospital with pneumonia leading to respiratory failure. The HCAP concept does not correlate well with the presence of infection due to a resistant pathogen. A simpler clinical decision rule based on select HCAP criteria performs as well as the HCAP concept for potentially guiding antibiotic decision making.