Discussion

In this study using a hypothetical clinical scenario of low clinical probability of VAP, we found that HCP were less likely to treat a patient for VAP when the respiratory culture report included verbiage suggesting potential colonization in lieu of reporting specific organism identification, similar to if the culture report demonstrated only normal respiratory flora. This is consistent with our prior qualitative study where HCP reported a lower propensity toward VAP treatment in the presence of normal flora compared to recovery of a potential pathogen. ^{Reference Kenaa, O’Hara and Richert1}

Asymptomatic respiratory tract colonization without pneumonia or “asymptomatic bacterisputia” is common, particularly among patients who have been mechanically ventilated. ^{Reference Albin, Pogue, Petty and Kaye7} In the ICU, hypotension, leukocytosis and isolated fever in the absence of respiratory specific symptoms contribute toward respiratory culturing as part of the early sepsis recognition campaign with subsequent identification of the colonizing organisms. ^{Reference Kenaa, O’Hara, O’Hara, Claeys and Leekha4,Reference Albin, Saravolatz, Petrie, Henig and Kaye5} The cognitive bias associated with a positive culture – without regard to the underlying clinical scenario – was emphasized in our survey results where we observed a “dose-response” of progressively increasing likelihood of treatment from no growth (least likely to treat) to identification of normal flora, modified reporting comment suggesting colonization, light growth in semi-quantitative culture, low colony count in BAL, and high colony count in BAL (most likely to treat). This is similar to previous survey findings assessing biases in treatment of asymptomatic bacteriuria (ASB) where HCP are more likely to treat a positive culture with certain organisms, even with the same underlying clinical scenario of ASB. ^{Reference Grigoryan, Naik and Horwitz8} The current survey is limited in generalizability due to use of a single hypothetical clinical scenario and did not include consideration for neutrophil presence or quantification, but findings of antibiotic prescribing bias in VAP are supported by other recent diagnostic stewardship studies. ^{Reference Tripathi, Kenaa and Claeys9,Reference Albin, Troost and Saravolatz10} Other limitations include a small sample size and low response rate; however, we believe that this is representative of practice at our, and potentially other academic institutions.

A few studies have reported pragmatically leveraging modification of culture reports to reduce treatment of potential colonization. In a randomized controlled trial, Daley et al., a modified urine culture report resulted in significant improved appropriateness of treatment (80% vs 53%). ^{Reference Daley, Garcia, Inayatullah, Penney and Boyd11} This report stated the presence of bacteria but did not identify or quantify the organism, instead instructing clinicians to call the microbiology laboratory if concerns for UTI persisted. In a recent proof of concept study, we found that the modification of the respiratory culture report to suggest colonization results in clinically eligible patients (those with low clinical suspicion of VAP based on a pre-defined algorithm) significantly decreased treatment for VAP (46% vs 18%) without downstream negative sequalae of sepsis or bacteremia. ^{Reference Tripathi, Kenaa and Claeys9} When combined with clinical judgment and applied to the appropriate patient population, modified reporting of respiratory culture results in patients in the ICU has the potential to be a safe and effective diagnostic stewardship intervention. The results of this survey further support the need to consider how culture reporting can be leveraged to improve diagnosis and management of VAP.