Abstract Background The management of patients with suspected infections requires identification of the infectious etiology to determine appropriate use of antibiotics. However, differentiating viral from bacterial infection (and co-infection) is often challenging as clinical presentations can be similar and existing diagnostics sometimes fail to identify a clinically relevant pathogen. A host-response test (MeMed BV®, MMBV) that relies on computational integration of three proteins (TRAIL, IP-10 and CRP) measured from blood or serum has demonstrated high diagnostic performance for differentiating bacterial from viral infections, with a negative predictive value >95% across multiple studies. This report evaluates real-world use of MMBV at a micro-hospital Emergency Department (ED) and associated antibiotic prescribing. Methods The study is a retrospective analysis of real-world data collected between January and June 2023. MMBV was ordered by providers at a micro-hospital ED in Tucson, AZ at provider discretion as part of routine care. Prescription among cases with MMBV score <35 was analyzed. This cutoff is indicated by the manufacturer to indicate a viral or other non-bacterial etiology. If an antibiotic was prescribed, the authors conducted a chart review to adjudicate whether the prescription was warranted. Results Data was evaluated between January and June 2023; 116 MMBV tests were ordered by six providers. Most tests (81.9% (95%CI: ±7.0%)) had an MMBV score <35 (mean score 11.3 with standard deviation 10.8). Among these, 92.6% (95%CI: ±5.3%) of cases with viral results were not prescribed antibiotics. MMBV use increased throughout the study period from 12 tests in the first month to 25 in the last month (r = 0.88; p<0.001). There were seven cases where providers chose to prescribe despite a viral MMBV result. Upon chart review, 2/7 were unwarranted prescriptions. Conclusions In the emergency settings, integrating MMBV can effectively guide clinical decision-making, potentially reducing unnecessary antibiotic use. Further research in diverse healthcare settings is needed to validate these findings.