Effect of an Artificial Intelligence-Assisted Antibiotic Susceptibility Test on Reducing the Mortality Rate of and Antibiotics Use in Patients with Bacteremia: A Prospective Observational Cohort Study

<div data-language="eng" data-ev-field="abstract">Background: In real world setting none of the mass spectrometry-based antibiotic susceptibility test (AST) AI system has been robustly validated and indeed deployed.&lt;br&gt;&lt;br&gt;Methods: The study extensively validated and deployed an AI-assisted AST prediction system XBugHunter. A total of 90,064 consecutive patients were included in the model&rsquo;s development. During the deployment, 155 consecutive patients with &lt;i&gt;Staphylococcus aureus bacteremia&lt;/i&gt; (SAB) was analyzed, whereas the historical control cohort included 455 patients.&lt;br&gt;&lt;br&gt;Findings: The predictive sensitivity and specificity for S. aureus (oxacillin) were 0&middot;95 and 0&middot;97, respectively. The reduction in AST turnaround time was 34&middot;06 h. Death within 28 days was 13&middot;55% for the SAB patients in the XBugHunter intervention group, which was significantly lower than 17&middot;58% in the historical control cohort. A total of 114&middot;77 defined daily doses of inappropriate antibiotics were avoided by deploying XBugHunter.&lt;br&gt;&lt;br&gt;Interpretation: This work showed that XBugHunter can prevent inappropriate antibiotic use, and the adjustment in antibiotic treatment can yield a lower mortality rate.&lt;br&gt;&lt;br&gt;Funding: This work was supported by Chang Gung Memorial Hospital [CMRPG3M0851], and the Ministry of Science and Technology, Taiwan [111-2320-B-182A-002-MY2].&lt;br&gt;&lt;br&gt;Declaration of Interests: The authors have no affiliations with or involvement in any organization or entity with any financial interest or nonfinancial interest in the subject matter or materials discussed in this manuscript.&lt;br&gt;&lt;br&gt;Ethics Approval: This study was approved by the institutional review board of the Chang Gung Medical Foundation (Nos. 201901578B0 and 202101884B1).<br/></div> &copy; 2023, The Authors. All rights reserved.