DESCRIPTION (provided by applicant): Inappropriate use of antibiotics is extremely common and results in increased antimicrobial resistance, increased medication-related adverse events (including Clostridium difficile-associated colitis), and increased healthcare costs. Antimicrobial stewardship programs (ASPs) encourage evidence-based decisions on initiating, streamlining, and discontinuing antibiotic therapy while ensuring optimal dosing and route of administration and limiting unintended consequences. However, best practices for ASP interventions are uncertain. Supporting evidence for most interventions is limited and formal studies on decision-making processes of ASP teams and targeted users are lacking. Numerous surveys (including work in the VA) indicate variations in ASP structure, implementation, and processes as well as a need for better informatics tools to guide and refine the development and management of ASPs. We propose a translational research program to develop, implement, and evaluate cognitive support-based population analytics to improve medication decision-making as applied to antimicrobial stewardship (AS). To do this we will accomplish the following Aims: 1. Characterize existing structural aspects and decision-making processes relevant to AS that are predictive of quality of care measures in antimicrobial use; 2. Develop new global and specific measures of quality for antimicrobial use that will be compiled into visual analytic displays; 3. Implement and iteratively test visual analytics as cognitive support tools to improve antimicrobial stewardship t four pilot facilities.; 4. Conduct a cluster randomized trial to assess impact of a refined antibitic stewardship informatics package on antimicrobial use. For Aim 1, we will collaborate with two VA operational partners, the VA Healthcare Analysis & Information Group (HAIG), which will conduct a nationwide survey on current ASP practices, and the Evidence- based Synthesis Program (ESP), which will review the peer-reviewed evidence on effectiveness of specific ASP elements, in order to determine which structural and process variables relating to ASPs best predict variation in five separate metrics that capture different qualities in antimicrobial usageat a facility level in order to develop recommendations for ASP structures and processes. For Aim 2, we will develop novel measures of antimicrobial use that reflect clinical decision-making and that can be used to improve stewardship management decisions. Using the VA system-wide VINCI database we will provide individualized facility-level data to stewardship teams regarding antimicrobial usage in three common disease states: pneumonia, urinary tract infection, and skin and soft tissue infection at 3 canonical branch points of treatment: Choice (i.e. initial choce of empiric antimicrobial therapy); Change (i.e. adjustment of therapy based on laboratory test results and clinical response); and Completion (i.e. antimicrobial duration and/or discontinuation). We will translate concepts of antibiotic quality derived in Aim 1 into quantitatie metrics within our decision framework and use these metrics in constructing visual analytics to serve as cognitive support tools for ASPs and other stakeholders. For Aim 3, we will iteratively test and implement the visual analytic tools at four VA facilities, with input from cognitive task analyses and focus groups. For Aim 4, we will test the implementation of a stewardship package consisting of recommendations from Aim 1 and the visual analytics from Aims 2 and 3 in a cluster-randomized trial of 20 VA facilities. Our primary outcome will be a comparison of antimicrobial usage in the first five days of hospital admission among patients with a principal ICD-9 diagnosis code suggesting a non-infectious cause for admission. Secondary outcomes will include comparisons of antimicrobial use quality metrics, rates of multidrug-resistant organisms, costs, and other global measures (i.e. changes in readmission rates or 30 day mortality). If the findings are supportive, we plan on implementation of our package across VHA, via our participation in the VA Antimicrobial Stewardship Task Force.

Over half of all veterans admitted to an acute-care VA medical center (VAMC) nationwide receive antibiotics. Unfortunately, antibiotics are frequently given for inappropriate reasons or dosed inappropriately for the clinical situation at hand, which can result in increased resistance to antibiotics, adverse events associated with antibiotic use, and increased costs. Antimicrobial stewardship programs (ASPs) that monitor antibiotic use at individual VAMCs can decrease inappropriate use, but the strategies for how they can best accomplish this goal are unclear. In this proposal, we will 1) determine the institutional factors that best predict the ability of ASPs to effect change in antibiotic use; 2) formulate metrics that reflect quality in antibiotic use that will be developed into visual displays that will allow for comparisons within and between VAMCs; 3) pilot test the visual displays as cognitive support tools at 4 VAMCs; and 4) test implementation of a package consisting of recommendations and cognitive support tools in a cluster-randomized trial of 20 VAMCs nationwide.