Abstract We sought to automate antibiogram creation at our multi-hospital institution with custom features including transparency into data filtering, expanding the granularity to include multi-resistant organisms, and additional conditional formatting based on sensitivity percentages. We automated creation of the antibiogram using a python framework with embedded SQL scripts. The sensitivity data for isolates from the prior year was retrieved from our lab information system. It was filtered to include only inpatient, emergency department, and observation patients across all seven hospitals in our system. Additionally, the data was filtered to only include the first organism per patient per year. Finally, organisms meeting specific sensitivity criteria (i.e. methicillin resistant Staphylococcus aureus) were included as separate rows beneath the parent organism. This data was inserted into an excel workbook with customized formatting to include the institutional logo, coloring based on sensitivity percentages, and customizable comments. Each of these steps are easily editable year-to-year to allow for changing requirements. At each step of filtering, the integrity of the data was extensively validated to make sure it was performing as intended. The antibiograms are generated within a few minutes in a fully automated manner (Figure 1). These will be distributed systemwide after review by the Medical Microbiology directors and Antimicrobial Stewardship Committee in January of each year. The code was made in an easily editable, modular way such that requests for additional separation of organisms by sensitivity, removing drugs not pertinent to our formulary, and editing what organisms appear is easily managed by the informatics team on a year-to-year basis.
Read full abstract