957. Interactive Antibiogram Decision Support Tool Predicated Upon Infection Site and Probable Pathogens

Laurel Legenza, Jared Zunenshine,Nasia Safdar,Nasia Safdar, Colin Guest, Harshvardhan Jadhav,Thomas R Fritsche,John D D Lee

Open Forum Infectious Diseases(2022)

引用 0|浏览2
暂无评分
摘要
Abstract Background Current methods for displaying summary antibiograms are static, have limited utility and are not user friendly. We present an interactive web application for visualization of antibiogram data. Methods Antibiotic susceptibility data for first, non-duplicate patient infection isolates with source/site data were collected from 2009–2018 at an integrated health system for the visualizations. Iterative user-centered design sessions were held with healthcare providers to guide prototype development, including assessment of source groupings and rational for aggregating data. Three providers participated in initial and follow-up sessions (6 total) with study personnel. Data were aggregated by agreed-upon infection-source and pathogen; antibiotic results with < 30 isolates were excluded. Results The multi-year dataset included 119,333 non-duplicative isolates from 194 unique infection sites; nearly half were E. coli (n=50,404). Final infection source groupings with interactive data visualization fields are shown in Figure 1. First an infection source is selected; the visualization then shows pathogen prevalence in descending order. Providers can view susceptibility results for all antibiotics or deselect all and view only those under consideration, such as the four displayed in Figure 1. Resulting bar charts allow providers to view which pathogens tested have greater than 80% or 90% susceptibility to each antibiotic. Susceptibilities for multiple pathogens can also be visualized with the dodged feature (Figure 2). The large dataset collected supports source-specific susceptibility results (e.g. eye infections). The visualization allows providers to see how susceptibility by pathogen may cross clinically meaningful thresholds of 80% and 90% for different infection sources, such as urinary tract compared to bloodstream infections. Feedback from providers on the prototype was overwhelmingly positive. Figure 1.Visualization of urine isolates, E. coli and user selected antibiotics Final source groupings: Urine, Skin/Soft Tissue, Blood, Respiratory Secretions, Fluid, Ear, Bone, Vaginal/Perianal, Eye, and Abscess. Figure 2.Selected antibiotic susceptibilities for multiple pathogens can be visualized simultaneously with the dodged feature. Conclusion Our interactive antibiogram data clinical decision support tool extends the utility of the traditional lab-generated antibiogram. Importantly, future iterations will include visualizing changes by year, by clinical setting (inpatient vs. outpatient) and an assessment of uptake of this tool in the clinical setting. Disclosures All Authors: No reported disclosures.
更多
查看译文
关键词
infection site
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要