Conversion of a Complex Legacy Antimicrobial Stewardship Clinical Decision Support System into Epic

Background. The hospital rules-based system (HRBS) was the first rules-based informatics program developed at Mayo Clinic in the 1980s in an effort to improve patient care through the reduction of medication errors, enhanced education, and clinician communication.¹ Computer-based antimicrobial monitoring (CBAM) was the first of the six HRBS subsystems to be developed and demonstrated positive patient outcomes and favorable cost reductions in medications. In January 2015, Mayo Clinic announced the selection of Epic as its partner for a single, integrated electronic health record (EHR). We describe the conversion of Mayo Clinic’s legacy decision support system CBAM into Epic’s antimicrobial stewardship program (ASP) module for eighty hospitals and clinics across four states.

Methods. The project kicked off in July 2015 and planning began between project analysts and subject matter experts. Epic’s antimicrobial stewardship strategy handbook was used to guide core implementation and build principles. In April 2016, the strategy shifted from using foundation Epic build to designing a customized ASP module that mirrored the complexity of Mayo’s CBAM system.

Results. The design, build, and implementation required substantial effort from key stakeholders from infectious disease physicians, infectious disease pharmacists, and project analysts. Bi-monthly meetings were set up in March 2016 to facilitate convergence and rapidly mitigate issues as it pertained to the Epic build. Over 200 legacy antimicrobial stewardship rules were reviewed, converged, and replicated into 75 customized rules for build in Epic. Estimated effort for rule build varied between 5 hours to 30 hours. Of the planned 75 rules, 17 were unable to be replicated by the first go-live in July 2017 due to complexity and constraints of the existing EHR tools. Project analysts innovated a solution that combined multiple tools in nontraditional methods to replicate minimum functionality that included rule interaction (i.e. dismiss and defer) and rule complexity (e.g. relative properties). A prototype of two rules using the new framework was introduced in the second go-live in November 2017. Given the successful implementation of the prototype, a full-scale conversion of existing rules and completion of the remaining 17 rules were implemented by the third go-live in May 2018. While this go-live marked the completion of converting the legacy systems into Epic, the initial weeks were characterized by false positives and usability issues. Significant optimization efforts of the ASP module persisted until the fourth go-live in October 2018 and the project officially closed on January 1, 2019. Despite the challenges, the ASP module is markedly improved from its earlier prototypes and its success and progression has served to inform Epic on multiple areas for future research and development.

Discussion. Implementation of Mayo Clinic’s legacy decision support systems for antimicrobial stewardship was achieved within project timelines. Innovative solutions with existing tools were required to fully replicate legacy functionality that allowed for interaction with individual rules.

Reference. 1. Wilson JW, Oyen LJ, Ou NN, et al. Hospital rules-based system: the next generation of medical informatics for patient safety. Am J Health Syst Pharm. 2005 Mar 1;62(5):499-505. PMID: 15745913