How to Develop and Implement a Computerized Decision Support System Integrated for Antimicrobial Stewardship? Experiences From Two Swiss Hospital Systems

Abstract

Background: Computerized decision support systems (CDSS) provide new opportunities for automating antimicrobial stewardship (AMS) interventions and integrating them in routine healthcare. CDSS are recommended as part of AMS programs by international guidelines but few have been implemented so far. In the context of the publicly funded COMPuterized Antibiotic Stewardship Study (COMPASS), we developed and implemented two CDSSs for antimicrobial prescriptions integrated into the in-house electronic health records of two public hospitals in Switzerland. Developing and implementing such systems was a unique opportunity for learning during which we faced several challenges. In this narrative review we describe key lessons learned. Recommendations: (1) During the initial planning and development stage, start by drafting the CDSS as an algorithm and use a standardized format to communicate clearly the desired functionalities of the tool to all stakeholders. (2) Set up a multidisciplinary team bringing together Information Technologies (IT) specialists with development expertise, clinicians familiar with "real-life" processes in the wards and if possible, involve collaborators having knowledge in both areas. (3) When designing the CDSS, make the underlying decision-making process transparent for physicians and start simple and make sure to find the right balance between force and persuasion to ensure adoption by end-users. (4) Correctly assess the clinical and economic impact of your tool, therefore try to use standardized terminologies and limit the use of free text for analysis purpose. (5) At the implementation stage, plan usability testing early, develop an appropriate training plan suitable to end users' skills and time-constraints and think ahead of additional challenges related to the study design that may occur (such as a cluster randomized trial). Stay also tuned to react quickly during the intervention phase. (6) Finally, during the assessment stage plan ahead maintenance, adaptation and related financial challenges and stay connected with institutional partners to leverage potential synergies with other informatics projects.

Keywords: antimicrobial stewardship; cluster randomized controlled trial; computerized decision support system; digital health; implementation; multidisciplinary; usability testing; user training.

Figure 1

(A) Workflow and rules of the COMPASS algorithm drawn in collaboration with business modelers. (B) Workflow and rules of the COMPASS algorithm initially drawn by the PI.

Figure 2

Screenshots of the CDSS. (A) Dropdown list of indications, (B) selection of free-text indication by selecting “Autre” (“Other”) in the dropdown list and (C) entering free-text. (D) List of justification provided for guideline rejection, (E) by selecting “Autre” (Other), the specific box for free-text appears.