Human-Centered Design and Usability Evaluation of an EHR-Integrated Clinical Decision Support Prototype for Pediatric...

Objectives To design and evaluate the usability of an electronic health record (EHR)-integrated clinical decision support (CDS) prototype supporting pediatric surgical site infection (SSI) risk estimation and preoperative optimization using a multi-phase human-centered design approach. Methods We conducted a three-phase study at a quaternary pediatric academic health system involving clinicians from anesthesia and seven surgical specialties. Phase 1 used contextual inquiry to examine preoperative information gathering, infection risk assessment, and coordination workflows to define user needs. Phase 2 used structured mock-up design review sessions to refine interface design and establish prototype requirements. Phase 3 involved formative usability testing of an interactive prototype using a think-aloud approach, followed by the Healthcare Systems Usability Scale (HSUS). Qualitative data were analyzed using thematic analysis. HSUS scores were analyzed descriptively. Results Nine clinicians participated in each phase (10 total participants). Contextual inquiry identified challenges including manual chart review burden, task coordination outside the EHR, and largely qualitative risk assessment practices. Design review translated these needs into a centralized preoperative workspace integrating patient summaries, contextualized SSI risk with contributing factors and benchmarks, and role-based task tracking. During usability testing, clinicians reported reduced chart review burden and better visibility of high-priority information, while identifying opportunities to improve visual prioritization, transparency of risk logic and recommendations, and mechanisms to correct or flag inaccurate content. HSUS results indicated good to excellent usability, with an overall usability score of 80%. Workflow integration received the highest ratings (90%), followed by patient safety and decision effectiveness (86%). Conclusions This study demonstrates that translating predictive models into clinical practice requires integration within real-world workflows, transparent presentation of information, and design aligned with the cognitive and operational realities of perioperative care. These findings provide a foundation for iterative refinement, implementation, and prospective evaluation of EHR-integrated pediatric SSI decision support.