Background Handoffs or care transitions from the operating room (OR) to intensive care unit (ICU) are fragmented and vulnerable to communication errors. Although protocols and checklists for standardization help reduce errors, such interventions suffer from limited sustainability. An unexplored aspect is the potential role of developing personalized postoperative transition interventions using artificial intelligence (AI)-generated risks. Objectives This study was aimed to (1) identify factors affecting sustainability of handoff standardization, (2) utilize a human-centered approach to develop design ideas and prototyping requirements for a sustainable handoff intervention, and (3) explore the potential role for AI risk assessment during handoffs. Methods We conducted four design workshops with 24 participants representing OR and ICU teams at a large medical academic center. Data collection phases were (1) open-ended questions, (2) closed card sorting of handoff information elements, and (3) scenario-based design ideation and prototyping for a handoff intervention. Data were analyzed using thematic analysis. Card sorts were further tallied to characterize handoff information elements as core, flexible, or unnecessary. Results Limited protocol awareness among clinicians and lack of an interdisciplinary electronic health record (EHR)-integrated handoff intervention prevented long-term sustainability of handoff standardization. Clinicians argued for a handoff intervention comprised of core elements (included for all patients) and flexible elements (tailored by patient condition and risks). They also identified unnecessary elements that could be omitted during handoffs. Similarities and differences in handoff intervention requirements among physicians and nurses were noted; in particular, clinicians expressed divergent views on the role of AI-generated postoperative risks. Conclusion Current postoperative handoff interventions focus largely on standardization of information transfer and handoff processes. Our design approach allowed us to visualize accurate models of user expectations for effective interdisciplinary communication. Insights from this study point toward EHR-integrated, flexibly standardized care transition interventions that can automatically generate a patient-centered summary and risk-based report.
- Continuity of care
- care transition
- intensive and critical care
- machine learning
- requirements analysis and design