With, not For: Co-Designing a Patient-Facing AI Companion Concept for the Emergency Department Waiting Area

Paper Title

With, not For: Co-Designing a Patient-Facing AI Companion Concept for the Emergency Department Waiting Area

Publication Info

• Topic area: Human-centered design of AI in healthcare settings
• Keywords: Emergency department, patient-facing AI, co-design, empathy, waiting area, clinical workflows, human-centered design, AI companion, participatory methods, patient experience

Background and Problem

• Problem / challenge: Emergency departments face overcrowding, staff shortages, and long waiting times, which negatively impact patient experiences. Existing AI solutions focus on clinical workflows, neglecting patient-centered needs such as empathy and reassurance.
• Significance: Addressing patient needs during ED waiting times can improve the overall care experience and reduce feelings of vulnerability and disconnection.
• Motivation and related work: Prior research has explored AI for triage, decision support, and documentation, but these systems often treat patients as workflow endpoints rather than partners. Human-computer interaction studies suggest AI can provide emotional support and meaningful engagement, but integrating these capabilities into ED workflows remains unexplored.

Solution

• Proposed approach: Development of an ephemeral AI companion to support ED patients during waiting times by providing reassurance, orientation, and empathetic presence without replacing human contact.
• Novelty:
  1. Empirical characterization of patient needs, concerns, and expectations for AI in the ED.
  2. Design principles for an AI companion concept balancing operational efficiency with relational needs.
  3. Translation of these principles into a conceptual prototype evaluated with patients during ED waits.
• Procedure and key techniques:
  • Four-phase study: preparatory fieldwork, co-creation workshops, design phase, and storyboard-guided interviews.
  • Iterative co-design process integrating patient and clinician perspectives.
  • Development of a storyboard to illustrate functionalities and empathetic qualities of the AI companion.

Results

• Concrete findings:
  • Patients valued AI functionalities such as preparation for consultation, navigation assistance, and process explanations.
  • Empathetic qualities like reassurance, agency, and reflective prompts were appreciated but raised concerns about authenticity and context.
  • Patients emphasized the need for human oversight and bounded use of AI.
• Advantage over baselines:
  • Unlike existing clinician-focused AI, the proposed companion addresses both operational and relational needs, transforming waiting times into purposeful experiences.
• Experiments / evaluation:
  • Co-creation workshops with 14 patients to identify needs and expectations.
  • Storyboard-guided interviews with 15 ED patients to assess the concept’s functionalities and empathetic qualities.
  • Comparative analysis of patient and clinician perspectives to align divergent needs.
• Limitations and future work:
  • Small sample size from a single hospital may limit generalizability.
  • Conceptual design without a functional prototype precludes clinical evaluation.
  • Future work includes prototyping, usability testing, integration into workflows, and longitudinal studies to assess systemic impacts.

Summary

This study proposes an ephemeral AI companion to support ED patients during waiting times by addressing both operational and relational needs. Through a four-phase co-design process, the researchers identified shared and divergent needs between patients and clinicians, translating them into design requirements for the AI companion. Evaluations with patients revealed appreciation for its functionalities and empathetic qualities, though concerns about authenticity and context were noted. While the concept remains at a conceptual level, future work aims to develop a prototype and evaluate its integration into clinical workflows. The findings highlight the potential of patient-centered AI to enhance care experiences without undermining human relationships.