Design Patterns of Human-AI Interfaces in Healthcare

2025-07-16

AI Summary

The absence of systematic design guidance for medical human–computer interaction (HCI) interfaces undermines clinical usability and user trust. Method: Adopting a design science research paradigm, this study conducted semi-structured expert interviews and participatory workshops to systematically identify—within the medical domain for the first time—12 prototypical information entities and their associated presentation and interaction patterns, yielding reusable design guidelines. Contribution/Results: The proposed patterns underwent iterative empirical validation, significantly improving interface usability and clinician trust; they received strong endorsement from frontline practitioners. This work fills a critical methodological gap in AI-enabled medical interface design and distills transferable cross-domain principles for trustworthy human–AI collaboration, thereby providing both theoretical foundations and practical tools for reliable human–machine coordination in healthcare.

Technology Category

• Humans and AI: Intelligent User Interfaces
• Philosophy and Ethics of AI: Safety, Robustness & Trustworthiness
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior

Application Category

• User Modeling, Personalization and Recommendation: Practical large-scale studies of user experience
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Human-AI interfaces play a crucial role in advancing practices and research within the healthcare domain. However, designing such interfaces presents a substantial challenge for designers. In this paper, we propose systematic guidance for designing human-AI interfaces in typical healthcare scenarios by summarizing the design patterns for presenting and interacting with common information entities. To deepen our understanding of these 12 design patterns, we interviewed 12 healthcare professionals to explore potential usage scenarios and important considerations. Furthermore, we conducted workshops with 14 participants recruited online to evaluate our design patterns. Finally, we discussed the generalizability of the design patterns to other application domains, the limitations, and the future work.

Problem

• Designing effective human-AI interfaces in healthcare
• Providing systematic guidance for healthcare AI interfaces
• Evaluating design patterns for AI in healthcare scenarios

Innovation

• Systematic guidance for healthcare AI interfaces
• 12 design patterns from professional interviews
• Workshops to evaluate design pattern effectiveness