Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

PAL: Designing Conversational Agents as Scalable, Cooperative Patient Simulators for Palliative-Care Training

Neil K. R. Sehgal, Hita Kambhamettu, Allen Chang, Andrew Zhu, Lyle Ungar, Sharath Chandra Guntuku

TL;DR

PAL addresses the scarcity of scalable, emotionally intelligent palliative-care training by deploying an LLM-based conversational agent that simulates nuanced patient interactions in text and voice. Grounded in the NURSE empathy framework, PAL provides scenario-specific feedback after sessions, leveraging GPT-4o, Whisper, and TTS to create realistic, multimodal practice environments. The study with 17 U.S. clinicians and trainees shows PAL is usable and beneficial for reflection and skill development, but highlights tensions around emotional authenticity and feedback adaptability. The work contributes empirical support for AI-assisted palliative communication training, offers design guidance for modality-aware, emotionally sensitive simulators, and discusses implications for integrating AI to support emotional labor and cooperative learning in high-stakes care settings.

Abstract

Effective communication in serious illness and palliative care is essential but often under-taught due to limited access to training resources like standardized patients. We present PAL (Palliative Assisted Learning-bot), a conversational system that simulates emotionally nuanced patient interactions and delivers structured feedback grounded in an existing empathy-based framework. PAL supports text and voice modalities and is designed to scaffold clinical skill-building through repeated, low-cost practice. Through a mixed-methods study with 17 U.S. medical trainees and clinicians, we explore user engagement with PAL, evaluate usability, and examine design tensions around modalities, emotional realism, and feedback delivery. Participants found PAL helpful for reflection and skill refinement, though some noted limitations in emotional authenticity and the adaptability of feedback. We contribute: (1) empirical evidence that large language models can support palliative communication training; (2) design insights for modality-aware, emotionally sensitive simulation tools; and (3) implications for systems that support emotional labor, cooperative learning, and AI-augmented training in high-stakes care settings.

PAL: Designing Conversational Agents as Scalable, Cooperative Patient Simulators for Palliative-Care Training

TL;DR

PAL addresses the scarcity of scalable, emotionally intelligent palliative-care training by deploying an LLM-based conversational agent that simulates nuanced patient interactions in text and voice. Grounded in the NURSE empathy framework, PAL provides scenario-specific feedback after sessions, leveraging GPT-4o, Whisper, and TTS to create realistic, multimodal practice environments. The study with 17 U.S. clinicians and trainees shows PAL is usable and beneficial for reflection and skill development, but highlights tensions around emotional authenticity and feedback adaptability. The work contributes empirical support for AI-assisted palliative communication training, offers design guidance for modality-aware, emotionally sensitive simulators, and discusses implications for integrating AI to support emotional labor and cooperative learning in high-stakes care settings.

Abstract

Effective communication in serious illness and palliative care is essential but often under-taught due to limited access to training resources like standardized patients. We present PAL (Palliative Assisted Learning-bot), a conversational system that simulates emotionally nuanced patient interactions and delivers structured feedback grounded in an existing empathy-based framework. PAL supports text and voice modalities and is designed to scaffold clinical skill-building through repeated, low-cost practice. Through a mixed-methods study with 17 U.S. medical trainees and clinicians, we explore user engagement with PAL, evaluate usability, and examine design tensions around modalities, emotional realism, and feedback delivery. Participants found PAL helpful for reflection and skill refinement, though some noted limitations in emotional authenticity and the adaptability of feedback. We contribute: (1) empirical evidence that large language models can support palliative communication training; (2) design insights for modality-aware, emotionally sensitive simulation tools; and (3) implications for systems that support emotional labor, cooperative learning, and AI-augmented training in high-stakes care settings.

Paper Structure

This paper contains 27 sections, 1 figure, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. System Design
  4. Methods
  5. Findings
  6. Challenges in Learning
  7. Strengths of the Chatbot
  8. Modality-Specific Experiences
  9. Recommendations for Improvement
  10. Discussion
  11. Conclusion
  12. Detailed System Design
  13. Patient Agent Design
  14. User Interface
  15. Control Panel
  16. ...and 12 more sections

Figures (1)

  • Figure 1: The PAL system: in the main display a provider can interact with the patient bot. On the right side panel, there is a speech modal key, a button to finish the session which triggers feedback report generation, and a patient profile which details information about the patient and their reason for visiting. On the left side panel are different patient personas.