BRAVE – A Generative AI Simulator for Practicing Breaking Bad News with Reflective, Protocol Based Feedback
Unmet need
- Breaking bad news (BBN) is a complex, emotionally charged skill, yet most physicians and medical students receive little structured training
- Traditional BBN training (bedside observation, role‑play, standardized patients) is costly, logistically demanding, and ethically constrained, limiting hands‑on practice
- Many learners feel unprepared and stressed, which can lead to avoidance of BBN encounters or poor‑quality communication that harms patients, families, and clinicians
The technology
- BRAVE (Breaking Bad News with AI Reflective Virtual Experience) is a GenAI‑based, text‑only simulator that enables learners to conduct a BBN conversation with a virtual patient and receive structured, protocol‑based feedback
- A large language model drives two personas: “Oron,” a realistic virtual patient, and “Dani,” an empathic feedback provider who delivers guided, pedagogical commentary
- BRAVE embeds experiential learning: simulated dialogue, brief dialogic reflection, and feedback mapped to the SPwICES BBN protocol, with the option to repeat practice
- The system generates responses dynamically in real time and was iteratively stress‑tested by clinical and AI experts to reduce hallucinations and ensure safe, consistent behavior.

Potential applications
- Integration into undergraduate and graduate medical curricula as an experiential BBN module to boost confidence and willingness to conduct difficult conversations.
- Use across health professions (e.g., nursing, genetics, mental health, veterinary medicine) where delivering difficult information is central to care.
- On‑demand practice tool for clinicians seeking a psychologically safe environment for refreshing or upgrading communication skills.
- Expansion to multiple diagnoses, patient profiles, cultures, and languages to support diverse training needs and institutional settings.
Reference:
Email
Linkedin
WhatsApp
Print/PDF