JMIR Publications

ABSTRACT

Background:

Pediatric trainees often feel underprepared for difficult conversations, especially around end-of-life care, prognostic uncertainty, and serious illness. Despite Accreditation Council for Graduate Medical Education (ACGME) requirements for communication competency, barriers such as limited faculty time, cost, and scheduling persist. Traditional approaches like simulation and standardized patient encounters improve confidence but are resource-intensive and offer limited opportunities for repeated practice.

Objective:

This study aimed to evaluate the feasibility and learner experience of a just-in-time, artificial intelligence (AI)–driven chatbot designed to support pediatric providers in practicing difficult conversations.

Methods:

We conducted a mixed-methods feasibility study across pediatric providers (residents and advanced practice providers) rotating through NICU, PICU, and ED settings. Participants engaged in an AI chatbot–based simulated conversation and received structured feedback. Pre- and post-intervention surveys assessed perceived competence, confidence, feasibility, and usability. Qualitative feedback was obtained through open-ended responses and optional debrief sessions. Descriptive statistics and exploratory paired analyses were performed.

Results:

A total of 31 participants completed the intervention. Pre and post survey responses were rated on a Likert scale of 1 to 5. Post-participation feasibility of the chatbot was rated a mean score of 3.81, while user-friendliness rated 4.16. The participants found the AI-generated feedback to be helpful with a mean score of 3.54. Most participants (67.7%) reported no change in self-confidence after a single interaction (mean change +0.06 on Likert scale). However, participants who repeated the simulation demonstrated consistent improvement in objective communication performance.

Conclusions:

AI chatbot–based communication training is feasible, scalable, and well accepted among pediatric providers. While a single exposure does not significantly improve self-confidence, repeated use may enhance objective communication performance. This approach shows promise as a supplement to traditional communication curricula.