PainTrain-AI Pilot: Feasibility, Usability (SUS) and Concept-Shift in Primary Care and a Specialized Setting

Overview

This pilot will assess feasibility and usability (SUS) and explore concept shift (reduction of neurophysiology misconceptions) after exposure to a behavioral educational intervention (PainTrain AI), a digital clinical simulation "fenced" by evidence (RAG) for safe, non diagnostic training. The study will be conducted in Primary Care and in a specialized comparative setting, without patient data collection. Primary endpoint: SUS score. Secondary endpoints: concept shift, adherence to micromodules, and interaction latency as friction metric. The platform does not generate clinical advice; it retrieves validated teaching content.

This feasibility pilot will evaluate the use of PainTrain AI, a behavioral educational intervention designed to improve clinical reasoning about chronic pain through simulated dialogues with virtual standardized patients. The platform uses a Retrieval Augmented Generation (RAG) safety architecture that restricts all generated responses to validated pedagogical content and does not provide clinical advice. No patient information is collected. The study includes health professionals from Primary Care and a specialized comparative setting. Participants will complete a baseline assessment and then interact with the PainTrain AI simulation for a defined training period. Usability will be assessed using the System Usability Scale (SUS). Concept shift will be measured as the reduction in predefined neurophysiology misconceptions scored through a standardized rubric. Additional metrics include adherence to micromodules and interaction latency as indicators of technological friction. This pilot aims to determine feasibility, acceptability, and preliminary signals of educational effectiveness to inform a future implementation trial in Primary Care. The intervention is educational only, non diagnostic, and GDPR compliant.