Student Performance (Conversation Data)
Description:
This database contains the full text of the conversations between students and AI-simulated patients during activities in which students take the medical history through interactive dialogue. It only applies to activities configured for text-based or voice-based interactions. Each row corresponds to a single message exchanged between the student and the simulated patient.
For text-based interactions, the transcript is fully accurate, reflecting exactly what the student typed and what the simulated patient responded. For voice-based interactions, the transcript is automatically generated through speech recognition and may contain transcription errors or omissions.
When analyzing communication skills or providing feedback on conversation content, instructors are encouraged to use text-based interactions to ensure data accuracy. If voice-based interaction is essential for the learning experience, it is recommended to assess student performance using complementary items, such as follow-up open-ended questions or evaluation of the student’s visit notes in the patient’s electronic medical record. As AI transcription models improve, the accuracy of voice-based conversation data will be reassessed in future updates.
This report is provided in long format, where each record represents one message within a conversation. Use clinicalActivityId to join to the Clinical Case Activities report for activity-level context (purpose, complexity, sections and items). Use attemptId to group the messages of a single session and to join consistently with student-response data. Use userId to join with the Users database; username is a convenience display name. The medicalHistoryDuration field provides the total duration of the medical-history interaction for the session (per attempt), repeated on each message row of that attempt and expressed in seconds.
Variables:
clinicalActivityId — Identifier of the Clinical Mind AI activity in which the conversation occurred. Join with Clinical Case Activities to retrieve activity-level details.
content — Text of the message exchanged. For text interactions, this is the exact written input and response; for voice interactions, this is the ASR transcription. Stored as plain text.
country — Country of the institution where the activity took place.
attemptId — Unique identifier for the specific conversation attempt by a student during the activity. Use to group messages into a session and to join with student-response data for the same attempt.
courseId — Identifier of the course in which the activity was assigned.
createdAt — Date and time when the message record was created in the system.
id — Unique identifier of the individual message within the conversation.
institution — Name of the educational institution associated with the activity.
patientName — Display name of the AI-simulated patient participating in the conversation.
role — Speaker role in the exchange: user (student) or assistant (AI-simulated patient).
timestamp — Date and time when the message was exchanged. Same format as createdAt; may differ slightly from createdAt due to processing latency. Use this field to order messages within an attempt.
updatedAt — Date and time of the last update to the message record (usually identical to createdAt unless post-processing occurred, such as transcription correction).
userId — Unique identifier of the student participating in the conversation. Join with the Users database to obtain user details.
username — Display name of the student who participated in the conversation (convenience field for browsing and reporting).
medicalHistoryDuration — Numeric duration of the medical-history interaction for the session, expressed in seconds (for example, 904 = 15 minutes and 4 seconds). Repeated on each row that shares the same attemptId.