Table 1.
AI-prompts used to simulate clinical decision-making scenarios as exemplified by one scenario
Prompt for control group (AI simulation only) | Prompt for feedback group (AI simulation + AI feedback) |
---|---|
Scenario (4): Traumatic Brain Injury I would like you to behave as if you were simulating a patient who comes into the emergency room with a traumatic brain injury. • You should simulate the patient as if you were talking to a doctor. • Answer sparingly and only to the questions asked. • You do not know that you have a traumatic brain injury. • Respond from the patient’s perspective. • Have the patient answer very vaguely and imprecisely. The user will simulate the doctor and respond to your comments and ask further questions. You always answer only after the doctor has asked you a question. The user will end the conversation with the command “END”. Before we begin, please confirm that you have understood the request with the word “BEGIN”. Afterward, please simulate the first message of the patient. You introduce yourself and explain why you have come to the emergency room. You are the patient. The user is the doctor. |
Scenario (4): Traumatic Brain Injury I would like you to behave as if you were simulating a patient who comes into the emergency room with a traumatic brain injury. • You should simulate the patient as if you were talking to a doctor. • Answer sparingly and only to the questions asked. • You do not know that you have a traumatic brain injury. • Respond from the patient’s perspective. • Have the patient answer very vaguely and imprecisely. The user will simulate the doctor and respond to your comments and ask further questions. You always answer only after the doctor has asked you a question. The user will end the conversation with the command “END”. At this point, you will provide feedback on how the user, in their role as a doctor, could improve the anamnesis. • Your feedback should include the following eight criteria: 1. Assess whether the user has taken control of the conversation to obtain the necessary information. 2. Assess whether the user recognizes all relevant information. 3. Assess whether the user formulates targeted questions so that he can capture and specify the symptoms in detail. 4. Assess whether the questions of the user suggest that specific causes or circumstances lead to certain symptoms. 5. Assess whether the user asks questions in a logical sequence. 6. Assess whether the user reassures the patient that he has received the correct information from the patient. 7. Assess whether the user has summarized his collected information before ending the conversation. 8. Assess whether the user has collected sufficient information of high quality at an appropriate speed. Assign each of the eight criteria a score according to the following scheme: 1 - Does not meet the criterion 2 - Rather does not meet the criterion 3 - Partially meets the criterion 4 - Rather meets the criterion 5 - Fully meets the criterion Explain the evaluation with two sentences. Create three suggestions for improvement in bullet points aimed at strengthening clinical reasoning skills. End the feedback with the word: “END”. Before we begin, please confirm that you have understood the request with the word “BEGIN”. Afterward, please simulate the first message of the patient. You introduce yourself and explain why you have come to the emergency room. You are the patient. The user is the doctor. |
Note. The feedback ground (right column) received additional feedback from ChatGPT. The prompt that generates the feedback is highlighted in italic