Skip to content

Evaluating the Impact of AI-Generated Misleading Information on Diagnostic Accuracy and Confidence Calibration in Novice Medical Learners: A Multi-Stage Randomized Controlled Trial

Evaluating the Impact of AI-Generated Misleading Information on Diagnostic Accuracy and Confidence Calibration in Novice Medical Learners: A Multi-Stage Randomized Controlled Trial

Status
Recruiting
Phases
Unknown
Study type
Interventional
Source
ChiCTR
Registry ID
ChiCTR2500111932
Enrollment
Unknown
Registered
2025-11-07
Start date
2025-04-10
Completion date
Unknown
Last updated
2025-11-11

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

None

Interventions

Control Group:Participants were exposed only to the test questions without any additional explanations.
Correct Explanation Group:Participants were exposed to authoritative, AI-generated explanations targeting the correct answer for each question.
Misleading Explanation Group:Participants were exposed to plausible but incorrect AI-generated explanations crafted to support a pre-defined incorrect answer.

Sponsors

Cancer Hospital, Chinese Academy of Medical Sciences
Lead Sponsor

Eligibility

Sex/Gender
All

Inclusion criteria

Inclusion criteria: 1. Meet the operational definition of "primary medical learner" (i.e., undergraduate medical students who have completed basic medical courses but have not yet started major clinical rotations); 2. Native language: Chinese or equivalent reading and comprehension ability; 3. Agree to voluntarily participate and be able to independently complete online experiments; 4. Have a stable network connection and compatible devices (PC or tablet is recommended).

Exclusion criteria

Exclusion criteria: 1.Have completed major clinical rotations or have more than 6 months of clinical practice experience; 2. Systematically participated in business preparation courses for major medical license examinations; 3. Has participated in research on highly similar topics; 4. Failure to demonstrate genuine participation in the task (judged by the preset answering time standard, such as a total completion time of less than 3 minutes or more than 60 minutes).

Design outcomes

Primary

MeasureTime frame
Diagnostic Accuracy;Confidence Calibration;

Countries

China

Contacts

Public ContactDan Zhao

Cancer Hospital, Chinese Academy of Medical Sciences

zhaodan@cicams.ac.cn+86 10 8778 7223

Outcome results

None listed

Source: ChiCTR (via WHO ICTRP) · Data processed: Mar 19, 2026