Medical Exam Question Answering Dataset from US and China

Medical Meadow MedQA is a large-scale open-domain question answering dataset derived from professional medical exams. The dataset was created by the authors of the paper 'What Disease does this Patient Have?', with contributions from the medalpaca organization, and was last updated in April 2023. It contains questions and answers sourced from US and Chinese medical licensing examinations, along with supporting textbook passages.

Use Cases

• Benchmarking model performance on US and Chinese medical exam questions.
• Training models to answer open-domain medical questions using provided textbook passages.
• Analyzing differences in medical knowledge assessment between US and China exam sources.

Strengths

• Derived from professional medical licensing exams, ensuring real-world relevance.
• Contains multiple sources (US and China), offering cross-regional perspective.
• Includes supporting textbook passages for context in question answering.

Limitations

• Specific row counts, column names, and dataset size are not provided.
• The dataset may have a class or topic imbalance reflecting exam structures.
• Data is static, last updated in 2023, and may not reflect the most recent exam content.

Provenance

Source

Original data and code from the GitHub repository for the MedQA paper by Jin, Di, et al.

Collection Method

Compiled from medical licensing examinations in the United States and Mainland China.

Time Range

null

Freshness

null

Geography

United States and Mainland China.