Using Decision Trees to Extract Information from RCT Reports
Author Information
Author(s): Chung Grace Y, Coiera Enrico
Primary Institution: Centre for Health Informatics, University of New South Wales
Hypothesis
Can decision trees effectively represent and extract critical information from randomized controlled trial reports?
Conclusion
Decision trees could be a suitable construct to guide machine summarization of RCTs and indicate report quality.
Supporting Evidence
- 73.8% of the analyzed abstracts were primary studies with a single population assigned to two or more interventions.
- 68% of primary RCT abstracts were structured.
- 84% reported the total number of study subjects.
Takeaway
This study looks at how decision trees can help computers understand important information in medical research papers, making it easier for doctors to find what they need.
Methodology
The study analyzed 455 RCT abstracts for decision tree elements and assessed their quality and structure.
Limitations
The study may not represent all RCTs as many are not indexed in Medline, and the method for identifying structured abstracts may have errors.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website