University of Illinois Urbana-Champaign

Comparing articles identified as Randomized Controlled Trials: MEDLINE, Cochrane, and the RCT Tagger

Kansara, Yogeshwar; Hoang, Linh; Schneider, Jodi

Permalink

https://hdl.handle.net/2142/102074

Description

Title
Comparing articles identified as Randomized Controlled Trials: MEDLINE, Cochrane, and the RCT Tagger
Author(s)
  • Kansara, Yogeshwar
  • Hoang, Linh
  • Schneider, Jodi
Issue Date
2018-10-31
Keyword(s)
  • Machine Learning
  • Error Analysis
  • Randomized Controlled Trials
  • evidence-based clinical practice
Abstract
Randomized Controlled Trials (RCTs) are considered the gold standard of medical knowledge about treatment effects. RCTs are used in evidence-based clinical practice and for the production of systematic reviews. Determining whether or not an article is a RCT, thus, can be useful for several search applications, including supporting clinicians in finding high-quality information, and providing high-specificity searches for systematic searching. In this study, we pilot a methodology for evaluating a machine learning tool called Tagger, that aims to distinguish Randomized Controlled Trials reports from other medical literature. The goal of our evaluation is to assess the feasibility of using the tool in real-life systematic review projects by examining Tagger's technical performance in identifying RCT reports included in a sample of 895 systematic reviews. We present the evaluation results and discuss possible improvements for the tool.
Type of Resource
  • text
  • other
Language
en
Permalink
http://hdl.handle.net/2142/102074

Owning Collections

Student Publications and Research - Information Sciences PRIMARY
Publications, conference papers, and other research and scholarship of iSchool students.