Traditional-intensive versus technology-enhanced approach for search and screening in systematic reviews (TITE project)




Poster session 2 Thursday: Evidence synthesis - methods / improving conduct and reporting


Thursday 14 September 2017 - 12:30 to 14:00


All authors in correct order:

Rada G1, Brandt L2, Bravo-Soto GA3, Rada G3, Edwards J3, Llovet V3, Lobos D3, Agoritsas T4, Lytvyn L3, Siemieniuk R5, Vandvik P2
1 Epistemonikos foundation; Centro Evidencia UC, Pontificia Universidad Católica de Chile, Chile
2 Department of Internal Medicine, Sykehuset Innlandet Hospital Trust, Gjøvik; Faculty of Medicine, Institute of Health and Society, University of Oslo, Oslo, Norway
3 Epistemonikos foundation, Chile
4 Divisions of Clinical Epidemiology and General Internal Medicine, University Hospitals of Geneva, Geneva, Switzerland
5 McMaster University, Hamilton, Ontario,, Canada
Presenting author and contact person

Presenting author:

Gabriel Rada

Contact person:

Abstract text
Background: The BMJ Rapid Recommendations project aims to rapidly produce evidence summaries and trustworthy clinical practice recommendations within 90 days after identification of potentially practice-changing evidence. The process involves conducting high-quality systematic reviews within 45 days. The TITE project is testing if a technology-enhanced approach can be as much or more accurate and efficient than the traditional-intensive approach.

Objectives: In TITE Q2I (Traditional-intensive versus technology-enhanced / question-to-inclusion time) we will compare both approaches for the process of developing a search strategy, independently screening records and selecting eligible articles in full text.

Methods:The traditional-intensive approach is based on a wide collaborative network of clinicians, patients and methodologists following methods and processes defined in the BMJ Rapid Recommendations protocol.
Technology-enhanced approach combines reuse of information from Epistemonikos database, software to facilitate search strategy creation, automated execution and deduplication, a screening platform (Collaboratron TM), machine learning among other technologies.
We will select 5 systematic reviews from the BMJ Rapid Recommendations project for which the question-to-inclusion process is already completed. We will set teams of at least two researchers that did not integrate the original review team which will receive the inclusion criteria and a list of electronic databases to be searched.
We will measure time, search efficiency and accuracy.

Results: Both traditional-intensive and technology-enhanced approaches have been pilot-tested, but not yet being compared for the same reviews.
We will present the results during the Summit.

Conclusions: Finding innovative ways of reducing the burden of the initial steps of systematic reviews are of relevance to the BMJ Rapid Recommendations project, but also to systematic reviewers in general.