Robust evidence synthesis methods are essential to collating and evaluating existing evidence rigorously and systematically. Retrieving all relevant literature is the first, fundamental step of evidence synthesis. This has become more challenging in many contemporary evidence syntheses due to 1) the volume and diversity of literature available; 2) the range and complexity of topics addressed; and 3) the ‘fuzziness’ of complex research questions. Collectively, this makes it challenging to effectively identify relevant information for broad, complex evidence syntheses, compared to conventional effectiveness reviews which draw on homogenous clinical trial reports.

This presentation will evaluate a range of citation analysis techniques as a primary search method utilizing tools to uncover multi-step links along a spectrum between exposure and outcome, the development of a novel classifier for complex study designs, and integrating these elements into an iterative process. The output will be a “search-classifier bundle” that can be used to optimize information retrieval by balancing sensitivity and specificity. This will be relevant and transferrable across a range of study designs and specialties and has the potential to facilitate the development of new types of evidence ecosystems where synthesis of existing evidence is integrated into continuous processes built around substantial questions that are of interest to the wider knowledge-user communities.

This method should enable more nuanced searching, reducing reliance on the “big” medical databases and facilitating more efficient retrieval of information pertinent to specific populations without having to search multiple databases, which should be beneficial to researchers in the Middle East.


Article metrics loading...

Loading full text...

Full text loading...

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error