Date created: | Last Updated:

Category: Project

Description: Background: Systematic Reviews (SRs) are a prosperous study type in modern medicine and beyond. SR authors aim at globally collecting and synthesizing the existing evidence concerning a research question. It is recommended to complement primary term-based searches in bibliographic databases by supplementary search techniques to find additional relevant articles. Among these supplementary techniques, citation-based search methods, known as citation searching (CS), are widespread. Based on known relevant input articles (seed references), current SR authors mostly use unranked direct CS (UDCS) to identify directly cited and citing literature. Ranked indirect CS (RICS) goes further and also collects co-cited and co-citing literature combined with a relevance ranking and cut off procedure to generate a manageable number of search results. However, RICS-associated benefits are unclear and workflows remain to be standardized. Our objectives are to implement a proposed workflow for RICS, make it available to the public in the form of a web application, and compare its performance to UDCS. Methods: On the basis of the existing Local Citation Network, we will develop the Ranked Co*Citation Network. This new web application will automate RICS by collecting citing, cited, co-citing, and co-cited literature of seed references. It will be equipped with a ranking algorithm and threshold, preferably matching the volume of literature retrieved by UDCS. We will run a case study in which we will use UDCS and RICS in parallel as supplementary search techniques in an ongoing SR, and then compare the effectiveness and efficacy in retrieving unknown relevant literature. Finally, we will expand this comparison to multiple cases by prospectively recruiting a consortium of authors who will agree to screen the results from both UDCS and RICS for eligibility in their SR projects. The effectiveness and efficacy measures "Number of unknown relevant records retrieved", "precision", and "number needed to read" will be calculated and synthesized across participating SR projects. Subgroup analyses according to research topic category, review type, and number of databases in the primary search will be run. Expected results: A new web application to perform RICS will be made available to the SR community. Our study will establish whether RICS is better suited to supplement primary SR searches than UDCS. It will also clarify the scope of research topics in which supplementary CS is likely to be beneficial.