Article type
Year
Abstract
Background: Reviewers could benefit from more efficient study retrieval tools than currently exist. Having an online, ‘1-stop’ federated search facility providing empirically derived and validated search filters and filtering aids to retrieve and collate all pertinent studies would help.
Objectives: To develop a ‘superfilters’ website that provides review authors with the opportunity to search across several electronic databases simultaneously with empirically derived, high performance, search filters. To add ‘capture-mark-recapture’ (CMR) statistical modeling to this superfilters site to help those searching for all available evidence to determine whether to continue or stop searching, for example, when conducting systematic reviews.
Methods and Results: We designed a website that has federated search capabilities, enabling users to select from a host of search filters and search in large, bibliographic databases including PubMed, Ovid Medline, Ovid Embase, EBSCO CINAHL, Ovid PsycINFO and MacPLUS. These filters retrieve articles of highermethodological rigor from various disciplines of medicine, and do so according to the research methods used. The user can turn the search filters on or off when performing a single search across multiple bibliographic databases simultaneously. Searches can be limited by type of article (e.g., treatment, diagnosis), age of study participants (e.g., adult, geriatric), and date of publication. Searches include options for breadth: broad (highly sensitive), balanced or narrow (highly specific). The retrieved set of citations is collated with duplicate citations removed. In the process of identifying duplicate citations, CMR statistical modeling is performed and an estimate of the total theoretical size of a collection of literature is provided. This automated statistical technique can provide searchers with evidence that their searching can stop or should continue.
Conclusions: A superfilters website has been developed that can aid researchers when conducting comprehensive and targeted searches of the medical literature.
Objectives: To develop a ‘superfilters’ website that provides review authors with the opportunity to search across several electronic databases simultaneously with empirically derived, high performance, search filters. To add ‘capture-mark-recapture’ (CMR) statistical modeling to this superfilters site to help those searching for all available evidence to determine whether to continue or stop searching, for example, when conducting systematic reviews.
Methods and Results: We designed a website that has federated search capabilities, enabling users to select from a host of search filters and search in large, bibliographic databases including PubMed, Ovid Medline, Ovid Embase, EBSCO CINAHL, Ovid PsycINFO and MacPLUS. These filters retrieve articles of highermethodological rigor from various disciplines of medicine, and do so according to the research methods used. The user can turn the search filters on or off when performing a single search across multiple bibliographic databases simultaneously. Searches can be limited by type of article (e.g., treatment, diagnosis), age of study participants (e.g., adult, geriatric), and date of publication. Searches include options for breadth: broad (highly sensitive), balanced or narrow (highly specific). The retrieved set of citations is collated with duplicate citations removed. In the process of identifying duplicate citations, CMR statistical modeling is performed and an estimate of the total theoretical size of a collection of literature is provided. This automated statistical technique can provide searchers with evidence that their searching can stop or should continue.
Conclusions: A superfilters website has been developed that can aid researchers when conducting comprehensive and targeted searches of the medical literature.