Identifying studies -information sources

Task: Describe all methods of identifying published and unpublished studies including, as applicable: which bibliographic databases were searched with dates of coverage; details of any hand searching including of conference proceedings; use of study registers and agency or company databases; contact with the original research team and experts in the field; open adverts and surveys. Give the date of last search or elicitation.

Text: In describing the extensive methods used to identify studies for the EBCTCG systematic reviews, it is important to consider its development over the last 30 and more years.* Briefly, when the overview began with the establishment of the EBCTCG in the early 1980s the focus was on the effects of chemotherapy and hormonal therapy (in particular, tamoxifen) on recurrence and death. Relevant randomised trials were sought through literature searches, contact with research groups around the world who might have done relevant studies and pharmaceutical companies. In the late 1980s, as part of the second cycle of the overview, the scope was expanded to cover all treatments for women with early breast cancer. The searching was also expanded to include a wider range of databases, including registers of trial protocols, conference proceedings, and specific efforts to ask researchers to provide information on trials that they had conducted or knew about. Over the subsequent decades, this approach to searching has continued, with regular searches of bibliographic databases including MEDLINE, Embase and the Cochrane Library and the checking of abstracts presented at the San Antonio Breast Cancer Congress, and conferences of the American Society of Clinical Oncology, European Cancer Organisation and European Society of Medical Oncology. This has led to the compilation of a database of more than 39,500 articles of relevance to the EBCTCG overview (as of November 2022), which continues to be populated on a regular basis, with additional targeted updates for any meta-analyses prior to submission for publication. As such, the date of last search or elicitation for each systematic review is likely to be close to the data on which the relevant analyses were finalized.

* Darby S, Davies C, McGale P. The Early Breast Cancer Trialists’ Collaborative Group: a brief history of results to date. In Davison AC, Dodge Y, Wermuth N (editors). Celebrating statistics. Oxford University Press, Oxford, 2005 pp.185-198.