Margaret Sampson1, Jessie McGowan2. 1. Children's Hospital of Eastern Ontario Research Institute, Ottawa, Ontario, Canada. msampson@cheo.on.ca. 2. Department of Medicine and Family Medicine, University of Ottawa, Ottawa, Ontario, Canada.
Abstract
UNLABELLED: We offer a new method of validation for the effectiveness of MEDLINE searches used in systematic reviews, the Inquisitio Validus Index Medicus. Validation is essential to ensure that relevant studies are not missed by the MEDLINE search strategy. METHODS: To demonstrate the validation method, a sample of six updated Cochrane reviews with comprehensive searches was used. The MEDLINE searches of both the original and updated reviews were tested to determine the percent of eligible MEDLINE-indexed studies retrieved by the search (recall). RESULTS: The validation method was robust and was able to demonstrate that the retrieval of relevant studies from MEDLINE was sub-optimal. The approach to revising searches in our sample appeared unsystematic. Some poorly performing searches were used unchanged in the updates, and of the two amended strategies, one performed worse than the original when tested against studies included in the original, while the other improved recall. CONCLUSION: There is a clear need for search validation. Using this validation method can determine whether the search of the main database performs adequately or needs to be revised to improve recall, allowing the searcher an opportunity to improve their search strategy. Validation of the search is recommended for systematic reviews, where the intent is to identify all relevant studies.
UNLABELLED: We offer a new method of validation for the effectiveness of MEDLINE searches used in systematic reviews, the Inquisitio Validus Index Medicus. Validation is essential to ensure that relevant studies are not missed by the MEDLINE search strategy. METHODS: To demonstrate the validation method, a sample of six updated Cochrane reviews with comprehensive searches was used. The MEDLINE searches of both the original and updated reviews were tested to determine the percent of eligible MEDLINE-indexed studies retrieved by the search (recall). RESULTS: The validation method was robust and was able to demonstrate that the retrieval of relevant studies from MEDLINE was sub-optimal. The approach to revising searches in our sample appeared unsystematic. Some poorly performing searches were used unchanged in the updates, and of the two amended strategies, one performed worse than the original when tested against studies included in the original, while the other improved recall. CONCLUSION: There is a clear need for search validation. Using this validation method can determine whether the search of the main database performs adequately or needs to be revised to improve recall, allowing the searcher an opportunity to improve their search strategy. Validation of the search is recommended for systematic reviews, where the intent is to identify all relevant studies.
Authors: Käthe Goossen; Solveig Tenckhoff; Pascal Probst; Kathrin Grummich; André L Mihaljevic; Markus W Büchler; Markus K Diener Journal: Langenbecks Arch Surg Date: 2017-12-05 Impact factor: 3.445
Authors: Vanessa Silva E Silva; Laura Hornby; Joan Almost; Ken Lotherington; Amber Appleby; Amina Regina Silva; Andrea Rochon; Sonny Dhanani Journal: BMJ Open Date: 2020-12-15 Impact factor: 2.692
Authors: Anab Rebecca Lehr; Soha Rached-d'Astous; Nick Barrowman; Anne Tsampalieros; Melissa Parker; Lauralyn McIntyre; Margaret Sampson; Kusum Menon Journal: Pediatr Crit Care Med Date: 2022-03-01 Impact factor: 3.971
Authors: Charles Hui; Jessica Dunn; Rachael Morton; Lukas P Staub; Anh Tran; Sally Hargreaves; Christina Greenaway; Beverly Ann Biggs; Robin Christensen; Kevin Pottie Journal: Int J Environ Res Public Health Date: 2018-09-20 Impact factor: 3.390