What databases to search for your systematic review?
[some personal COMMENTS from your librarian]
- PUBMED/MEDLINE (NCBI): probably the most popular database for information retrieval and systematic reviews in biomedical and health sciences.
Apart from its sheer size - 25 million records from 1950-present, covering some 5000 journals - PubMed/Medline features a powerful hierarchical thesaurus module named Medical Subject Headings (MeSH).
PubMed contains all Medline records, but remember that not all PubMed records have already been fully processed as Medline records. E.g. recently added PubMed records may only receive adequate MeSH terms after several months, so they won't show up in a controled vocabulary search. Don't expect too much non-journal material (e.g. books and chapters) either, although this has been improving for the last few years.
- EMBASE (Elsevier): PubMed's major biomedical and health sciences competitor, comparable in size, but partly different in content.
Embase should also be part of any serious literature review. Unfortunately it is currently too expensive for ITM to subscribe to.
- ISI Web of Knowledge (Thomson Reuters): again, an essential tool for serious literature reviewing, although keywords are added far less systematically than in PubMed. The number of journal indexed is far larger (over 10.000), but comprises all scientific disciplines, not limited to biomedical and health sciences. But because of the ample citation linking in both directions, for information retrieval ISI WoK has the added advantage of forward linking (i.e. find out which more recent articles have cited a particular interesting item).
- SCOPUS (Elsevier) could somehow be considered to be for ISI WoK what Embase is for PubMed: a very serious competitor (actually with even more source journals), but alas too expensive for ITM to subscribe to.
- COCHRANE LIBRARY (Wiley) contains over 9,000 reviews and protocols on medical topics. As it aims to publish one comprehensive review on each specific topic it is more an evidence-based medical knowledge database than a bibliographic one (like the four above). Cochrane Library is a popular resource used in many systematic reviews, especially for clinical topics. Then again, one can wonder whether performing a new systematic review on a specific topic makes much sense if a recent Cochrane review is already available. Whatever, next to reviews and protocols the Cochrane Library also contains a substantial database on methodological literature relating to systematic reviews.
- GOOGLE SCHOLAR: contains an awful lot of data, but the exact universum you're searching in is not defined rigorously as with the previous databases. Yet useful to find additional (e.g. non-journal) literature, especially in social science context like e.g. public health.
- GLOBAL HEALTH (CABI): Although less than 10% the size of PubMed, Global Health is a fine complement relating to infectious diseases in developing countries and fully incorporated the Tropical Diseases Bulletin. Expect to find a lot of useful additional material - especially non-journal.
- VETERINARY SCIENCE (CABI): Like Global Health, a useful complement when animal health is your concern, but no longer subscribed to by ITM.
- BDSP - Banque de Données en Santé Publique (EHESP): seems like a valuable resource for French language literature on public health.
- SSRN - Social Science Research Network: seems like a valuable resource for e.g. public health.
- DOAJ – Directory of Open Access Journals: obviously limited to open access journals, but tends to be used regularly in systematic reviews.
- EDS Discovery Service from ITM library:
EDS is fine for optimizing bibliographic information retrieval, but not really suited for systematic reviews in a standardized publications universe, as every institute makes its own selection of databases and full-text collections to include in their EDS. The EDS of e.g. Antwerp University will yield very different results from the ITM library EDS. As each EDS is built upon a number of databases and collections, duplicate occurrences of the same items are unavoidable - the same is true when for your literature review you combine several of the databases listed above. And as each EDS features a balance between discoverability of bibliographic data and availability of online or print full-text, finding useful bibliographic references does not guarantee actual availability of the publication.
- Finally, there are the ITM library databases (http://lib.itg.be/literature/databases.php) to consider, such as ITG Staff Publications (TropMed Central Antwerp and Pure contain no additional content) , ITG Student Dissertations, Health Care in Developing Countries, Tropical Endemic Diseases Control – and, why not, the ITM Library Books Catalog (http://lib.itg.be/literature/catalogs.php ). Generally, you'll find extra useful publications on your topic, but can you rhyme this with the principles of systematic literature reviews?
- Publications that are not immediately available online or in the library can be ordered from the ITM Document Delivery Service => http://lib.itg.be/services/interlibrary-services.php.
Keywords or free-text?
- Keyword systems tend to be specific to each database. The MeSH thesausus is an admirable example of some 30,000 controlled terms placed in a hierarchical context and applied to bibliographic records by professional medical indexers. But most other databases have their own thesaurus systems, or no controlled keywords system at all. Contents searching may be based on computer generated keywords, title words, abstract words or even full-text indexing - each option having its own implications for recall (as much items as possible) and precision (only the relevant items). Generally some trial and error is implied in finding the ideal combination between controlled and free text searching in each of the databases selected. Always consider synonyms for your prefered keywords. Fashionable vocabularies tend to change over time.
- Boolean operators: not all databases automatically combine your search terms in an implicit 'OR' (any of these terms) or 'AND' (all of these terms) relation. Some databases (e.g. our own WebSpirs databases) expect explicit 'AND', 'OR', 'NOT' operators (or more specific 'NEAR', 'WITH', etc.). When such Boolean operators are called for it is safest to apply brackets '(...)' for a correct interpretation, as there may be system setting, e.g. systematically prioritizing 'AND' above 'OR'. Whatever, do beware!
- See also: http://lib.itg.be/literature/information-retrieval.php for more information retrieval issues.
- Egger M, Smith GD, Altman DG. Systematic reviews in health care; meta-analysis in context; 2nd ed. London, BMJ Books, 2001. 487 pp. [ITG Library: QA/16981]
- Glasziou P, Irwig L, Bain C, Colditz G. Systematic reviews in health care; a practical guide. Cambridge, Cambridge University Press, 2001. 137 pp. [ITG Library: QA/17335]
- Greenhalgh T. How to read a paper; the basics of evidence-based medicine; 4th ed. New York, Wiley-Blackwell, 2010. 238 pp. [ITG Library: BA/21649]
- Khan K, Kunz R, Kleijnen J, Antes G. Systematic reviews to support evidence-based medicine; 2nd ed. London, Hodder Arnold, 2011. 201 pp. [ITG Library: QA/22171]
- Liamputtong P. Research methods in health; foundations for evidence based practice; 2nd ed. South Melbourne, Oxford University Press, 2013. 521 pp. [ITG Library: ZVP/23012]
- Nelson HD. Systematic reviews to answer health care questions. Philadelphia, PA, Wolters Kluwer / Lippincott Williams & Wilkins, 2014. 230 pp. [ITG Library: QA/23132]
- Oliver P. Succeeding with your literature review: a handbook for students. Maidenhead, Open University Press, 2012. 152 pp. [ITG Library: BA/23326]
- Shoyania KG, Sampson M, Ansari MT, Ji J, Garritty C, Rader T, Moher D. Updating systematic reviews. Rockville, MD, Agency for Healthcare Research and Quality (AHRQ), 2007. (AHRQ Technical Reviews and Summaries; 16). [ITG Library: QA/Online]