In today's video, I'm going to show you how to use the filters in PubMed, do a search using the MeSH database, as well as the advanced search builder. As we have seen, a basic all field search often returns large numbers of papers, many of which are irrelevant. The MeSH database is a controlled vocabulary of terms that are used to index the papers for Medline. By searching using the MeSH terms, you will retrieve only those papers that have been indexed with those terms. Thereby creating a smaller set of results that contains more relevant papers. You will find the link to the MeSH database on PubMed's homepage. I've entered the term aspirin in the MeSH database. Notice that a definition of the subject heading is given to ensure that you choose the most appropriate, and I added it to the search folder instructing PubMed to search for articles that were indexed with this subject heading. As you scroll down, you will see entry terms, that had you searched on them they would have linked to the MeSH heading aspirin. Notice also the MeSH hierarchy of categories showing the position of aspirin. Again, I will search for heart attack in the MeSH database, and it links me to myocardial infarction. I have the option of narrowing my search to a subheading, which I do in selecting prevention and control. I add my choice of myocardial infarction, prevention and control to the search builder, and select the boolean operator. And in this way, I'm now instructing PubMed to search for articles that were indexed by both the MeSH headings aspirin and myocardial infarction, prevention and control. The list of results from this search is now displayed, and I can now decide to filter to a particular publication or article type by clicking on Customize and selecting Systematic Review. This is now option for filtering which I do. I then filter the search to human, last 10 years, and if necessary to full text. Please not that there is a time delay between the papers being input into PubMed and their being indexed by Medline. So, by searching on MeSH terms alone, you could miss some very recent material, as well as material from PubMed Central that has not been included in Medline. As you have seen, by using filters you can narrow a search to a publication or article type. However, by doing this you run the risk of missing some recent additions to the database. To ensure that very recent material has not been missed, it is sometimes necessary to rerun your search doing an all fields search using keywords. I enter my search terms using free language as done previously, aspirin heart attack prevention, and if I want to locate a particular study type, I can enter that as well. So I add systematic review. I then customize my date range to 2015, or even just the last few months. A basic all fields search, which we saw in the previous video, gave a high recall of articles but included match irrelevant material. The MeSH search returned a smaller but more relevant set of results. Now we will look at the Advanced Search Builder where we can choose the fields to be searched. In the advanced screen, it is possible to control exactly which fields must be searched. For example, we can narrow a search by searching only the title abstract field, or an author's name if we want to locate articles by a particular author. In the advanced screen, it is now possible to see a history of my searches. And at this stage, I can decide to save a search. In order to do this you need to register for an account with My NCBI, sign in, and then select the search set that you want to save. You're also given the option to set up an alert, so that should anything new be added to PubMed that matches your search strategy, you will receive an email to alert you to new material. For a literature search to be comprehensive, it is customary to search other databases as well. Some important subscription based databases include the Cochrane Library for randomized control trials and systematic reviews of trials. Embase, the European medical database, which indexes many journals from developing countries and compliments your search in Medline. Citation indexes. These are important as one can follow a particular paper by viewing all the references used for that paper, as well as other authors who may be citing the paper in hand. Citation indexes include Web of Science and Scopus, which in turn include citations from Medline and Embase. Useful if your institution does not subscribe to Embase. Then there are indigenous databases, such as Africawide, African Index Medicus, and Silo, the database of open source journals from Latin America Spain, Portugal, and South Africa. Lastly, if your topic covers other disciplines, you might want to consider a subject-specific database, like PsycINFO, the psychology database, or CINAHL, the nursing and allied-health database.
