I make no apology that this chapter is the longest in the book.You can apply all the rules for reading a paper correctly but if you’re reading the wrong paper you might as well be doing something else entirely. There are already over 15 million medical articles on our library shelves.Every month, around 5000 medical journals are published worldwide and the number of different journals which now exist solely to summarise the articles in the remainder probably exceeds 250. Only 10–15% of the material which appears in print today will subsequently prove to be of lasting scientific value. A number of research studies have shown that most clinicians are unaware of the extent of the clinical literature and of how to go about accessing it.1, 2 Dr David Jewell, writing in the excellent book Critical reading for primary care,3 reminds us that there are three levels of reading. 1. Browsing, in which we flick through books and journals looking for anything that might interest us. 2. Reading for information, in which we approach the literature looking for answers to a specific question, usually related to a problem we have met in real life. 3. Reading for research, in which we seek to gain a comprehensive view of the existing state of knowledge, ignorance, and uncertainty in a defined area.
In practice, most of us get most of our information (and, let’s face it, a good deal of pleasure) from browsing.To overapply the rules for
critical appraisal which follow in the rest of this book would be to kill the enjoyment of casual reading. Jewell warns us, however, to steer a path between the bland gullibility of believing everything and the strenuous intellectualism of formal critical appraisal.
The Medline database:
If you are browsing (reading for the fun of it), you can read what you like, in whatever order you wish. If reading for information (focused searching) or research (systematic review), you will waste time and miss many valuable articles if you simply search at random. Many (but not all – see section 2.10) medical articles are indexed in the huge Medline database, access to which is almost universal in medical and science libraries in developed countries. Note that if you are looking for a systematic quality checked summary of all the evidence on a particular topic you should probably start with the Cochrane database rather than Medline, which uses very similar search principles. However, if you are relatively unfamiliar with both, Medline is probably easier to learn on. Medline is compiled by the National Library of Medicine of the USA and indexes over 4000 journals published in over 70 countries. Three versions of the information in Medline are available.
• Printed (the Index Medicus, a manual index updated every year, from which the electronic version is compiled).
• On-line (the whole database from 1966 to date on a mainframe computer, accessed over the Internet or other electronic server).
• CD-ROM (the whole database on between 10 and 18 CDs, depending on who makes it).
The Medline database is exactly the same, whichever company is selling it, but the commands you need to type in to access it differ according to the CD-ROM software.
Commercial vendors of Medline on-line and/or on CD-ROM include Ovid Technologies (OVID),
Silver Platter Information Ltd (WinSPIRS),
Aries Systems Inc (Knowledge Finder),
The best way to learn to use Medline is to book a session with a trained librarian, informaticist or other experienced user. Unless you are a technophobe, you can pick up the basics in less than an hour. Remember that articles can be traced in two ways.
1. By any word listed on the database including words in the title, abstract, authors’ names, and the institution where the research was done (note: the abstract is a short summary of what the article is all about, which you will find on the database as well as at the beginning of the printed article).
2. By a restricted thesaurus of medical titles, known as medical subject heading (MeSH) terms.
To illustrate how Medline works, I have worked through some common problems in searching. The following scenarios have been drawn up using OVID software4 (because that’s what I personally use most often and because it is the version used by the dial up service of the BMA library,to which all BMA members with a modem have free access). I have included notes on WinSPIRS5 (which many universities use as a preferred system) and PubMed (which is available free on the Internet, comes with ready made search filters which you can insert at the touch of a button, and throws in a search of PreMedline,the database of about to be published and just recently published articles6). All these systems (Ovid, WinSPIRS and PubMed) are designed to be used with Boolean logic, i.e. putting in particular words (such as “hypertension”,“therapy”and so on) linked by operators (such as “and”, “or” and “not”, as illustrated on pp 19 and 20). Knowledge Finder7 is a different Medline software which is marketed as a “fuzzy logic” system; in other words, it is designed to cope with complete questions such as “What is the best therapy for hypertension?” and is said to be more suited to the naïve user (i.e. someone with little or no training). I have certainly found Knowledge Finder’s fuzzy logic approach quick and effective and would recommend it as an investment for your organisation if you expect a lot of untrained people to be doing their own searching.The practical exercises included in this chapter are all equally possible with all types of Medline software.