| Site Index |
|
|
         |
|
|
|
2004 Vohs and Lawrence Awards; A Focus on Evidence-Based Medicine |
|
|
|
|
|
|
Acquiring
Evidence-- Effective, evidence-based literature searching is easier if it's kept simple. The basic steps are: A Clinical Question Arises Out of Care of the Patient A 60-year-old, postmenopausal woman presents with a diagnosis of arthritis. Can glucosamine be recommended in place of more conventional therapy? Formulate the Question Well-built clinical questions are constructed using PICO format. The key is to start with the most basic question. More facets can always be added if necessary. In this example: P (Patient/Disease)--Arthritis Are the age, sex and menopausal status of the patient of primary importance? Probably not, so leave it out for now. I (Intervention)--Glucosamine C (Comparison)--Conventional Therapy Many clinical questions such as this one don't have a searchable comparison. It is okay to ignore it. O (Outcome)--Effectiveness of therapy, side effects, etc If one particular outcome was of interest, it could be built in the search. In this case, outcome definition is broad enough to leave it out. Select the Appropriate Resource There are many resources available. Don't automatically assume that an exhaustive MEDLINE search is needed. Sometimes a book is the best place to look. For example, in searching for the use of a particular drug in pregnancy, it is perfectly reasonable to use Drugs in Pregnancy and Lactation by Gerald Briggs.1 But most of the time, you will be looking for a systematic review or randomized controlled trial. Cochrane databases (www.cochrane.org) or Clinical Evidence (www.clinicalevidence.com) make a good first pick. These types of resources do the evidence analysis for you. ACP Journal Club (www.acpjc.org) or the Database of Abstracts of Reviews of Effectiveness (DARE) (www.york.ac.uk/inst/crd/darehp.htm) produces critical reviews of systematic reviews and other key studies. If these resources do not answer the question, a literature search will need to be performed.
Develop a Search Strategy The search strategy will depend on what type of resource is used. But in all cases, try to enter each of the PICO elements as separate concepts. It will be easier to manipulate the search if necessary. All the evidence-based
medicine (EBM) review databases (eg, Cochrane) are not indexed. There
are no MESH headings or controlled keywords. All synonyms must be included.
The steps of a typical search might be (Figure 1):
2. "Glucosamine"
3. "1 and 2" Adding facets, such as age or sex, can be tricky. Don't try to do this unless absolutely necessary. It is easier to browse through 30 citations rather than try to limit the search somehow and potentially miss good references. MEDLINE is an indexed database. That means that every citation is assigned headings. The indexer does the work so synonyms become less important. The steps of a MEDLINE search might be (Figure 2): 1. "Exp Arthritis"
2. "Glucosamine" 3. "1 and 2" Sex and age are much easier to search in MEDLINE. Both OVID and PubMed offer not only sex and age limits but also human/animal and subject subsets, such as AIDS. Evaluate Results and Try the Search Again if Necessary Did you get the information you need? If not, try different search terms. Was your retrieval too large? Think about adding other facets of the clinical question, such as age or sex. If the results seemed too small, think about taking search terms out. Many times, there really isn't much information on a topic. Which is Better, OVID or PubMed? Using OVID or PubMed is largely a matter of personal preference, but there are some key differences: OVID
PubMed
These basic tools make searching for the evidence much less complicated than it sounds. The last ingredient is practice. Happy searching! Reference
.
|
|
|
|
|
|
|
|
|
|
|
 |
|
