Because not all articles indexed in Medline contain evidence that could be directly applied to patient care, you have to apply special filters that would retrieve only EBM documents.
To find the best research evidence available in Ovid Medline, search Medline as explained in the Medline tutorial, then click on "Limits", then "Additional Limits" and under the section "Clinical Queries" select the appropriate EBM filter.
Suppose you want to see if there is evidence about effectiveness of needle biopsy in diagnosing prostatic cancer.
Search Medline OVID as discussed in Medline tutorial; click "Limits" then "Additional Limits". One of the limiting options displayed is "Clinical Queries" that allows filtering of evidence-based citations.
Clinical Queries has options according to the following categories: therapy, diagnosis, prognosis, reviews, clinical prediction guides, qualitative studies, etiology, costs, and economics. For each one of these you have three choices: sensitivity (produces largest number of records, from which some are not relevant), specificity (produces highest level of relevancy and the least number of citations with the possibility of missing one), and optimized (results are in between the sensitivity and specificity options).
Limit the last result from search history ("prostatic neoplasms" and "needle biopsy") with the diagnosis (optimized) filter to limit retrieval to EBM literature. If not satisfied, select to limit by "diagnosis (specificity)" or "diagnosis (sensitivity)".
 |
When searching for EBM it is advisable not to Focus the MeSH terms, and to OR MeSH with its keyword, unless the number of hits retrieved is too high |
From SML homepage click on PubMed@SML. There you find a link "Clinical Queries" that has built-in search filters to retrieve EBM citations that could be applied to provide best patient care.
This Clinical Queries page is divided into 3 sections that could be searched for a particular question:
For example to see if there is evidence for treatment of otitis media with antibiotics in children, type in the box otitis media antibiotics and click "Search".
In the first category and by default, "therapy" is selected as "Category" and "Broad" is selected under "Scope".
 |
For diagnosis, etiology, prognosis, or prediction question, select the category option accordingly. |
You will see directly the number of records retrieved with the first five displayed and a See all link at the end of displayed results. If you retrieve too many results, change "Scope" option to "Narrow" to get fewer hits that are more relevant and focused.
Similarly the middle section "Systematic Reviews" displays the first five results and at the end there is a See all link. The last section is for "Medical Genetics".
Note: remember that the hits you retrieve from PubMed or Medline need critical appraisal or evaluation.
Cochrane is an international organization that started in 1993. It was developed after Archie Cochrane, a UK epidemiologist who noticed that health-care is not always based on good evidence. It prepares, maintains and promotes the accessibility of systematic reviews of the effects of health care intervention. It has more than 15 centers worldwide, with 8000 members. Each subject area in health care is covered by a Specialized Cochrane Review Group.
Cochrane reviews are high quality, up-to-date systematic reviews across a broad range of health topics, intended to help decision makers to cope with the large volume of literature by summarizing the reliable evidence of the benefits and risks of healthcare. Systematic reviews limit bias and reduce chance effects, thus providing more reliable results for drawing decisions about health care. Cochrane contains a number of high-quality evidence databases the most important being:
Cochrane Database of Systematic Reviews (CDSR): are full text documents reviewing effects of health care. These are highly structured and systematic and have the highest level of evidence.
Database of Abstracts of Reviews of Effects (DARE): provides structured abstracts of systematic reviews that have been initially appraised by reviewers giving comment on the quality of methodology. It complements CDSR by offering a selection of high-quality assessed reviews in those subjects where there is currently no Cochrane review.
Cochrane Central Register of Controlled Trials (CENTRAL): is a bibliography of controlled trials, it also includes reports published in conference proceedings and other sources not indexed in Medline or other bibliographic databases.
Cochrane Methodology Register (CDMR): contains articles on research synthesis and systematic reviews.
Health Technology Assessment Database ( HTA ): contains information on healthcare technology assessment, including details of ongoing projects and publications from health technology assessment organizations.
NHS Economic Evaluation Database (NHS EED)
Cochrane Collaboration Logo:
Cochrane Logo illustrates a systematic review of data from 7 randomized controlled trials (RCT) comparing one health care treatment with a placebo. Each horizontal line represents results of one trial. The shorter the line means the more certain the result. The diamond is the combined result, if it is to the left then treatment is beneficial. If a horizontal line touches the vertical one then the trial found no clear difference between the treatments. If the horizontal line is to the right or the diamond is to the right of the vertical line then the treatment in not good, and may do more harm than good.
A Cochrane systematic review consists of:
The first stage in preparing a review is the development of a protocol which includes an introduction, objectives, inclusion criteria, and methods (including search, strategy). It takes about 12 -18 months for a protocol to become a full review. If it remains as a protocol for 2 years, it is removed from the database. Updating reviews is very essential, this may be done in light of comments from others or as new evidence shows up.
Cochrane Search Options
Below are some Cochrane search options:
◦Default is Boolean AND; use OR and parenthesis ( ).
◦May 2000 abstracts were added to Medline/PubMed.
◦Many records in Cochrane are not indexed using MeSH. So search using MeSH “OR”ed with textwords.
◦Protocols are not yet indexed by MeSH, so use textwords. Two thirds of CENTRAL records are indexed by MeSH, and some of the DARE database.
.Truncation(*), could be internal, external or at beginning of word.
◦NEXT ( is adjacent ); NEAR ( is maximum 6 words)
◦Allows "Phrase Search"
