Project 171359

Health-care workers often search the MEDLINE database for journal articles to keep up with the latest advances in medicine. MEDLINE, which is freely available online through the PubMed service (www.pubmed.gov), is the largest database in the world for medical information and currently has records on over 16 million scientific articles. However, because the database is so large, finding the right information can be frustrating and time-consuming; a search often returns articles that aren't relevant to the clinical problem, and important articles that should be found are often missed. One way to improve these results is to search using a filter that weeds-out unwanted information, leaving a higher concentration of relevant articles. This method is known as filtering the database. We have previously created three types of filters to use in MEDLINE. The first filters identify clinical research that meets quality standards. These filters are now part of the PubMed service and are widely used by the medical community. The second filters find information related to nephrology (the care of patients with kidney disorders) and the third filters limit the database to journals that publish nephrology information. We will now test these filters by applying them to real searches created by doctors to answer clinical questions related to kidney patient care. We will compare the use of the filters to the clinician's own, unaided search. If we show that these filters produce improved results, we can make them available online to health professionals around the world and educate them on their use. Using our example, filters can be created and tested for other medical disciplines such as cardiology and cancer care. We expect helping health professionals keep up-to-date with the latest advances will improve the transfer of research into practice, medical decision-making and patient care.

testing the impact of MEDLINE filters on evidence retrieval for nephrology-related clinical questions