Microsoft word - research project.doc

Karolina Kucerova Charles University in Prague Evidence-Based Medicine Course The University of North Carolina at Chapel Hill Spring 2004 Introduction
This work intends to concentrate on comparison of two sources of medical information included in the Medline database, PubMed and OVID. Its first part should consist of some “academic information” about these resources, their history and contents. The other part would give some examples of searches in both of these sources, compare their strategy and outcomes. It should be user-oriented, given the point of view of a person that is not meeting these resources in everyday practice. About Medline
Medline (Medical Literature, Analysis and Retrieval System Online) is the largest bibliographic database of journal articles in the field of medicine, nursing, dentistry, veterinary medicine, the health care system, and some other related sciences. It contains over 12 million records in retrospective from year 1966. Over 4600 journals in 30 languages are treated. Vast majority of cited articles is in English. Medline’s provider is U.S. National Library of Medicine, who also makes the database available via internet free of charge (some database centers – e.g. Dialog, DataStar, STN International, OVID – offer Medline as well, charged of course). NLM also allows searching in documents older than 1966 (NLM Gateway), offers document delivery service for articles cited in Medline, and publishes a consumer-oriented MedlinePlus service which is also free. NLM uses and handles the widespread classification system Medical Subject About PubMed
PubMed is a special service of U.S. National Library of Medicine and National Center for Biotechnology Information (NCBI). It is part of an Entrez system that offers also databases of Nucleotide and Protein Sequences, Protein Structures, Complete PubMed offers not only Medline records but also links to full-text articles where available; number of journals offering this service is growing. PubMed contains also many additional services: you may use Journal database, refine search terms in MeSH, search for particular article, use clinical queries – filters for more efficient searching, and personalize PubMed and save searches and Medline via PubMed was launched in 1997. About OVID
OVID Technologies is a database center, founded in 1988 in New York. In 1994 it bought BRS Online, a large database center oriented on medicine and other life sciences, and until now OVID belongs to leaders in medical information services. In 2001 OVID acquired another large database provider, the SilverPlatter. OVID contains not only different bibliographic databases but also e-books, full-text journals and several SilverPlatter databases (e.g. Health & Safety Publishing). Medical specialist could well use the Clinical Decision Support section of OVID content. Its Tools are aimed at helping clinicians in their daily work and Contents involve resources including EBM reviews. The use of PubMed and OVID
Even before performing any search I can state some differences between the PubMed and OVID. They either result from the “theory,” or they can be experienced during the use of tutorials (PubMed and OVID). 1. PubMed is free, and so it is more accessible for all possible users. 2. Concerning the interface, most OVID features are concentrated on one page, they can be displayed at the same time, or they are referred in a simple way; PubMed offers more utilities on new pages, many of whom are automatically combined with the main search (but some are not). 3. Because of this I appreciate the possibility of saving the PubMed search as an URL; if I got lost in the search, I can find the most relevant page in my 4. On the other hand, PubMed integrates more features for searching. These include not only MeSH search etc., but also clinical queries filters or availability of other Entrez resources (protein, genome information etc.) 5. OVID integrates another groups of resources – other databases, and also special resources for EBM (Cochrane databases etc.) and clinical decisions 6. PubMed uses automatic explosion and automatic mapping which both can help a lot in formulating strategy: man does not have to be so careful in Of course there could be many more qualities found but I consider these the most important ones that could be noticed without profound knowledge of the systems. Other characteristics would appear in the searches. Search No. 1
In the first search I am solving a problem of a 50-year-old woman coming to her clinician with urinary incontinence. She wants to know whether she could benefit from advertised medicines tolterodine or oxybutynin. PICO question: In middle-aged women with urinary incontinence, are tolterodine or oxybutynin an effective treatment? This question obviously focuses on therapy, and so we will be looking for randomized controlled study. The first thing I do is look for terms in MeSH. The drugs are not MeSH terms, urinary incontinence is. Putting these terms in PubMed Search ((tolterodine OR oxybutynin) AND urinary incontinence), I receive the expanded query ((("tolterodine"[Substance Name] OR tolterodine[Text Word]) OR oxybutynine[All Fields]) AND ("urinary incontinence"[MeSH Terms] OR urinary incontinence[Text Word])) and 84 results. Filtering it for middle-aged women, humans, RCT’s and English language, I get 16 results. Another possibility is to search the three main terms in Clinical queries. It gives a very long query with 76 results. Imposing the same filter, 16 results appear again. Looking at the results, it is obvious that many of them are very relevant, often head- to-head studies. But maybe even more useful would be searching for systematic The main query ((tolterodine OR oxybutynin) AND urinary incontinence) gives me 5 citations, three or four of them seem to be highly relevant. Their titles: Is tolterodine (Detrol) or oxybutynin (Ditropan) the best for treatment of urge urinary incontinence?; Anticholinergic drugs versus placebo for overactive bladder syndrome in adults (this one is from Cochrane Database of Systematic Reviews); Tolterodine versus oxybutynin in the treatment of urge urinary incontinence: a meta-analysis. and Tolterodine use for symptoms of overactive Entering OVID, in this and the other search I choose only the Medline 1966-2004 database. Although I could earn more from the EBM publications, this project In OVID I have to follow many steps. I search urinary incontinence as a MeSH term (exploded), combine it with textwords tolterodine and oxybutynin. I have 242 results at the moment. Limiting the search to systematic reviews and humans, I retrieve 10 citations. The ones retrieved via PubMed are among them. Trying to get closer to my patient, I limit the search on English and middle-aged females: 2 results remain. Their titles are Is tolterodine (Detrol) or oxybutynin (Ditropan) the best for treatment of urge urinary incontinence? and Behavioral vs. drug treatment for urge urinary incontinence in older women: a randomized controlled trial. We can see that even this refined search brought one citation PubMed did not notice. Search No. 2
In this case a woman comes and wants to know whether her newborn child has higher risk of allergy when sharing a flat with a pet cat. PICO question: In infants, does intense cat exposure augment the risk of developing allergies? This is a question of etiology/harm, I will be looking especially for cohort In PubMed, I search for the terms in MeSH. I find the term Cats (not cat) and hypersensitivity (not allergy). Knowing this, with the use of etiology/harm strategy I formulate the query (cat OR cats) AND (allergy OR hypersensitivity) AND ("Case-control studies"[MeSH] OR cohort), it gives me 156 relevant citations. Filtered for humans, infants and English, 57 of them are left. The same query in clinical queries/harm gives 116 citations, 51 of them stay after the limit is imposed. Adding the word review to the query, 7 citations are left. Their titles appear to be relevant: Early exposure to allergen: is this the cat's meow, or are we barking up the wrong tree?; Early pet exposure: friend or foe?; The development of childhood asthma: lessons from the German Multicentre Allergy Study (MAS); The atopic child and the environment. If they were not relevant, the larger set of citations can be sought. Using the experience from PubMed I search terms cat and cats, hypersensitivity and allergy, in Mesh and text, exploded where possible. Combining them results in 1902 citations. Adding the limit for case-control studies (MeSH term) and text word cohort, 135 matches remain. Of them 52 are in English, relate to humans under 2 years of age. These appear to be the ones found by PubMed. Limiting to systematic reviews is not useful as only 2 results remain. Trying to repeat the PubMed strategy, I search for the term review. OVID automatically searches for publication type and answers 0. Search for review as text word in combination with the 52 documents gives only one result. The only useful limit could possibly be the publication year – 2000 through 2004 there are 34 articles matched. Concluding Remarks
Apart from the differences mentioned above, both the Medline providers offer Searching via OVID might be more synoptic as every step is clearly visible in the table. You can – or you even must – clearly define very step of your search. It means, for example, that especially searching synonyms or similar terms is exceedingly time-consuming; while paying for connect time this could be extremely unpleasant. And what really annoys me? It is the interaction with the web browser: the “Forward” and “Back” buttons do not work here, and I as I am using them automatically all the time I often have to perform the search more than once. PubMed can very easily make you lost. Either you forget to send the term from MeSH to the search field, or you forget to delete the last search’s limits, or you search for clinical query instead of systematic review… But otherwise the PubMed tries to do as much work as possible for you, which is a great help especially for a non-professional bibliographic researcher. Useful links
National Center for Biotechnology Information

Source: http://kajinka.info/UISK/EBM.pdf