 |
 |
Hi Patricia, During the last Cochrane Colloquium in Auckland, I attended a very presentation that showed depending on the algorithm used you will get different results from the same database. They were presenting results of searching the basic vs. advanced search panels in Clinicaltrials.gov and it was surprising how the results differed. Now taking that a step further, obviously the algorithms in place for PubMed, Central, and Embase are different (even though all three contain Medline with the exception of Central which only includes specific citations from Medline rather than the full Medline database). In a perfect world we could search Central only, but in the real-world, you can’t cut those corners (even though I see people trying to do this every day). Hopefully, as reporting of trial reports improves, so will our accuracy in returning the needed citations with more precision. Until then, at least for systematic review, we need to spread our nets as wide as we can to make sure we capture as much as we can. Ahmed From: Evidence based health (EBH) [mailto:[log in to unmask]] On Behalf Of Patricia Anderson Sent: Tuesday, December 04, 2012 9:43 AM To: [log in to unmask] Subject: Re: RCTs Hi, Susan, You are correct that ORIGINALLY the Cochrane CENTRAL database was comprised of information identified by the systematic review researchers in the Cochrane Collaboration. A few years ago they enriched the database with content scraped from MEDLINE and EMBASE through automatic searches based on filters such as the example you share. Thus, while CENTRAL still contains all the content identified by the human experts, it no longer contains only that information. I haven't encountered a situation in which anyone has found a citation that doesn't belong, but it is possible. We had a huge debate for one Cochrane study I was working on where another librarian felt it was inappropriate to search either MEDLINE or EMBASE, since we were already searching CENTRAL and CENTRAL presumably contains the content scraped from the other databases. That was what brought this to my attention. We did end up agreeing to search all three databases, but for me the whole conversation highlighted the subtleties of the overlaps and uniqueness of the different databases. <http://www.cochrane.org/handbook/6332-searching-medline-and-embase-specific -issues> "Despite the fact that both MEDLINE and <http://www.cochrane.org/glossary/5#term220> EMBASE have been searched systematically for reports of trials and that these reports of trials have been included in CENTRAL, as described in Sections <http://www.mrc-bsu.cam.ac.uk/cochrane/handbook/chapter_6/6_3_2_1_what_is_in _the_cochrane_central_register_of_controlled.htm> 6.3.2.1 and <http://www.mrc-bsu.cam.ac.uk/cochrane/handbook/chapter_6/6_3_2_2_what_is_in _the_cochrane_central_register_of_controlled.htm> 6.3.2.2, supplementary searches of both MEDLINE and EMBASE are recommended. Any such searches, however, should be undertaken in the knowledge of what searching has already been conducted to avoid duplication of effort." - Patricia On Mon, Dec 3, 2012 at 11:48 AM, Susan Fowler <[log in to unmask]> wrote: Hi: You will find more RCTs and controlled clinical trials in PubMed using the following filter: ------------------- Groups[tiab] OR trial[tiab] OR randomly[tiab] OR randomized[tiab] OR "Controlled Clinical Trial" [Publication Type] OR "Randomized Controlled Trial" [Publication Type] ------------------ Using just "Controlled Clinical Trial" [Publication Type] OR "Randomized Controlled Trial" [Publication Type] has the benefit of only pulling those citations that have been indexed specifically as RCTs and Controlled Clinical Trials but has the drawback of missing a lot of RCTs and Controlled Clinical Trials that, for one reason or another, were not indexed as those publications types. Mendeley and Connotea are not databases. They are personal citation management software programs like EndNote and Zotero, used to organize citations found in databases like Medline. CENTRAL stands for "Cochrane Central Register of Controlled Trials (Central; Clinical Trials). It is a database (bibliography) of controlled trials identified by contributors of the Cochrane Collaboration and others, as a part of an international effort to hand search the world´s journals and create an unbiased source of data for systematic reviews. Central includes reports published in conference proceedings and in many other sources not currently listed in MEDLINE or other bibliographic databases." CENTRAL is searchable via many interfaces including Wiley, Ebsco, Ovid, etc... I think CENTRAL is available to search for free although I cannot verify that. The content, however, mostly wont be free. When thinking of indexes/databases/catalogs and the concept of "free" content it may help to think back to the print indexes you accessed in your public library (if you are old enough to have had them). It is free to walk into the library and search the print indexes because those were provided to your library for free (or almost for free) but your library has to pay for the journal articles and books referenced by those indexes/bibliographies/catalogs. When you search a database now, like PubMed, PubMed is like the piece of furniture the cards or indexes were filed in, Medline is like the collection of index cards or print indexes being kept in that case. The actual journal articles and books were not also kept in that piece of furniture - they were kept on shelves somewhere else in the library. When you click a link in PubMed to access an article, you are essentially walking to the part of the library where that paid-for-content is being stored and pulling it from the shelf. There is some content that is available for free but very little compared to the amount of content actually indexed. Furthermore, let's say your library decided they wanted to get a really fancy piece of furniture and special indexing to store citations, that fancy piece of furniture and special indexing system would be analogous to Embase. It costs the library extra money to have it but stores more then usual and makes finding what is stored easier to. Embase is still just a piece of furniture and most of the actual content referenced is still stored elsewhere. Citation Management systems like Mendeley and Connotea are like your personal index cards that you create while working on a paper. Back in the day, we would write the citation we found down on an index card and then the quotes and such other information we found in the article or book to be useful. When we had to create our bibliographies, we had all the information wen needed on those cards (this is assuming we were students with excellent study habits). Then we usually kept those cards when we were done writing our papers and would maybe use them again for other projects later on. Anyway, hope this helps. Sorry if I over-explained things people already knew. It is hard to know how much people "get" when it comes to the differences among all the different sorts of information resources and management systems. -- Susan Fowler, MLIS Medical Librarian Evidence at Becker: http://beckerguides.wustl.edu/ebm Mobile Resources Guide: http://beckerguides.wustl.edu/mobileresources Becker Medical Library, Washington University in St. Louis 314-362-8092 [log in to unmask] On Mon, Dec 3, 2012 at 8:09 AM, Bastian, Hilda (NIH/NLM/NCBI) [C] <[log in to unmask]> wrote: Hello! Does anyone know of a report on what percentage of records in CENTRAL are unique records and what percentage aren't trials of any kind? (Ahmed, you possibly live in one of the regions of Canada that has a license. I live in an area with no license at all, and you can't see the bibliographic details for anything in CENTRAL. What you can see without a license seems to me to have got more restricted over the years, and most countries don't have national licenses.) Hilda From: "<Ahmed Abou-Setta>", "M.D." <[log in to unmask]<mailto:[log in to unmask]>> Reply-To: "Ahmed Abou-Setta, M.D." <[log in to unmask]<mailto:[log in to unmask]>> Date: Monday, December 3, 2012 3:34 AM To: "[log in to unmask]<mailto:[log in to unmask] AC.UK>" <[log in to unmask]<mailto:[log in to unmask] AC.UK>> Subject: Re: RCTs Hi Jon, For a PubMed search, you get: Randomized controlled trial[PTYP] >> 333596 Controlled Clinical trial[PTYP] >> 84639 Combined >> 413488 This will get you the RCTs/ NRCTs. As for Central being free, I just searched it without being logged on through www.thecochranelibrary.com<http://www.thecochranelibrary.com>. I had no issues with searching the library or CENTRAL. You can also download any number of citations from Central as needed, but you can’t access the full text of the reviews in the CDSR, which is password protected. I don’t think Canada has a nation-wide access policy so if it works fine here then it should also work in the UK. I definitely used to do this regularly when searching while still in Egypt a few years ago and never had any problems searching CENTRAL. Hope this helps. Ahmed From: Evidence based health (EBH) [mailto:[log in to unmask]] On Behalf Of Jon Brassey Sent: Monday, December 03, 2012 2:24 AM To: [log in to unmask]<mailto:[log in to unmask] C.UK> Subject: Re: RCTs Hi Ahmed, Thank you for that important distinction. You highlight that the number from PubMed is 413,000 which suggests around 40,000 are controlled trials (The RCT filter finds around 375). I wonder if there is a more 'relaxed' filter to find controlled trials, not just RCTs. As for Central being free, is that correct? I'd assumed you needed a subscription to Cochrane to get it, as we have in the UK. Also, it might be free access but they control the access. For what I want the RCTs for it requires a more liberal definition of 'free'. BW jon On Mon, Dec 3, 2012 at 8:14 AM, Ahmed Abou-Setta, M.D. <[log in to unmask]<mailto:[log in to unmask]>> wrote: Hi Jon, There is an important distinction here… not all citations in the Cochrane CENTRAL are RCTs. They are RCTs + Controlled Clinical Trials (i.e. reports of randomized controlled trials (RCTs) and quasi-randomized controlled trials (QRCTs) - Publication Types “Randomized Controlled Trial” and “Controlled Clinical Trial”). Therefore the actual total number from PubMed is over 413,000. Now Central also receives input from Embase, and other sources including hand-searching, but it’s also free to search. I doubt there is a one-stop shop for all RCTs/ NRCTs anywhere in the world that is free and kept up-to-date in a short time after the original databases are updated. If there is then I would love to hear about it. Hope this helps. Ahmed From: Evidence based health (EBH) [mailto:[log in to unmask]<mailto:EVIDENCE-BASED-HEALTH@JI SCMAIL.AC.UK>] On Behalf Of Jon Brassey Sent: Monday, December 03, 2012 1:49 AM To: [log in to unmask]<mailto:[log in to unmask] C.UK> Subject: RCTs Hi! Can anyone advise me the best open/free resources for RCT identification. The obvious one being PubMed and using various filters/hedges I can identify around 375,000 RCTs, but this is much smaller than Cochrane CENTRAL's 650,000+. So, I'm looking at ways to minimise that gap. Embase isn't free and open so that's ruled out. I was wondering about sources such as Mendeley and Connotea. I assume these are rich sources of RCTs?? Do people know if there are any search filters for these? Any help would be appreciated. BW jon -- Jon Brassey TRIP Database http://www.tripdatabase.com Find evidence fast -- Jon Brassey TRIP Database http://www.tripdatabase.com Find evidence fast -- Patricia Anderson, [log in to unmask] Emerging Technologies Librarian University of Michigan http://www.lib.umich.edu/users/pfa