Tuesday, May 31, 2005

Full author name searching enhanced: PubMed

You can now search full author names in PubMed!

The National Library of Medicine has included full author names in MEDLINE records since 2002 - except until now the full author field has only been a display field - you could never search the field.

Just type the full author name on the query line: julia s wong

Some tips:
  • You can use the Single Citation Matcher but not the new first author name search feature.
  • Searching by full author name limits to citations to articles published from 2002 forward, and to journals that publish using the full names of authors.
  • You can type the name in natural (julia s wong) or inverted order (wong julia s).
  • You can browse full author names in the Full Author Name index available on the Preview/Index screen. Select Full Author Name on the fields pull-down menu, enter a last name in the box, and click on Index.

Give it a try!

For more details, see the NLM Technical Bulletin

Wednesday, May 25, 2005

Related Records: Web of Science

Try out the 'Related Records' feature from any of the Web of Science databases - Science Citation Index, Social Science Citation Index, and the Arts & Humanities Citation Index.

Once you click on the related records link for a particular article, the Summary page displays a list of articles whose cited reference lists include at least one of the sources cited by the original article (the original articles title always appears at the top of the page -- just to help keep you oriented). Articles that share the largest number of sources with the original article are listed first.

Related Records searching assumes that articles citing the same works have a subject relationship, regardless of whether their titles, abstracts, or keywords contain the same terms. The more cited references two articles share, the closer this subject relationship is.

You can use this feature to find "more like this" articles. For example:

1. After searching on the topic otitis media AND diagnosis, you find the article by K. Blomgren titled " Current challenges in diagnosis of acute otitis media".
2. Click on the Cited References to examine the article's cited reference list.
3. Most of the 49 cited references look relevant so you click on Find Related Records.
4. The Related Records - Summary page displays articles whose reference lists include at least on of the articles cited by the Blomgren article.

    The articles that share the greatest number of references with the original article appear at the top of the list.

    You can also click on the related records link for any individual article as well.

    To view the references shared by the two Related Records, click the link in the Shared Refs column. The number in this column is the number of references shared by the two Related Records. The number in the Cited Refs column is the total number of references in the record's cited reference list.

    Give it a try!

    Tuesday, May 24, 2005

    Related Articles: PubMed

    The related articles link in PubMed retrieves a pre-calculated set of citations that are similar to or related to the original citation.

    Some of the advantages of related articles are that it's a rapid, easy way to identify like citations. You can also pick up citations that other strategies might miss. All in all, a good way to quickly find relevant citations.

    Some disadvantages are that it is not comprehensive. Any limits you applied to your original search will be not be applied to the pre-calculated related articles. By using History you can reapply limits (e.g., english) but this removes the ranked order and may remove citations that are most relevant. You will also notice that relevancy can drop off quickly. And, in order to use this feature, you need to start with a relevant citation.

    The bottom line is that you should definitely add 'related articles' to your arsenal of strategies as you search for information.

    For more information:
    Related Articles in PubMed
    Computation of Related Articles in PubMed

    Monday, May 23, 2005

    Related Articles

    There are several systems that provide a 'related articles' or 'related records' feature. This feature can be incredibly handy for many reasons.

    'Quick and Dirty' Searching *
    You need a few more articles on a particular topic and already have in your hands a highly relevant article. You want to find more like it. By using the 'Related Articles' link, you will come up with closely related citations.

    'Needle in a Haystack' Searching
    You are looking for something in a context that very little seems to be written about. You do a textword search, identify an appropriate article from the retrieved list and then click on the 'related articles' link. Now, you should get other citations that are closely related.

    'Looking for MeSH Terms' Searching **
    Sometimes it's hard to figure out the best MeSH terms to use when doing a comprehensive search but once you've located a relevant article, click on the 'related articles' link. Then methodically examine the subject heading fields of the best citations to identify a list of potential subject headings. With that list in hand, conduct the search again using your list of appropriate MeSH terms.

    Now for an example of how this works. Go to PubMed and, on the query line, type: killer pop machines
    You'll get one citation which doesn't give you much information. Click on the 'related articles' link.

    Tune in tomorrow for an in-depth look at PubMed's Related Articles.

    * As with any 'Quick and Dirty' method, this will NOT be comprehensive but will, more often than not, retrieve very relevant citations.
    ** Only works in databases that utilize MeSH (oddly enough!)

    Friday, May 20, 2005

    Where's my favorite journal indexed?

    If you've ever wondered how to find out where a particular journal is indexed, the perfect tool for that job is Ulrichs Periodical Directory (ulrichsweb.com).

    There are two reasons why this information can be helpful:

    1. You don't want to submit an article to a journal that isn't indexed anywhere. No one will ever find your article. Check Ulrichs to make sure that the journal is indexed and that the databases that it is indexed in make sense.

    2. You're searching in your favorite database and realize that there aren't any articles from a particular journal that you know covers the topic. Check in Ulrichs to see if that journal is indexed in the database you're using. If it isn't, find another database where that journal is covered.

    You can find Ulrichsweb.com on the Library's homepage under 'All Major Resources A-Z'
    Once you get to Ulrichs, type the journal name in the query box. Then click on the correct journal title. At the next screen, click on the tab 'Abstracting/Indexing and Article Access' and that's where you'll find the list of database services that index the journal.

    Try it out!

    Thursday, May 19, 2005

    PubMed's Single Citation Matcher Enhanced

    First author search and auto-complete for journal titles added! This is great news!

    Now you can search for the first author of an article by simply checking the 'Only as a first author' box in the Single Citation Matcher. Alternatively, you can use the query box and type the author's last name + initials followed by the tag [1au].

    And even more helpful - an autocomplete feature for journal titles has been added. This feature suggests titles when you type in an abbreviation or a full journal title. When you see the title you want, just stop entering and select it.

    • You can find the Single Citation Matcher on the blue scan column under PubMed Services.
    • For more information, see the NLM Technical Bulletin

    Wednesday, May 18, 2005

    Quick and dirty searching

    You know, sometimes you just need to find something on a topic really quickly. Here's a way to do that. After determining just what it is that you looking for and prioritizing the concepts, look for the highest priority concepts to be in the title of the article.


    You're looking for the effects of children's exposure to violence. Here's what you might try:

    OVID: violen$.ti. and (expos$ or witness$).ti. and child$.ti.
    [alternate: (violen$ and (expos$ or witness$) and child$).ti.]

    PubMed: violen*[ti] AND (expos*[ti] OR witness*[ti]) AND child*[ti]

    You're looking for the management of abruptio placentae. Here's what you might try:

    OVID: manage$.ti. and abruptio placenta$.ti.

    PubMed: manage* AND abruptio placenta* Only this time, click on Limits and choose Title from the All Fields drop-down. Then click on Go.

    • This technique shouldn't be used when doing an exhaustive search.
    • This technique is quick with a high degree of relevance.
    • Tip for authors: don't use 'cute' titles for your articles.
    • This technique also works well when you're trying to identify the appropriate subject headings for an elusive topic.

    Tuesday, May 17, 2005

    Where do I search, part 2

    There are other things to consider as you're thinking about your topic and where to search.

    Are you in a hurry?
    Are you familiarizing yourself with a new area?
    Are you doing a comprehensive, exhaustive search?
    Are you looking for cutting edge information?

    These questions help you plan your method of attack. Picking the right resource to match type of search you're doing will help ensure a higher level of success.

    For example, if I'm in a hurry, I'm going to look at resources that synethize or synopsize information (textbooks, encyclopedic handbooks) or a couple of good recent articles. If I'm trying to familiarize myself with a new area, I'm going to look for a recent book, review articles, a textbook on the topic or check resources that synopsize information. For a comprehensive, exhaustive search, I'll use bibliographic and citation databases. If I'm looking for cutting edge information, I'd check the news and web resources.

    Monday, May 16, 2005

    I've got a question...where do I search?

    Has this happened to you?

    You have a question, go to your favorite database and complete your search but the results surprise you - maybe they're too general or they don't seem to address what you think should be the key issues surrounding the topic or too few show up although you know that this is a hot topic.

    Choosing the 'right' place to search for information is sometimes difficult. How do you know which database or databases are the most appropriate for the topic?

    Consider the scope of your question. For example, if my question was "What are effective interventions for children exposed to violence?", then I'd be thinking about what disciplines would be wrestling with this topic. Here's where I would consider looking for information: psychiatry/psychology, medicine/health, social sciences, religion (faith-based interventions), criminal justice/law.

    Consider the scope of the database. Here's where you need discover what subjects/topics that the database covers. And the best way to do that is to look at information pages describing the details about the database. At our library, check out the databse information page by going to the All Major Resources A-Z and clicking on the name of the database.

    Match your question to the appropriate database(s)!

    Friday, May 13, 2005

    PubMed for Handhelds

    In case you're a PDA or other handheld device user, don't forget that PubMed is available for your device. Here are two articles from the NLM Technical Bulletin that spell out the details:

    PubMed Available for Use on Handheld, Wireless Devices

    PubMed for Handhelds offers Searching via PICO

    Thursday, May 12, 2005

    Auto Explode

    The term Auto Explode is used by Ovid. It means that the subject heading that you've selected will automatically explode - the system will automatically search the particular subject heading and any others that are more specific. Auto Explode is the default choice on the Mapping Display screen.

    PubMed automatically explodes the subject heading from the query line as part of the searching algorithm. The only way to NOT explode is to use the MeSH Database and select "Do Not Explode this term".

    Wednesday, May 11, 2005

    Explode and Focus in PubMed

    The MeSH Database is where you control Explode and Focus in PubMed.

    1. Select MeSH Database from PubMed Services on the blue scan column on the left.
    2. Type in your subject heading.
    3. On the next screen, click on the subject heading (not the check box)

    Below the definition or scope note of the subject heading, you'll see the list of subheadings that you can choose from. Further down you'll see:

    Check Restrict Search to Major Topic headings only to focus your search.

    PubMed automatically explodes the subject heading so check Do Not Explode this term to turn off the explode feature.

    Tuesday, May 10, 2005

    Explode and Focus in Ovid MEDLINE and CINAHL

    It is possible to both Explode and Focus a MeSH descriptor. And here's what you do:

    1. In the query box, type: multiple sclerosis
    2. Here's what you see on the Mapping Display screen:

    3. The subject heading and Auto Explode are automatically selected. If you want to focus the subject heading, then click in the Focus check box. If you are interested in all subheadings then click in the Include All Subheadings check box.

    Making these selections from the Mapping Display screen is very handy, especially when you know already that you want to explode and focus. If you decide not to check the Include All Subheadings box, then the next screen will be the Subheadings Display screen. At this point you can select specific subheadings for your search.

    If you have clicked on a subject heading because you want to see if there are more specific subject headings in the tree structure, then you'll need to check the Explode box. As soon as you are in the Tree Structure Display screen, Auto Explode is turned off.

    Monday, May 09, 2005

    NLM Mobile

    A new Web site, NLM Mobile, is now available. This site will maintain information about tools for handhelds being developed at NLM. Topics available:
    • NCBI Bookshelf
    • PubMed® for Handhelds
    • PubMed on Tap
    • Wireless System for Emergency Responders (WISER)

    Friday, May 06, 2005

    PubMed Health Services Research Queries

    This new Health Services Research (HSR) queries page will be extremely helpful for searches on healthcare quality and costs. There are six queries (aka hedges or filters) for you to choose from:

    Process assessment
    Outcomes assessment
    Qualitative Research

    You can also select a broad, sensitive search or a narrow, specific search.

    As with the Clinical Queries page, you can type your search terms in a query box, select the query category and the scope, and then click on 'Go'. The results page is the PubMed results page and from that point on, you can use any of the features that you would normally use (limits, related articles, etc.).

    You really should give this a try!

    Definitions of the queries are included on the HSR queries page.

    Thursday, May 05, 2005

    Focus that subject heading

    To explain the use of Focus it's necessary to go back to those busy subject indexers who are reading and analyzing journal articles in order to assign the most specific descriptors available to describe the article's subject content.

    The subject indexers assign between six and fifteen descriptors; at the end of this process, they select up to four that signify the major points of that article. When you look at the MeSH field in a MEDLINE record, any descriptor marked with an asterisk indicates a major point. Focusing restricts a subject search to only those articles where that subject is one of the major points.

    Here are two scenarios where you can use Focus to your advantage:

    1. You're looking at diabetes mellitus and coronary artery disease. Beyond that you're not sure how you want to limit the search or what other concepts you might want to add - you're just sort of looking at what's out there. In this case, it's best to focus both diabetes mellitus and coronary artery disease and even then you'll probably have a rather large results set. But in the results set that you do get, you can be certain that your two topics will be the main thrust of the citations. Then, of course, you'd add your other limits (reviews, English, human, year range, etc.). This is a good way of scanning citations to get an idea of what's out there on your topic.
    2. You're doing a search on diabetes mellitus and articles about hip replacement appear in the results set. After looking at a few citations you realize that the main point of the these citations is hip replacement and diabetes seems to be a side issue. By Focusing diabetes mellitus, all of the citations where diabetes is NOT one of the main points will drop out. It's a way of choosing which aspect of your search question has a higher priority and the results set will reflect that. Also, you can always 'unfocus' if that strategy doesn't seem to work for your question.

    You can explode and focus at the same time.

    Wednesday, May 04, 2005

    Explode that subject heading

    While this sounds a tad violent, I assure you it's not. And to adequately describe what happens when you 'explode' the subject heading we need some illustrations.

    Let's look at the tree structure for antibiotics:

    The tree shows that within the category of Chemicals and Drugs there is a subcategory of Pharmacologic Actions and a further subcategory of Anti-infective Agents. Our term ‘antibiotics’ mapped to the MeSH descriptor Anti-bacterial Agents which is a more specific subcategory within Anti-infective Agents. [For the purposes of this illustration, this tree has been severely pruned].

    If I search on ‘Anti-bacterial Agents’ as a descriptor then only the articles that were assigned the descriptor ‘Anti-bacterial Agents’ will be retrieved. Now we know from the last post that the subject indexers assign the most specific term available. The citations retrieved will discuss ‘Anti-bacterial Agents’ as a group or in general (maybe this set will also include a new type of anti-bacterial agent).

    What if our search question is “Which antibiotics are most effective in treating acute otitis media”?

    If we use the descriptor ‘Anti-bacterial Agents’ then there’s the potential of missing relevant articles that have been assigned specific drug terms, such as Amoxicillin, and not the broader term ‘Anti-bacterial Agents’. This, finally, is where EXPLODE comes in to play.

    Let’s look at some numbers:

    1. exp Otitis Media/dt (2956 citations retrieved)
    2. Anti-bacterial Agents/ (124687 citations retrieved)
    3. exp Anti-bacterial Agents/ (329921 citations retrieved)
    4. 1 and 2 (1249 citations retrieved)
    5. 1 and 3 (2183 citations retrieved)
    6. 5 not 4 (934 citations retrieved)

    In search statement (SS) 1, I exploded ‘Otitis Media’ picking up ‘otitis media’ or ‘otitis media, suppurative’ or ‘otitis media with effusion’ or ‘mastoiditis’. And because this is a drug therapy question, I attached the subheading ‘drug therapy’ (more about this later) to the descriptor thereby narrowing the types of citations to only those that discuss drug therapy of otitis media.

    SS 2: I looked for ‘anti-bacterial agents’ in the descriptor field and retrieved only those articles with that descriptor assigned.

    SS 3: I exploded ‘anti-bacterial agents’ picking up ‘antibacterial agents’ or ‘alamethicin’ or ‘amdinocillin’ or ‘amodinocillin pivoxil’ or ‘amikacin’ or ‘amoxicillin’ or [160 more named substances].

    SS 4: I combined my otitis media set and anti bacterial set

    SS 5: I combined my otitis media set and the exploded anti bacterial set

    SS 6: Using the NOT operator, I looked at what I would have missed if I used only the descriptor ‘anti-bacterial agents’ without exploding and picking up all the named drug terms.

    Here’s some titles from the set of citations I would have missed:

    • Primary care based randomized, double blind trial of amoxicillin versus placebo for acute otitis media in children aged under 2 years
    • Five-day versus ten-day treatment of acute otitis media with cefprozil
    • Comparison of five-day cefdinir treatment with ten-day low dose amoxicillin/clavulanate treatment for acute otitis media
    • Use of oral cephalosporins in the treatment of otitis media in children

    Tuesday, May 03, 2005

    MeSH: Hierarchical List

    This post examines the Hierarchical List of the Medical Subject Headings (MeSH).

    MeSH descriptors are organized in 15 categories. For example, category A for anatomic terms, category B for organisms, C for diseases, D for drugs and chemicals, etc. Each category is further divided into subcategories and within each subcategory, descriptors are displayed hierarchically from the broadest to the most specific.

    Because of the branching structure of the hierarchies, these lists are sometimes referred to as "trees". Each descriptor appears in at least one place in the hierarchy but may also appear in other categories. A good example is Breast Neoplasms. This descriptor appears in the Diseases category 'Neoplasms' in the subcategory 'Neoplasms by Site'. It also appears in the Diseases category 'Endocrine Diseases' in the subcategory 'Breast Diseases'. Here is what it looks like:

    Why is this important?

    The subject indexers are instructed to find and use the most specific MeSH descriptor that is available to represent each indexable concept. For example, articles concerning Streptococcus pneumoniae will be found under the descriptor Streptococcus Pneumoniae rather than the broader term Streptococcus, while an article referring to a new streptococcal bacterium which is not yet in the vocabulary will be listed directly under Streptococcus.

    It makes sense then that the searcher would consult the trees as well. Here's where you find out how your topic fits in the hierarchical structure and whether you need to select a broader or more specific descriptor to tease out the most relevant articles.

    Monday, May 02, 2005

    New in PubMed: Special Queries

    A new link called Special Queries has been added to PubMed's blue side bar. Like the Clinical Queries link that takes you to a page providing specialized PubMed searches for clinicians, the Special Queries link provides access to a directory of topic-specific PubMed queries. The directory includes links to:

    • Queries targeted to Clinicians and Health Services Researchers, including a listing for the standard Clinical Queries, Health Services Research (HSR) queries, and the Cancer Topic Searches (offered by the National Cancer Institute).
    • Subject queries, such as AIDS, Bioethics, Systematic Reviews
    • Special Search Queries or interfaces for complementary medicine, History of Medicine, and MedlinePlus health topics.
    • Journal Collection Queries

    The Special Quereis directory provides links to either special search interfaces, or pre-sets the Limit function of the standard PubMed interface to the selected subject or journal subset.