Transcript Slide 1

The Database at the End of
the Universe: Making the Most
of PubMed
The Database at the End of The
Universe
Introduction and Overview to the
Course
Aim & Objectives
Aim
To introduce features and provide experience in using PubMed.
Objectives
Participants will be able to:
– Understand differences between PubMed and Dialog/Ovid
– Navigate PubMed and MeSH controlled vocabulary.
– Carry out free-text/field searching
– Build search strategy, use limits, locate single citations
– Manipulate result sets (i.e. saving, printing, emailing)
– Set up personalised accounts with PubMed.
– Use special queries and clinical queries functions.
– Describe other useful PubMed/MEDLINE related resources
PROGRAMME
10.15
10.30
11:10
11.30
12.00
12.15
12.30
13:00
13:45
14:15
14:45
15:15
15:.55
16:00
Introduction and Overview to the Course
Making the Most of PubMed MeSH
Exercise: Finding MeSH on PubMed
Guided ATM Tour
Cites for Sore Eyes
Exercise: Tracking Elusive Citations
LUNCH
Using Clinical and Special Queries on PubMed
Exercise: Answering Clinical Questions
Using Added Features from PubMed
Exercise: Getting added value from PubMed
Other PubMed/Medline related tools and projects
Evaluation
CLOSE
PubMed the Basics
• National Library of Medicine’s own version
of ‘MEDLINE’ with some additions
• Started in 1997 partly as replacement for
Internet GratefulMed
• First free access to ‘Medline’ via Internet
• Accessible anywhere in the world without
password at http://www.pubmed.com
Did you spot the
deliberate mistake?
Why do you use PubMed?
• Single Citation Matcher? – for checking
inaccurate references
• For finding most up to date references? –
more up to date that Dialog?
• Because you do not need a password? –
quick access
• Because some of your users use it? –
Why do they use it?
How do you react to users
using PubMed?
• Help them as best you can with your
knowledge of searching??
• Throw your hands in the air and tell them
to use DIALOG (as it is a more intuitive
interface, other biomedical databases are
available through it and full text of journals
can be accessed)??
• Somewhere in between??
Why do people use PubMed?
• Simple to use??
- demonstrate that it is different rather than
simpler than Dialog
• Currency of the records??
- probably only slightly more current (more
in a moment)
• You can use it anywhere in the world
without a password??
- cannot argue with that!
Is PubMed Medline?
• PubMed is not synonymous with Medline
• Medline records since 1966 make up the
vast majority of PubMed’s records
• All but 2% of PubMed’s records are
included in Medline produced by those
who ‘lease’ Medline such as Ovid and
Dialog
• Also includes other records whose status
is slightly different
What are these different status
records?
• In process records without full MeSH etc
• Records sent to publishers by NLM in
advance of publication of issue
• Out of scope articles – non biomedical
articles from Medline journals, e.g.
geology articles from Scientific American
• OLDMEDLINE – partial records, e.g. they
do not have full MeSH have been added
covering 1950 to 1965
Currency of PubMed
• Completely unscientific survey of journals
comparing currency of PubMed and Dialog
• Ten randomly selected journals – 6
monthly journals and 4 weekly journals
• For ‘In process’ i.e. basic record
- two journals were more up to date in
PubMed than Dialog
- Only by only one issue
Currency of MeSH
• However, for full record i.e. with MeSH,
publication type etc.
- six journals more up to date in PubMed.
- All four weeklies (including Lancet &
JAMA) more up to date by three or four
issues
• Different comparison in OVID as has
separate ‘in process’ database
The Database at the End of The
Universe
Making the Most of PubMed MeSH
The Perfect Search- Recap
• Clearly
focused question (e.g. using
PICO).
•Plenty of planning!
•Exploiting database features…MeSH
controlled vocabulary, exploding
/focusing, subheadings, special queries
etc.
•Free-text searching when appropriate
•Add it together using Boolean Logic
(AND, OR, NOT)
MeSH in PubMed- the same as
Dialog/Ovid??
• Like Dialog, PubMed has a thesaurus or controlled
vocabulary called Medical Subject Headings (MeSH)
• Same principles as Dialog/Ovid thesaurus with which you
are familiar i.e. controlled list of terms to describe what a
database record is about
• As you already know....This allows for different spellings
and synonyms
For e.g. Using the MeSH term for
myocardial infarction will retrieve
articles that talk about either
myocardial infarction or heart attack
MeSH in PubMed-some differences
• Theory behind the Thesaurus/MeSH is the same
but accessed a little differently
• MeSH is accessed by choosing the MeSH
database on PubMed: either by the PubMed
side bar or the dialog box at the top of the page
• MeSH terms are automatically exploded (more
on this later)
• Perhaps less intuitive than Dialog/Ovid?
MeSH terms for
cardiovascular disease
Searching MeSH
• Searching MeSH database is equivalent to ticking
Thesaurus Mapping checkbox in Dialog
• To search for a term, enter a term in the query box
• Click on go
List of MeSH terms….
• Like Dialog, searching thesaurus/MeSH produces a
list of terms that match your search term
• Scope notes for each term are provided under the
MeSH term
• If you use the PubMed database rather than the
MeSH database, PubMed automatically selects
MeSH considered most relevant – there is no choice!
(more on this later)
• PubMed also automatically runs a textword search for
the term you have entered
Using a MeSH term: Send to
Search box
• This next step is a little different to Dialog!
• In Dialog to use a MeSH term, you tick the box
next to the term, then click on search
• In PubMed, you use the “Send to Search Box”
feature.
• Tick the box next to the MeSH term you wish to
use
• Go to “Send to” dialogue box- click Send to
Search box with AND/OR/NOT
• Click on “Search PubMed” in new dialogue box
that appears
Supplementary Concept
Records in PubMed
• The MeSH controlled vocabulary contains
Supplementary Concept Records (SCRs) as well
as main MeSH headings.
• (You won’t use these terms to search but they
are occasionally retrieved when you carry out a
MeSH search so it is useful to know what they
are).
• SCRs are typically chemicals for e.g. calcium
acetate.
• They are NOT true MeSH headings but are
mapped to an appropriate MeSH term.
Why have Supplementary
Concept Records?
• SCRs exist so that the names by which
chemicals are indexed in the database are
controlled i.e. following same principle of
MeSH vocabulary.
• For e.g. The substance name IGSF4
protein is used for TSLC1 protein and
tumor suppressor in lung cancer 1
protein which are alternative names for
this substance.
Using MeSH terms
• Display terms in a hierarchy i.e. Thesaurus tree
• Exploding
• Majoring/Focusing
• Subheadings
• Combining terms and building a search
strategy
Displaying terms in a hierarchy
• Exactly the same as Dialog- click on a MeSH
term and it’s position in the thesaurus tree is
displayed
• At the bottom of the page, the MeSH term is
displayed in the relevant part of the thesaurus tree
Exploding
•PubMed automatically explodes MeSH terms whilst
Dialog gives users the option to explode a term
• It is possible to override automatic explosion by
entering a field tag
•For example:
Angina [MeSH:NOEXP]
Or another way to prevent
explosion of MeSH terms...
• Tick the box “Do Not Explode this term
(i.e., do not include MeSH terms found
below this term in the MeSH tree). ”
• This can be found once you’ve clicked on
a MeSH term
Focusing/Majoring
• Focusing/majoring a MeSH term in PubMed is very
different from Dialog but the theory is the same
• Need to use the “Links” functions
• “Links” can be found in tiny (!) writing to the right of the
MeSH term
• Click on “Links” and a menu comes up
• Choose PubMed Major Topic and this automatically
carries out a MeSH (major) query on the term
• Alternatively, use the appropriate field tag : Angina
[MeSH Major Topic]
To focus a MeSH term, click on “Links” which
can be found to the right of the screen
Or another way of focusing
MeSH terms...
• Tick the box “Restrict Search to Major
Topic headings only”
• This is found once you’ve clicked on a
MeSH term
Subheadings
•As we’ve seen, clicking on the
MeSH term produces a list of
applicable subheadings
•As in Dialog, choose one or more
subheadings by ticking appropriate
box(es)
Automatic explosion of
subheadings
• Choosing a particular subheading automatically
explodes that “family” of subheadings (e.g.
Ticking subheading “therapy” searches for
therapy, drug therapy, surgery etc)
• In Dialog, each subheading is chosen
individually
• Automatic explosion is overidden by ticking the
“Do Not Explode this term (i.e., do not include
MeSH terms found below this term in the MeSH
tree)” box
In summary...
• Theory behind PubMed
MeSH is exactly the
same as Dialog MeSH
• Same tools: Exploding,
focusing, subheadings
etc
• Just need to know which
button to press!
Mixing it all together
•
To combine MeSH terms and build a search
strategy, there are two ways:
1. Using the SEND to SEARCH BOX feature
2. Using the SEARCH HISTORY (to be covered
later)
Send to Search Box Feature
• You may also construct a search
strategy by using the Send to
Search Box feature.
• Whenever you carry out a search
on the MeSH database, each
MeSH term will have a little
checkbox to the left.
• Select a MeSH term by ticking the
checkbox
Send to Search Box Feature
cont...
• The Send to Search box feature is found below
Limits/Preview/Index/History/Clipboard/Details tabs
• In the box which says “Send to”, use the drop down
list to choose appropriate Boolean operator.
• Click Go, and search box appears with beginnings of
your search strategy.
• Add as many MeSH terms as you like by searching
the MeSH database- each time sending your chosen
MeSH term to the Search Box
•
Then click on Search PubMed
When to build a search strategy
using the send to search box feature
• The only time the send to search box
feature can be used is when using MeSH
database
• MeSH terms can be AND/OR/NOT-ed
through the Send to Search Box feature
Now your turn!
• Find Exercise 1: Finding MeSH on Pubmed/mixing it together
• Have a go at finding the MeSH terms and building the
following search strategies by attempting the worked
examples
• We’ll then take you on a guide ATM tour...!
Cashing in on ATM Searching
and Finding the Right Phrase
The Database at the End of The
Universe
ATM and Phrase Searching
• Demonstrate searching when PubMed is
selected as the database.
• This is the default and so demonstrates
how untrained users will search Medline.
• Includes:
- Automatic Term Mapping (ATM)
Searching
- Phrase Searching
ATM Searching??
Automatic Term Mapping!!
Automatic Term Mapping
• Standard searching with PubMed selected
from the database list
• Uses the Translation Table through which
it searches for MeSH, Supplementary
concept record, author and journal title
• Automatically applies Boolean Logic
• Comprehensive search but may be a
problem if it finds the wrong MeSH
Phrase Searching
PubMed’s Phrase List
• PubMed searches from the Phrase List if
“inverted commas” are inserted around the
search term
• The Phrase List contains several hundred
thousand phrases the majority of which
come from the NLM’s Unified Medical
Language System (UMLS)
• In some cases matching the phrase list
can be an advantage but not in others
End Users and PubMed
• Thus this section covers how the untrained
end user will search PubMed (though it is
also useful for library staff).
• Generally end user will use ATM searching
as it is the default though they can easily
learn how to insert inverted commas to
use Phrase Searching
Examples to Work Through
• The focus of this day is on library staff
rather than end users use of PubMed
• Thus there are examples to work through
of the basics of ATM searching for the next
twenty minutes
• However, the examples continue and you
can work through them at your leisure in
the future to understand how your end
users are searching PubMed.
The Database at the End of The
Universe
Cites for sore eyes?
Author Index and Journal
Index Searching
Translation Table
As we have seen PubMed looks in order for:
- MeSH
- Substance Name
- Journal Name
- Phrase List –
- Author List – (similar to journal name)
- All fields searching
This is a simplified version of the Translation
Table
Author Index
• If you enter an author’s name it is unlikely
to match a MeSH or Substance Name
• PubMed looks for name in Author index
• You can enter surname and first name as
well as simply the surname
• However, results depend on how name is
entered. “No exact match” leads to
separation and ATM searching
Journal Index
• If you have entered a journal name into
the search box then it is unlikely to have
matched any of the steps from the
Translation Table
• PubMed looks for the journal in the journal
index
• However, you need to be careful how you
enter a journal e.g. British Medical Journal
not BMJ
Finding a specific journal article
• Single citation matcher is a tool
to find a specific article of
interest.
• Located on the PubMed sidebar
• Very useful for finding out correct
details for inaccurate references
Finding a specific journal article 2
•Searching for a citation by
- author
- first page
-publication year
will produce most accurate
results
•For e.g. “Thomson”, page 299,
date 2003
Finding a specific journal article 3
• Avoid using journal names or journal
abbreviations unless journal title is
straightforward.
• e.g. PubMed is not able to retrieve
citations for the journal Bioorganic
Medicinal Chemistry if abbreviation
“Bio Med Chem” is used. This is
because the correct abbreviation is
“Bioorg Med Chem”.
Standard Journal abbreviations
• All that JAS: Journal Abbreviation
Sources. Registry of Web resources that
list or provide access to full titles for
journal abbreviations.
http://www.public.iastate.edu/~CYBERSTA
CKS/JAS.htm
What is the batch citation matcher?
• The batch citation matcher is primarily a
tool for publishers
• It allows publishers to match online
references (e.g. in the bibliography of an
online article) to PubMed records) i.e. can
find more than one citation at a time
• Little functional use for librarians/clinical
staff
Journals database
• Can be used to find citations from
a particular journal
• Basic information about journal is
also provided
• Accessed from PubMed sidebar
Now your turn!
• Find Exercise 2: Tracking elusive citations
• Have a go at identifying the missing
components for the journal references
12:30 LUNCH
The Database at the End of The
Universe
Clinical and special (including
subject) queries
Many types of clinical question
Types of question - Types of article primary studies
secondary
studies
• Reviews
• Diagnosis
• Economic Evaluations
• Etiology
• Guidelines
• Prognosis
• Therapy
[and now Qualitative]
How can I limit to a particular
study type/subject area?
• By using search filters or hedges
• Search filters are standardised search strategies
designed to retrieve studies of a particular
methodology type e.g. randomised controlled
trials, systematic reviews
• Hedges are standardised search strategies
designed to identify studies in a particular
subject area
How filters/hedges work…
• Filters and Hedges work in one of two ways:
– by identifying particular publication types or
study designs most likely to answer a question
– by isolating subject or free-text terms most likely
to be associated with high-quality studies
– Publication type e.g. Randomized controlled trial
– Subject heading e.g. Clinical trials, diagnosis
– Free text e.g. “systematic review”
In PubMed, we don’t need to worry
about devising our own filters!!
• PubMed has built-in search “filters”
• Based largely upon work of Health
Information Research Unit at McMaster
University
• Allow you to search for publications
according to study type, subject area or
other special queries
• Quick way of homing in on information
In-built search filters in PubMed
• Clinical queries
• Special queries
• Subject queries
Clinical Queries
• PubMed can limit searches to following
clinical study type categories
-
Aetiology
Diagnosis
Therapy
Prognosis
Clinical prediction guides
Systematic reviews
• Can apply these concepts in a broad,
sensitive search or narrow, specific search
Broad, sensitive clinical queries
• Ensure you retrieve every citation that is
applicable to a particular query
• For e.g. A broad, sensitive search using
the “Therapy” clinical query will ensure all
relevant articles that discuss therapy,
treatment etc are returned
• Larger result set, increased frequency of
irrelevant articles
Narrow specific clinical queries
• Are focused and increase the number of
relevant citations applicable to a clinical
query
• Reduce the number of results but these
results will be more relevant
• Since search is not as sensitive, possibility
of missing relevant citations
Searching using the clinical
queries
• Basic search box
• Don’t forget to use Boolean logic and
truncation in basic “Clinical Queries”
search box
• Can use “Search History” to build
searches
For example…
• Is Zanamivir an effective treatment
for influenza?
What type of question is this?
• Treatment/therapy question
• Therefore use a treatment/therapy filter to
hone in on relevant studies
• As this is a clinical question, not a
special/subject query, choose PubMed
“Clinical queries”
Special queries
•
•
As well as “Clinical Queries”, PubMed
has a “Special Queries” function.
The two most useful special queries are:
- Health services research
- Cancer topics
Health Services Research
Special Query
• This feature limits search to various topics
on health care quality and costs
• 6 sub-categories:
» Appropriateness
» Process assessment
» Outcome assessment
» Costs
» Economics
» Qualitative research
• Again, searches can be broad & sensitive
or narrow & specific
For example…
• Is Zanamivir cost-effective in the
treatment for influenza?
Cancer Topics
• Another special query is “Cancer Topic”
• If you choose this option, you will be
redirected to another webpage
• From this page, you can choose a
particular cancer topic
• This will then redirect you back to the
PubMed search interface and the cancer
topic query will appear in the search box
Using the Search History
• PubMed keeps a record of every query
you carry out.
• History remains until there is 8 hours of
inactivity or until searches are manually
deleted from “History”.
• To access history, click on “History” tab.
History Tab
Search History
Using the search history
• Click on the hyperlink for the
appropriate search query
• Choose the appropriate Boolean
operator from the menu
• Once you have constructed your
query, click GO
Each search is hyperlinked. Click on a search
number and a menu will appear
Choosing AND, OR or NOT
will send your query to the
search box
When to build a search strategy
using search history
• If you are using a combination of free-text
terms and MeSH terms, the only way to
build a search strategy is by using the
History Tab
• When using the clinical, special and
queries function- using the history is a
good way to build a search strategy
Using Search History...
Can artery injury occur after elbow
dislocation?
Applying PICO...
•
•
•
•
Patient/population/condition: Artery injury
Intervention/Exposure: Elbow dislocation
Comparison: None
Outcome: Morbidity, mortality
Using the Search History...
• Both terms are not included in the
thesaurus/MeSH
• Therefore, must enter these terms using
free-text searching and PubMED database
(rather than MeSH database)
• Step 1: Search for “Elbow dislocation”
• Step 2: Search for “ Artery Injury”
• Step 3: Use “History” tab to AND these
terms together
Use “History” to build a search
• Record of terms and phrases used to
search for “Cancer topic” (i.e. filter/hedge)
kept in “History” tab
• Applies to any PubMed queries: clinical,
special or subject
• Use this to build a search strategy
Subject queries
• PubMed also has built-in filters to limit a
search to a particular subject area
• Includes:
»
»
»
»
»
»
»
»
AIDS
Bioethics
Cancer
Complementary medicine
History of medicine
Space life sciences
Smallpox
Toxicology
Clinical and Special Queries:
Pros and Cons
• Clinical
and Special queries are fast
and effective ways of limiting your
search to a particular research
methodology e.g. Therapy, Diagnosis
etc
• Not comprehensive and may omit
potentially useful results
• For a more thorough way of limiting
your search to a particular clinical query,
use a search filter.
Now your turn!
• Find Exercise 3: Answering clinical questions
• Have a go at using the clinical, special and subject
queries to find answers to the scenarios
The Database at the End of the
Universe
Using Added Features from
PubMed
What we’ll cover...
• Limits- including the personalised limits
that can be set up using My NCBI
• Related links
• Manipulating results: i.e. saving, printing,
emailing
• Some hints and tips
Limits
• You can limit searches to one of a number
of fields
• Similar categories to Dialog Medline
• Will limit term that is presently in search
box
• Tick appears next to “Limits” tab when a
limit is currently applied.
• Below this, the limits being applied will be
listed
Remember to deselect limits!
• Limits remain applied until you deselect
them
• Click on “Limits” tab, scroll to bottom and
choose: “Clear all limits”
Related Articles
• PubMed provides links to “Related Articles” for
each database record
• To work out degree of similarity between
database records, PubMed applies a special
algorithm
• Basically this looks at (a) terms used,(b) number
of times terms are used and (c) importance of
each term (i.e. more important if in “Title” field)
• To find out more about this algorithm read:
http://www.ncbi.nlm.nih.gov/entrez/query/static/c
omputation.html
Manipulating results and
searches
• Same options as Dialog
–Printing, emailing and saving
searches
–Saving and re-running
searches
Printing, Emailing and Saving
Results
• Select results you wish to keep by ticking
check boxes
• Then use “Send to Search Box” feature
• The options you need are:
» File (to save results as a text file)
» Printer
» Email
Saving results
• After selecting results you wish to keep,
choose “Send to File”
• You are then prompted to save the file (as
a text file)
• If this fails to work, pop-ups may be
blocked on your computer
• Simply, choose “Send to File” again and
this time you will be prompted to save the
file
Printing
• After selecting results you wish to keep,
choose “Send to Printer”
• The Printer dialogue box will appear
• If this fails to work, it’s just because popups may be blocked on your computer
• Simply, click Printer if this fails to work and
the dialogue box will appear
Emailing results
• After selecting results you wish to keep,
choose Send to Email
• Fill out details for e-mail destination
address
Saving searches
• Need to set up a My NCBI account
• To do this click on “My NCBI” on PubMed
services toolbar on left hand menu
• Two options:
– To rerun the search in its entirety
– To update a search i.e. retrieve only those results that
have been added to PubMed since you last ran the
search
Re-running an entire search
• Click on Save Search (next to the search
box)
• This prompts you to log into your “My
NCBI” account
• Name the search and that’s it!
• To rerun search, log into your “My NCBI”
account
• To re-run entire search, click on search
name itself.
Updating a search
• Select search by using checkboxes on the
left side
• Click on “What’s new for selected”
Hints and Tips
•
•
•
•
Personal Name as Subject, not Author
Stopwords
Systematic Review and Full-Text Subsets
Combining “Related Articles” with Subject
Searching
Personal Name as Subject
• Most names on PubMed are Author
Names
• However occasionally you might be
interested in a person as the SUBJECT of
an article
• E.g. Articles on David Sackett or Muir Gray
• You use the PS field e.g. sackett d[ps]
Stopwords
• Pubmed assigns 132 words as stopwords
• Dialog has 9 [AN, AND, BY, FOR, FROM,
OF, THE, TO, WITH] Ovid 160,
Silverplatter 66
• These include the usual suspects but also
words such as “especially” and
“significantly”
• “and” is a stopword but “or” and “not” are
not!
Systematic Review and
Full-Text Subsets
• Systematic Review filter on PubMed
Clinical Queries does not use a string of
search terms.
• Instead it uses a special subset
systematic[sb]
• This means it can be used as a command
from the search box OR
• It can be applied as a “Limit”
Combining “Related Articles”
with Subject Searching
• Many databases now have “Related
Articles” features
• However PubMed stores the set of
“Related Articles” as a step in the search
“History”
Related Articles for PubMed (Select nnnnnnnn)
• This means results can be combined with
a subject or text word or other search term
for added relevance
Now your turn!
• Find Exercise 4: Getting added value from PubMed
• Have a go at practising your search skills again this time try and build your strategy using the
search history. Have a go at manipulating the
results, saving and re-running searches and looking
at some of the hints and tips just mentioned
The Database at the End of The
Universe
Other PubMed/Medline
related tools and
projects
Some PubMed Related Tools
and Projects
•
•
•
•
•
•
•
SUMSearch
NLM Gateway
MEDLINE Plus
Interact (formerly Slim Slider)
AskMedline
ReleMed
PubMed Informer
SUMSearch
What is it?
Where can I access it?
• Meta-search engine from http://sumsearch.uthscsa.edu/
Society for General
Internal Medicine that
searches high quality
resources, depending on
question type.
How does it work?
• Relies on MeSH.Quality
control includes checking
for MeSH and prompting
for truncation.
NLM Gateway
What is it?
• Web-based system to allow
"one-stop searching" for
many NLM information
resources or databases.
How does it work?
• Connects users with
multiple NLM retrieval
systems and also provides
search interface for meeting
abstracts. Meeting
Abstracts collection is found
only in the Gateway.
• Where can I access it?
http://gateway.nlm.nih.gov/
MEDLINE PLus
What is it?
• Patient-focused resource
that brings together
authoritative information
from NLM, the National
Institutes of Health (NIH),
and other government
agencies and healthrelated organizations.
How does it work?
• Preformulated MEDLINE
searches are included in
MedlinePlus giving easy
access to medical journal
articles.
• Where can I access it?
• http://medlineplus.gov/
Interact
What is it?
• Alternative search
interface for
MEDLINE/PubMed to
provide assistance to
novice user and added
convenience to advanced
user.
• You can “Delete” results
or mark “High” relevance
***Has a new PubMed
Interact Tutorial***
How does it work?
• Formerly 'Slider Interface for
MEDLINE/PubMed
searches' (SLIM) with
JavaScript slider bars to
control search parameters.
New version utilises recent
developments in Web
technologies.
Where can I access it?
http://pmi.nlm.nih.gov/interact
askMedline
What is it?
• Search tool for MEDLINE/
PubMed that uses freetext, natural language
queries (without
specialized vocabularies
an expert searcher might
use), to find relevant
citations.
How does it work?
• Translates a question into
an efficient search. It has
an (optional) PICO
Interface.
Where can I access it?
http://askmedline.nlm.nih.gov/
ask/ask.php
askMEDLINE
ReleMed
What is it?
Where can I access it?
• A search engine for MEDLINE http://www.relemed.com
that increases specificity and
precision of retrieval by
searching for query words
within sentences rather than
the whole abstract.
How does it work?
• Uses sentence-level
concurrence to statistically
analyse relationship between
search terms. Also estimates a
relevance score and sorts
results on this basis, shifting
irrelevant articles lower down
the list
PubMed Informer
What is it?
Where can I access it?
• An online tracking tool • http://pmi.nlm.nih.gov/
that monitors keyword
searches for
MEDLINE/PubMed.
How does it work?
***Requires Registration***
• Sends email alerts,
RSS feeds (or SMS
texts - U.S. only)
References - 1
• Booth A (2002) MEDLINEplus: a golden gateway to
health information resources. Evid Based Med 7:
136.
• Booth A & O'Rourke A (2000) Resource corner.
SUMSearch and PubMed: 2 Internet-based
evidence-based medicine tools. ACP Journal Club.
132(3):A-16
• Fontelo P, Liu F, Ackerman M (2005).
askMEDLINE: a free-text, natural language query
tool for MEDLINE/PubMed. BMC Med Inform Decis
Mak. 5(1):5.
• Haynes RB, Wilczynski N. Finding the gold in
MEDLINE: clinical queries. ACP J Club.
2005;142:A8–9.
References - 2
Muin M, Fontelo P. Technical development of PubMed interact:
an improved interface for MEDLINE/PubMed searches. BMC
Med Inform Decis Mak. 2006 Nov 3;6:36.
Muin M, Fontelo P, Liu F (2005) Ackerman M. SLIM: an
alternative Web interface for MEDLINE/PubMed searches - a
preliminary study. BMC Med Inform Decis Mak.1;5:37.
Muin M, Fontelo P, Ackerman M (2005) PubMed Informer:
monitoring MEDLINE/PubMed through e-mail alerts, SMS,
PDA downloads and RSS feeds. AMIA Annu Symp Proc.
1057.
Muin M, Fontelo P, Ackerman M. PubMed Interact: an
Interactive Search Application for MEDLINE/PubMed.
AMIA Annu Symp Proc. 2006;:1039.
Siadaty MS, Shu J, Knaus WA (2007) Relemed: sentence-level
search engine with relevance score for the MEDLINE
database of biomedical articles. BMC Med Inform Decis Mak.
7:1.
The Database at the End of The
Universe
Evaluation and Close