Medical Literature Analysis and Retrieval System

(MEDLARS)

By
Pallab Das
MLIS (DL)
Roll – 001900803004
JADAVPUR UNIVERSITY
• MEDical
Literatures

Analysis

Retrieval

System
• M EDLARS, Medical Literature Analysis and
Retrieval System.
• Newest, most sophisticated, and costliest
method yet devised for controlling the vast
flood of medical literature published
throughout the world.
 Time coverage

MEDLINE includes literature published

from1966 to present, and selected coverage
of literature prior to that period.
 COVERAGE ITEMS
•  25 million references to journal articles in life
science
• 18,000 separate journals were initially
received at the National Library of Medicine.
• 2,300 are indexed into MEDLARS
 COVERING AREAS

1. Medicine,
2. Nursing,
3. Pharmacy,
4. Dentistry,
5. Veterinary Medicine,
6. And Health Care.
 Development of MEDLARS
• In 1957 the staff of the NLM started to plan the
mechanization of the Index Medicus.
• By 1960 a detailed specification was prepared and
by the spring of 1961 a request for proposals was
sent out to 72 companies to develop the system.
• MEDLARS cost $3 million to develop and at the
time of its completion in 1964, no other publicly
available, fully operational electronic storage and
retrieval system of its magnitude existed.
NATIONAL LIBRARY
OF MEDICINE (USA)
NLM OFFICIAL WEBSITE
 MEDLARS Online

• In late 1971, an online version called MEDLINE

("MEDLARS Online") became available.
• This early system covered 239 journals and boasted
that it could support as many as 25 simultaneous online
users
• This situation continued through the beginning of the
1990s and the rise of the World Wide Web.
• a free public version of MEDLINE was instigated
• This system, called PubMed, was offered to the
general online user in June, 1997
Official Website of MEDLINE
 Utility of MEDLARS
• : MEDLARS is known as the biggest
bibliographical database of international level.
• It is important not only for the experts of
medical science, but also for other scientists ,
sociologists, trades and businessmen etc.
• It has become more and more useful and its
scope is very much wide.
• It also has been useful for the people which are
not concerned with medical sciences.
 National Library of Medicine classification
System
• The National Library of Medicine (NLM) classification system is a library 
indexing system covering the fields of medicine and preclinical basic
sciences.

• The NLM classification is patterned after the Library of Congress (LC) 

Classification system
 INDEXING
• The MEDLARS data base is created by indexers
• who analyze and describe the contents of
journal articles by means of carefully selected
terms.
• These terms, called subject headings or
main headings, are contained in an authorized
list.
 MeSH : Key to MEDLARS
Known as MeSH, for Medical Subject
Headings,this controlled vocabulary is
essential for effective use of the system, since
only MeSH terms are used to index articles for
MEDLARS. MeSH is updated and published
annually.
 MeSH types
MeSH

8,000 terms are listed Rank Order , categorized list

alphabetically and appear in showing
bold, upper case type. The rank from broad to narrow
 The 14 broad subject categories of
MeSH
1. Anatomical Terms
2. Organisms Sociology, and
Social Phenomena
3. Diseases
4. Chemicals and Drug
5. Analytical, Diagnostic, and
Therapeutic techniques
and Equipment
6. Psychiatry and Psychology
7. Biological Sciences
8. Anthropology, Education
9. Technology, Industry and
Agriculture
10. Humanities
11. Communication, Library
Science and Documentation
12. Named Groups of Persons
13. Health Care
14. Physical science.
 In-House MeSH
• This is the more current and updated version
of MeSh
• This in-house MeSH is revised auarterly
• and it contains about 1,000 more terms than
the published MeSH .
 Types of In House MeSH
• This are of two types :

 geographical headings

 provisional headings
The geographical headings
• are used to describe articles with
epidemiological, political,
sociological, or geographic
interest. For instance, consider
articles . Example:
• Public Health in China.
Provisional headings
• Are the terms available for
indexing and machine searching
which have not yet been
approved as main headings for
inclusion in Index Medicus.
• To describe a drug under the
heading ACTINOSPECTCIN -
The term ANTIBIOTICS can also
be assigned in the Index
Medicus.
 Index Medicus
• Library's most widely known MEDLARS
publication.
• Approximately 18,000 articles are cited in
each monthly issue. under about 8,000 MeSH
headings.
• Each issue contains an author index.
• Cumulated Index Medicus is published yearly.
 Various subsets of Index Medicus
• Abridged Index Medicus
It is a monthly listing of references from 100 key English-
language journals in clinical medicine.

• Monthly Bibliography of Medical Reviews

This publication contains references to those journalarticles
which are well documented surveys ofrecent literature.
INDEXING OPERATION
 Essential Requirement for indexers

• Most analysts are able to read at least one

foreign tongue.
• Many were able to index literature in four or
five languages.
• One can index efficiently without the perusal
of every single word of a text.
Analysts use a read/scan method based on the
following instructions
a) Read and understand the title.
b) Read the text down to the point at which the author states
the purpose of his paper.
c) Read every word of the summary.
d) Closely scan the abstract.
e) Scan the bibliographic references.

 
 Check- Tags
• A set of routine items that must be accounted for in
the indexing of every article by the indexers.
• These items are known as check-tags.
• eg: Studies describing controlled clinical research on human beings
are identified by the check-tag clinical research ,
while,
studies of the comparative effects of two or more drugs, or two or more
procedures or techniques, are identified by the tag comparative study.
 Coordinate Indexing
• In MEDLARS complex concepts are expressed by
combinations or coordination’s of two or more terms.
Example :

• An article discussing "community health services

under Regional Medical Programs" would be indexed
with the main headings –
1. COMMUNITY HEALTH SERVICES and
2. REGIONAL MEDICAL PROGRAMS,
• both to be printed in Index Medicus.
 Indexing Revision
• Conducted by a fully experienced indexer.
• known as a"reviser.“
• The reviser rapidly scans the bibliographic description and then
concentrates on the assigned subject headings,

• by following questions:
• Do the main headings reflect the true content of the
article?
• Do the Index Medicus ( print ) headings cover the
central points of the article?
• Are the headings spelled correctly, in exactly the form
appearing in MeSH ?
• Are the correct subheadings used?
• Are the relationships expressed correct?
 The Analyst's Reference Tools

• Best authority for his indexing

of an article is the text of the
article itself..
• The Integrated Authority File
EXAMPLE OF THE INDEXING PROCEDURE
• Sample article is entitled :--
"Positive Sputum Cytologic Tests for Five Specific Detection of
Bronchial Carcinoma”
The indexer decides after reading this article this is primarily related to the diagnosis
of pulmonary and bronchial neoplasm’s by means of cytological tests of the sputum.
Heading/subheading combinations:-
 X SPUTUM*cytology
 X CARCINOMA, BRONCHOGENI C *diagnosis
 X BRONCHIAL NEOPLASMS*diagnosis
 X LUNG NEOPLASMS –diagnosis.
• (Indicating by the X that these are print terms, representing headings
under which the citation should be printed in Index Medicus. )
 The additional terms

 BRONCHIAL NEOPLASMS* pathology

 CARCINOMA, EPIDERMOID*pathology

 CARCINOMA, BRONCHOGENI C*pathology

 ADENOCARCINOMA*pathology

 CYTO DIAGNOSIS
 PROCESSING THE CITATIONS

 The magnetic tapes produced by these input

procedures ,in the initial stages.
 For each citation the following data are
recorded:
1. citation number
2. Author
3. title of article
 CONDUCTING THE DEMAND SEARCH

• Retrieval in MEDLARS, as in any other

information system, is essentially a matching
process.
• A request for literature on a particular topic is
matched against the file of citations to
biomedical articles that have been input to the
system. When a match occurs, an article is
retrieved.
 Request Analysis

• A search analyst goes through the same two-stage

process:
1. The analyst must decide what the request is about

2. The analyst must translate his interpretation of the

request into a search statement, in MeSH terms,
that can be processed against the citation file.
 
• In the analysis of a request, it is logical to begin by breaking the request into its various
aspects, or facets.
• Example,
• Consider a search for literature on the subject
• Renal Amyloidosis.
• This request has two facets:
• (1) the organ facet (kidney) and
• (2) the disease facet (amyloidosis).
• The requester is not interested in all articles on the kidney, and he is not interested in all
articles on amyloidosis.
• He is interested only in articles that discuss both facets of his request, i.e., kidney and
Amyloidosis,
• and which presumably deal with renal amyloidosis.
• This relationship between the two facets may conveniently be represented by a
diagramof overlapping circles, as
This relationship between the two facets may conveniently be represented by a
diagram of overlapping circles, as
Renal Amyloidosis

Kidney Amyloidosis
 A Simple Search Formulation
• This is a very simple search formulation for a request on renal
amyloidosis.
• By giving each term in this strategy a unique identifying number, and
by using symbols that are recognized by the MEDLARS computer,
we can reduce this strategy to a simple algebraic search equation.
• Considering the following :
• M1 KIDNEY
• M2 KIDNEY GLOMERULUS Related to Kidney
• M3 KIDNEY PELVIS
• M4 KIDNEY TUBULES
• M8 AMYLOID 
• M9 AMYLOIDOSIS Kidney Related Dieseses
 Contd’….
• Here it shows that M1 to M4 all the terms are
related to Kidney and
• M8 and M9 is related to Amioloidosis
 Developing the Lists of Search Terms

• The most specific available term is always used

to describe a concept.
• Thus,
an article on sunburn is indexed under
SUNBURN and one on eye burns under EYE
BURNS rather than under the more generic
term BURNS.
 Cross-references
• The cross-references in MeSH lead the searcher to many of those headings
which might be considered for searching.
• for example : burns (nonspecific) in children
• We find four terms:
1. BURNS
2. BURNS, CHEMICAL
3. BURNS, ELECTRIC
4. BURNS, INHALATION

 The great advantage of the tree structure is that it minimizes the

time needed by the search analyst to conduct comprehensive generic
searches.
 
 The Machine Search and Its Output
• Several searches are batched together and processed
simultaneously,
• The search formulation is matched sequentially
against the complete data base of indexed citations.
• A citation is retrieved from this data base when its
index terms match one of the combinations of index
terms demanded by the search strategy.
• Qualifying citations are copied from the main
citation file onto a retrieved citation tape.
 Web Versions
• Around mid-1990s, the development of MEDLINE took another new
dimension.
• Several web entrepreneurs such as HealthGate etc. purchased MEDLINE
data and mounted it on their site with a search engine for free use for its
visitors.
• The underlying aim of this move was to make the site popular.
• Among the different databases available, it was the obvious choice
because of its low cost and wide popularity
• free MEDLINE sites that are available on the Web are,
• - HealthGate (http://www.healthgate.com); N4eds.cape
(http://www.medscape.com); Biomednet (http://www.biomednet.com).
Health Gate Website
Medscape
Biomednet
 It uses a simple search interface. The search
features are
1. It uses a simple search interface. The search features are
a) Basic search - a simple keyword search.
b) Advanced search - allows use of Boolean operators and other search
aspects and limit options .
2. Searches can be made using different fields like, author, text word in title,
subject term, etc
3. MeSH terms can be selected with the help of MeSH browser
4. Mapping terms to MeSH is limited
5. Searches can be limited to clinical queries
6. Search strategy is automatically displayed
7. Search results can be marked and printed or saved
8. Default display of 20 citations per page
 Paid MEDLINE SERVICE

• Knight Ridder (now Thompson-Dialog),

• Silver Platter,
• Ovid, and
• STN.
• Thus a MEDLINE plethora of online options are
available; most web versions of MEDLINE have
introduced links to full-text items wherever
available.
 MEDLINE Today
• Scope and Coverage
1. 40 online databases
2. in the field of bio-medical sciences with more
than 20 miiiion records.
3. It is estimated that about 20000 citations are
added.
 OVID Technologies (Now available from
Wolters Kluwer)
• Ovid is a search software designed to search more than 80 science,
medical, technological databases including MEDLINE, Current contents,
BIOSIS, ABI/Inform, and Psycholnfo.
• It is described as fully integrated full text linked database.
• Ovid is available on stand-alone, Windows, Novell NetWare, WinNT, Unix,
via Internet .
• As per the assessment by DUKE UNIVERSITY MEDICAL CENTER LIBRARY,
Ovid search engine js_ an excel lent..interface to MEDLjNE .
 FACTORS AFFECTING PERFORMANCE IN
MEDLARS
• The success or failure of MEDLARS, or any other information
retrieval system, is dependent less on machine capabilities
than on intellectual effort.
• This intellectual effort is exerted by the indexers, the search
analysts, the vocabulary specialists and - very importantly - by
the user himself.
• The great majority of articles in the data base will not be
relevant.
• There are 999,980 non-relevant articles.
 Evaluation of operating efficiency of
MEDLARS
• conducted in 1966-67
• Purpose:
i. To study the demand search requirements of MEDLARS
users.
ii. To determine how effectively and efficiently and the present
MEDLARS services are meeting its requirements .
iii. To determine factors adversely effecting the performance.
iv. To discover ways in which more effectively or more
economically user requirements can be satisfied.
Prime requirements of demand search users

• The coverage of MEDLARS

• Its recall power.
• The precision power.
• The response time of the system.
• The format in which the search results are presented.
• The amount of effort the user must personally expend in
order to achieve a satisfactory response from the system.
• The evaluation program was conducted in order to meet user requirements
and tolerances in relation to these various factors and to determine
MEDLARS performance in regard to the user requirements.
 The two most critical problems faced in the evaluation
of MEDLARS were

1. Ensuring that the body of test request was as far as possible

representative of the complete spectrum of kinds of
requests processed.

2. Establishing methods for determining recall and precision

performance.
 Methodology

• Queries were submitted to the MEDLARS and on

receipt of the queries; MEDLARS staff prepared a
search formulation (i.e. query designation) for each
query using an appropriate combination of MeSH
terms. A computer search was then carried out in the
normal way. At this stage each user was also asked to
submit a list of recent articles that he judged to
berelevant to his query.
• Result of a search was a computer printout of
references. Since the total items retrieved might be
high (some searches retrieved more than 500
references), 25 to 30 items were selected randomly
from the list and photocopies of these were provided
to the searcher fro relevance assessment.
• Each searcher was asked to go through the full
text of the articles and then to report about
each article on the following scale of relevance:
• H1 – of major value (relevance)
• H2 – of minor value
• W1 – of no value
• W2 – of value not assessable (for example, in a
foreign language).
 EACH SEARCHER WAS ASKED TO GO THROUGH THE FULL TEXT OF THE
ARTICLES AND THEN TO REPORT ABOUT EACH ARTICLE ON THE FOLLOWING
SCALE OF RELEVANCE

• If L items were in the sample, the overall precision ratio was 100(H1
+ H2)/L and the ‘major value’ precision ratio were 100H1/L. It was
obviously not feasible to examine the whole MEDLARS database in
relation to each search in order to establish recall ratio. Therefore, an
indirect method was adopted for calculation of recall ratio: Each user
was asked to identify relevant items for his query before receiving
the search output and then search was carried out to find out
whether those items were indexed in the database and retrieved
along with other items that are both relevant and irrelevant. If t such
relevant items were identified by the user and available on the
database for a given query, and H were retrieved in the search, the
overall recall ratio and ‘major value’ recall ratio was estimated as
100H/t and 100H1/t1respectively.
• The next stage of the evaluation was an elaborate
analysis of retrieval failures,
i.e., examining, for each search, collected data
concerning failures include:
1. Query statement
2. Search formulation;
3. Index entries for a sample of ‘missed’ items
(i.e.relevant items that are not retrieved) and
‘waste’ items (i.e. noise—retrieval of irrelevant
items);and
4. Full text (c).
 RESULTS
• The average number of references retrieved for each search was 175,

• with an average or overall precision ratio of 50.4%; that is, of the average
175 references retrieved,
• about 87 were found to be not relevant.
• The overall recall ratio was 57.7% as calculated by an indirect method.
Taking the average search, and assuming that about 88 of the references
found were relevant, with an overall recall ratio of 57.7% implies that
about 150 references should have been found, but 62 were missed.
• However, the recall and precision ratios for each of the 302 searches were
analyzed and individual ratios were then averaged in the MEDLARS test.
• The results were:
 The results were
over all Major value
Recall ratio Recall ratio
• 57.7% 65.2%

• Precision ratio Precision ratio

• 50.4% 25.7%
 CONCLUSION

MEDLINE functions as an important resource for biomedical

researchers and journal clubs from all over the world. Along
with the Cochrane Library and a number of other databases,
MEDLINE facilitates evidence-based medicine. Most
systematic review articles published presently build on
extensive searches of MEDLINE to identify articles that might
be useful in the review MEDLINE influences researchers in
their choice of journals in which to publish.
 Bibliography
1. LANCASTER, F. W. (n.d.). MEDLARS: Report on the Evaluation of Its Operating Efficiency.
2. MEDLINE on CD-ROM: An Analysis of User Search Behaviour.
3. ATWOOD, R. (n.d.). A Grass-Roots Look at MEDLARS.
4. CHARLES J. AUSTIN, H. (n.d.). Data Processing Aspects of MEDLARS. National Library of
Medicine .
5. (1968). EVALUATION OF MEDLARS DEMAND SEARCH SERVICE. U.S. DEPARTMENT OF
HEALTH, EDUCATION, AND WELFARE.
6. Lancaster, F. W. (n.d.). EVALUATION OF ON-LINE SEARCHING IN MEDLARS.
7. LANCASTER, F. W. (n.d.). MEDLARS: Report on the Evaluation of Its Operating Efficiency.
8. MEDLINE on CD-ROM: An Analysis of User Search Behaviour.
9. NATIONAL LIBRARY MEDICINE. THE PRINCIPLES OF MEDLARS. NATIONAL LIBRARY
MEDICINE.
10. ROGERS, F. B. (n.d.). The Development of MEDLARS.
11. Swargiary, D. (n.d.). Assignment on MEDLARS, Pondicherry University.
THANK YOU