CHAPTER 4			INDEXING OPERATION 
 

4.1		DEFINITION 
 
The indexing or subject heading operation is the process of assigning to an  
article the headings from MEDICAL SUBJECT HEADINGS (MeSH), the MEDLARS  
authoritative vocabulary, which best describe the content and substance as  
written by the author. 
 

4.2		COORDINATE INDEXING 
 
MEDLARS uses a system of coordinate indexing.  In coordinate indexing, the  
concepts in the text of articles are expressed by the combination or  
coordination of two or more indexing terms, some destined for publication in  
INDEX MEDICUS and others destined for storage within the computer for future  
retrieval. 
 
The indexing of terms destined for INDEX MEDICUS (IM) will reflect the point of  
the article, while the computer store will garner either their coordinates  
(terms which limit or describe or qualify the IM terms) or significant items  
discussed by the author but which are not necessarily the major point.  These  
latter headings are termed NON-INDEX MEDICUS or NIM headings. 
 
Coordination in MEDLARS is achieved by 
 
-	two or more main headings 
-	a main heading and a subheading 
-	a main heading and a check tag (a pre-determined concept routinely checked  
in every article) 
-	a pre-coordinated main heading, i.e., one created from a past frequent co- 
occurrence of two concepts 
 
Any article can be indexed by any type of coordination or by a combination of  
any two or more of the above types.  See Figure 4.2 for an illustration of the  
types using hypothetical titles of articles. 
 
The elements of coordination requiring most explanation are the Check Tags  
(Chapter 18) and the Subheadings (Chapter 19).  Many of the following pages will  
be devoted to explanations of these essential coordinations. 
 
				TYPES OF COORDINATION IN MEDLARS 
 
Two or more main headings 
 
	Statistics on shock therapy:		 
		CONVULSIVE THERAPY (IM) 
		STATISTICS (NIM) 
 
	Design of intensive care units in psychiatric hospitals: 
		HOSPITALS, PSYCHIATRIC (IM) 
		INTENSIVE CARE UNITS (IM) 
		HOSPITAL DESIGN AND CONSTRUCTION (IM) 
 
A main heading and a subheading 
 
	Metabolism of the liver:	 
		LIVER / metabolism 
 
	Liver pathology in hepatitis:	 
		LIVER / pathology (IM) 
		HEPATITIS / pathology (IM) 
 
A main heading and a check tag 
 
	Gastritis in children:		 
		GASTRITIS (IM) 
		CHILD (NIM) 
		HUMAN (NIM) 
 
	Experimental arthritis in dogs:	 
		ARTHRITIS (IM) 
		DOGS (NIM) 
		ANIMAL (NIM) 
 
A pre-coordinated main heading 
 
		LIVER GLYCOGEN		from	LIVER + GLYCOGEN 
		BREAST NEOPLASMS	from	BREAST + NEOPLASMS 
		HEPATITIS, ANIMAL	from	HEPATITIS + ANIMAL 
		TUBERCULOSIS, PULMONARY	from	TUBERCULOSIS + LUNG 
 
					Figure 4.2 
 
 

4.3		CONTENT ANALYSIS 
 
What are the major points of the article?  That is, 
What concepts in the article are worthy of being called to the attention of the  
INDEX MEDICUS user? 
 
What are the minor points of the article?  That is, 
What concepts would better be stored for future use although they are not the  
major point? 
 
What is discussed as opposed to what is "merely mentioned"? 
 
What is the slant of the article with regard to the specialty of the journal,  
and how does it relate to established indexing principles; i.e., experimental?   
clinical?  pediatric?  geriatric?  obstetric? veterinary?  microbiological?   
endocrinological?  etc. 
 
What is the relative status of the general aspects versus the specific? 
 
Are the conclusions based on the text or on suppositions or speculations not  
discussed? 
 
Are the concepts picked up for indexing based on the substance of the text and  
not on the indexer's subjective or personal evaluations? 
 

4.4		ARTICLE EXAMINATION 
 
The steps outlined below should be followed in sequence by logic and should be  
followed routinely for every article to be indexed: 
 
Read carefully and understand the title. 
 
Read the introduction down to the point where the author states the purpose of  
his article and correlate it with the title.  Absorb but do not necessarily  
attempt to index the introductory material since this is usually a statement of  
known facts upon which his present study is based. 
 
Note chapter headings, section headings, paragraph headings; italics, boldface;  
charts, plates, tables, illustrations; laboratory methods, case reports, etc.   
Headings supplied by the author usually herald the content of the section  
headed. 
 
Select for indexing only those subjects actually discussed (and therefore worthy  
of retrieval as INDEX MEDICUS citations or in on-line searches) as opposed to  
those subjects merely mentioned (and of little or no value in retrieval). 
 
Read the summary or conclusions of the author to determine whether he achieved  
the aims set forth in his stated purpose.  Weigh conclusions based on the text  
but do not index implications or suggested future applications.  Do not index  
conclusive statements not supported by discussion in the text. 
 
Scan the bibliographic references supplied by the author for clues and further  
corroborations. 
 
Scan the abstract, if there, for items missed in indexing, being careful,  
however, to locate actual discussions within the text of the article; ignore  
mere implications. 
 
Scan the author's own indexing if supplied or the keywords supplied by the  
publisher to see whether the concepts chosen are actually discussed in the text  
and already picked up by the indexer. 
 

4.5		QUALITIES OF THE INDEXING OPERATION 
 
In assigning subject headings after following the routine outlined above, the  
conclusions reached by the indexer should be characterized at all times by these  
qualities to follow.  They represent the basic philosophy of indexing. 
 

4.5.1	Conformity: 
 
The indexer must adhere to the content of MEDICAL SUBJECT HEADINGS, the MEDLARS  
thesaurus, in analyzing the literature, and to its directions in the form of the  
cross-references provided.  He must follow the hierarchy of these headings (the  
Trees) and the policies governing their use. 
 

4.5.2	Consistency: 
 
The indexer must apply consistently the practices he follows in conforming to  
MeSH and to rules of indexing.  He must consistently use headings the same way  
each time so that a searcher may rely on his practices and even predict their  
application.  Only in this way can the searcher expect to find the same concepts  
in the same place with reasonable predictability. 
 

4.5.3	Impartiality: 
 
The indexer cannot predict what aspect of an article will be of interest to what  
researcher.  He must set aside personal evaluations and opinion and must  
approach all subjects in all their aspects impartially and without prejudice. 
 

4.5.4	Specificity: 
 
MeSH provides both general concepts (like PHYSIOLOGY) and specific concepts  
(like LIVER REGENERATION).  The indexer is committed to the greatest possible  
specificity:  where the subject of an article is specifically on LEUKOCYTES, it  
is wrong to assign to the article a general heading like BLOOD or BLOOD CELLS.   
If an article entitled "Parasiticides in gynecology" is in fact on the treatment  
of trichomonal vaginitis with metronidazole, the main headings will be  
TRICHOMONAS VAGINITIS and METRONIDAZOLE and not ANTIPROTOZOAL AGENTS and  
GYNECOLOGY. 
 

4.5.5	Multiplicity: 
 
Since MEDLARS is used by many types of research worker and clinician, the  
indexer will provide for each article as many subject headings as he deems  
necessary to index the article adequately from all its aspects.  This is called  
multiple-indexing.  The specialty approach of the journal being indexed does not  
restrict the indexer from indexing other aspects in each article. 
 

4.5.6	Veracity: 
 
The headings chosen by the indexer must truly reflect the content of the  
article.  The searcher will go unhesitatingly to an article because the headings  
selected assure him that he will find there information on the subject concerned  
with the heading used.  In other words, the use of the heading DURA MATER  
implies that in the article cited is noteworthy information on the dura mater. 
 
 

4.5.7	Judgment: 
 
The indexer is obligated to disregard irrelevant or impertinent data without  
sacrificing impartiality, specificity, multiplicity or veracity.  He must draw  
upon his familiarity with the subject field, the journal in which the article is  
printed, its quality, its editorial policy, peculiar language or national  
practices -- any feature of his knowledge and experience with the subject,  
field, or journal in hand.  Especially noteworthy must be his judgment in  
differentiating worthy discussion from mere mention in passing.  He must judge  
objectively the substantive amount of text worthy of retrieval; he must refrain  
from an evaluative judgment of the author's work. 
 

4.6		SELECTION OF ARTICLES 
 
In selecting articles for indexing, the following instructions will govern the  
Index Medicus scope and coverage practices: 
 

4.6.1	Index all original articles on medical and biomedical subjects  
regardless of form or length. 
 
Index any original article on any subject -- medical or non-medical -- written  
by an employee of the United States Public Health Service or the National  
Institutes of Health.  Index also any article -- in scope or not -- carrying a  
statement that the research was supported by PHS or NIH.  See also 18.15.9. 
 

4.6.2	Index all substantive biographies and obituaries (see sections 32.16  
and 32.17). 
 

4.6.3	Index all clinico-pathological conferences (see sections 18.11.8  
through 18.11.15). 
 

4.6.4	Index reports of scientific material presented at congresses and  
conferences, of acceptable substance.  If a paragraph, or at least several  
sentences are devoted to each speaker, the report is substantive enough to be  
indexed. 
 
Do not index reports of congresses and conferences concerning committee matters,  
organizational details, financial reports, etc. 
 
Do not index articles in which an author summarizes and comments on what took  
place at a congress or meeting if the summary includes no more than the names of  
the speakers and the titles or topics discussed. 
 

4.6.5	Index so-called "editorials" that are in fact lead articles or  
special articles on substantive material, generally supplied with a bibliography  
(see section 14.5.1.4). 
 
Do not index editorials that resemble the newspaper concept of editorial with  
possibly only local interest.  Do not index "editorializing" editorials. 
 

4.6.6	Index letters to the editor that are substantive and that are  
documented and that in fact constitute "little articles" (see section 14.5.1.5). 
 

4.6.7	Index all reports of Councils of the American Medical Association  
and of all national councils or commissions. 
 

4.6.8	Index all directories and membership lists (see section 33.13). 
 

4.6.9	Index all brief articles and short communications in such journals  
as Nature, Science, and the like, that are on medical and biomedical subjects,  
despite their brevity. 
 

4.6.10	Index individual items from "Questions and Answers" sections if  
substantive.   
 
Do not index any Question and Answer that is less than one-quarter page long. 
 
Items one-quarter to one-half page long may be selected if at least two  
references are cited. 
 
Items with no references given may be selected only if they are at least one- 
half page long. 
 

4.6.11	Index presidential addresses if they are substantive. 
 

4.6.12	Index society proceedings within a given issue if they are of  
substance and with no great variation in length within the issue.  
 
Do not index proceedings that are little more than expanded abstracts.  Such  
proceedings are indexed in a block with an "overall". 
 

4.6.13	Index formal discussions of articles.  A formal discussion is one  
with a title more distinctive than the designation "Discussion" and with authors  
indicated formally.  This type is listed as a separate item in the table of  
contents of the issue. 
 
Do not index an informal discussion.  An informal discussion is one without a  
distinctive title, identified by only the word "Discussion".  Include this  
discussion within the pagination of the article to which it is attached.  See  
10.8. 
 
Do not index as a separate article a discussion, formal or informal, that  
follows the last of a series of articles and discusses them all jointly.  See  
10.8. 
 

4.6.14	Index translations of articles if the original appeared in a journal  
not in the List of Journals Indexed in Index Medicus. 
 

4.6.15	Index reprints only if the original appeared in a journal not in the  
List of Journals Indexed in Index Medicus. 
 

4.6.16	Index reprints of classic articles or other classic landmarks in the  
history of medicine (see sections 14.5.1.10, 17.20 and 32.9). 
 

4.6.17	Do not index non-medical articles in journals indexed selectively  
(these are indicated by an "s" -- for selective -- in the List of Journals  
Indexed in Index Medicus). 
 

4.6.18	Do not index signed or unsigned presentations that appear to be  
nothing more than a recapitulation of another article previously indexed for  
Index Medicus when it appeared originally.  Such articles are brief and  
recognizable by a statement of the reference or references early in the text.   
This is usually verified at the end by a full bibliographic reference. 
 

4.6.19	Do not index such regular features as "X-ray of the month", "What is  
your diagnosis?", "The ECG of the month", "Dermatological iconography" if the  
presentation is merely an illustration accompanied by little more than a brief  
reply. 
 
Index such features only if the explanation or analysis constitutes a "little  
article" with substantial accompanying text and references. 
 

4.6.20	Do not index abstracts, condensations or summaries of books, of  
articles or of proceedings and presentations at congresses, symposia or other  
meetings. 
 

4.6.21	Do not index book reviews.  "Essay reviews" in history of medicine  
journals, however, are in scope. 
 

4.6.22	Do not index current bibliographies printed in a journal as a  
regular feature under a monthly rubric if they are derived from a MEDLINE  
database.  Regularly occurring bibliographies may be selected and indexed if  
they are not explicitly defined as being derived from MEDLINE. 
 

4.6.23	Do not index monthly epidemiological reports. 
 

4.6.24	Do not index hospital reports. 
 

4.6.25	Do not index business or organizational or administrative reports of  
institutions. 
 

4.6.26	Do not index medical society news or news notes. 
 

4.6.27	Do not index government news releases or news reports of state or  
national activities published as regular features of supplementary matter within  
an issue. 
 

4.6.28	Do not index announcements of new drugs by pharmaceutical companies. 
 

4.6.29	Do not index poster papers if they are little more than expanded  
abstracts. 
 

4.6.30	Do not index advertisements even if they contain substantial text.