The Medical Subject Headings (MeSH®) thesaurus is a controlled vocabulary produced by the National Library of Medicine and used for indexing, cataloging, and searching for biomedical and health-related information and documents.
MeSH includes the subject descriptors appearing in MEDLINE®/PubMed®, the NLM catalog database, and other NLM databases.
Many synonyms, near-synonyms, and closely related concepts are included as entry terms to help users find the most relevant MeSH descriptor for the concept they are seeking. In NLM's online databases, many terms entered by searchers are automatically mapped to MeSH descriptors to facilitate retrieval of relevant information.
Various online systems provide access to MeSH and the vocabulary is available in several online systems. These include the MeSH Browser, containing the complete contents of the vocabulary; the MeSH Entrez databases, which are designed to assist those searching MEDLINE®/PubMed®; and the UMLS Metathesaurus® with links to many other controlled vocabularies. Additional information about MeSH and direct access to MeSH data is provided on the Web at //www.nlm.nih.gov/mesh.
The first official list of subject headings published by the National Library of Medicine appeared in 1954 under the title Subject Heading Authority List. It was based on the internal authority list that had been used for publication of Current List of Medical Literature which in turn had incorporated headings from the Library's Index-Catalogue and from the 1940 Quarterly Cumulative Index Medicus Subject Headings. With the inception of Index Medicus (New Series) in 1960, a new and thoroughly revised Medical Subject Headings appeared.
With the 1954 Subject Heading Authority List, there appeared a "Categorical Listing" of standard subheadings. "Abnormalities," for instance, was listed as a standard subheading for use with terms for organs, tissues, and regions, and "anesthesia and analgesia" was to be used under surgical procedure headings. But such subheadings could be used only for subject headings which fell within the category of headings to which they were to be applied. There were over 100 such subheadings, some of which varied only slightly according to the category of main heading with which they were used. For instance, "therapeutic use" was used under physical agents and drugs and chemicals, and "therapy" was used with diseases. In the 1960 Medical Subject Headings, the number of subheadings was reduced to sixty-seven. They could be used under any kind of main heading if the combination was not patently foolish or impossible. These sixty-seven subheadings were applied with more generalized meanings. For instance, the subheading "therapy" was used to mean "therapy of," "therapeutic use of" or just "therapeutic aspects." Though this solution was simpler, many problems still remained. The use of one subheading might prevent the use of another. For instance, if a paper covered the etiology, pathology, and therapy of a disease, it might occur without further subdivision, or it might occur under the subheading which seemed most appropriate to the indexer. If "therapy" was chosen, the article would be lost to the searcher looking for the etiology of the disease under the subheading "etiology." In addition, if the subheading "diseases" had been appended to the term for an anatomic part, it would not be possible to subdivide further for the therapy or complications of such diseases. A related problem was the overlap in meaning of the subheadings themselves. It was difficult, for example, to decide whether a paper on chemical biosynthesis fit best under "chemistry" or "metabolism."
Categorized lists of terms were printed for the first time in the 1963 Medical Subject Headings and contained thirteen main categories and a total of fifty-eight separate groups in subcategories and main categories. These categorized lists made it possible for the user to find many more related terms than were in the former cross-reference structure. In 1963, the second edition of Medical Subject Headings contained 5,700 descriptors, compared with 4,400 in the 1960 edition. Of the headings used in the 1960 list, 113 were withdrawn in favor of newer terms. In contrast, the 2015 edition of MeSH contains 27,455 descriptors.
In 1960, medical librarianship was on the cusp of a revolution. The first issue of the new Index Medicus series was published. On the horizon was a computerization project undertaken by the National Library of Medicine (NLM) to store and retrieve information. The Medical Literature Analysis and Retrieval System (MEDLARS®) would speed the publication process for bibliographies such as Index Medicus, facilitate the expansion of coverage of the literature, and permit searches for individuals upon demand. The new list of subject headings introduced in 1960 was the underpinning of the analysis and retrieval operation. MeSH was a new and thoroughly revised version of lists of subject headings compiled by NLM for its bibliographies and cataloging. Frank B. Rogers, then NLM director, announced several innovations as he introduced MeSH in 1960.
The adoption of a single subject authority list for both books and periodical articles is a departure from traditional practice. We take the view that subject cataloging and periodical indexing, as exemplified in the Index Medicus and in the NLM Catalog, are identical processes in their major dimensions. A single list can and should be used for both purposes. This has two major virtues: simplicity for users, in requiring familiarity with only a single scheme; and economy to the Library in the development and maintenance of a single scheme.
There is another departure from traditional practice represented in this list. This is the adoption of standard topical subheadings for cataloging books, as well as for indexing periodical articles. The topical subheading is in effect a substitute for a phrase heading, and on the whole it is a preferable substitute.
The main heading-topical subheading combination is a pre-coordination of terms, reducing the problem of term permutation, which looms large in most manual retrieval systems in book form.
From its beginning, MeSH was intended to be a dynamic list, with procedures for recommending and examining the need for new headings. The content of the vocabulary related to the usage of terms in the literature itself and evolved to meet new concepts in the field. The use of the computer made revisions more practical and systematic, despite the difficulty in updating printed indexes and card catalogs.
The following organizing principles, reprinted from the Medical Subject Headings, 1st edition, 1960, continue to be used in the creation and maintenance of the MeSH vocabulary. This text was prepared when the Index Medicus was the primary finding tool of the National Library of Medicine for article citations and was an annual printed work. It is now replaced by the various online databases.
Here is presented the subject heading authority list of the National Library of Medicine. These subject headings are used in the compilation of the new Index Medicus and in the National Library of Medicine Catalog, beginning 1960.
The list represents a combination, rationalization, and extensive modification of two previous authority lists. One was a list, maintained only as a card file, which was utilized in the subject cataloging of books. The other was the Subject Heading Authority List used in the old Current List of Medical Literature; this was published in 1954 and underwent several internal revisions over the course of the years.
The adoption of a single subject authority list for both books and periodical articles is a departure from traditional practice. The rationale of this position was set forth in "Applications and Limitations of Subject Headings: The Pure and Applied Sciences" a paper printed in the Subject Analysis of Library Materials, edited by Maurice F. Tauber, published by the School of Library Service, Columbia University, 1953. Much has been made of the presumed differences between headings used for cataloging and headings used for indexing; most of the difficulty lies in the ambiguity of the word "indexing". This has ordinarily thought of in terms of indexing a book. A book index is made up on an ad hoc basis; there is a brand new conceptual scheme evolved for every book indexed. It should be clear that the construction of continuing indexes in multiple periodicals is quite a different matter. We take the view that subject cataloging and periodical indexing, as exemplified in the Index Medicus and in the NLM Catalog, are identical processes in their major dimensions. A single list can and should be used for both purposes. This has two major virtues: simplicity for users, in requiring familiarity with only a single scheme; and economy to the Library in the development and maintenance of a single scheme.
Medical information "has been drawn from such a wide span of time and such a diversity of specialized fields that its doctrines belong to several different systems and its language problem is almost as bad as that of India. There is at least one major language for each major department, and each of these has several dialects. The situation is made even worse because in each language we teach a mixture of doctrines which range from Newtonian absolutism to Einsteinian relativism, including additive, reciprocal, exponential, and circular structures. It is tragic to contemplate the amount of effort we now waste because of our conflicting doctrines, and intriguing to wonder to what heights we might soar, each in his own way, once we manage to resolve the internal contradictions in the system by which we live and work."1 There will be less frustration on the part of librarians and other users of catalogs, indexes, and bibliographies if it is realized that the complexities of the field are such that simple, unequivocal solutions to the problem of the form and substance of medical subject headings are not easy to find.
From one point of view, subject headings may be looked at as an artificial language which bears only superficial resemblances to the natural language. Subject headings are more stilted, more stereotyped. From another point of view, in subject headings conceived of as pointers, rather than as labels, a certain amount of ambivalence is tolerable. Suppose, for example, we find a sign reading "San Francisco" pointing west on Independence Avenue in Washington, DC. This would not be very helpful. If, however, we see the same sign on the outskirts of Sacramento, it is likely to be very helpful indeed. And if we see the same sign on the approaches to the bridge outside of Oakland, we know we have arrived. It does not prejudice our case at all if some wise man comes up to us and says, "Exactly what do you mean by San Francisco? Do you mean the City of San Francisco, or the metropolitan area of San Francisco? Exactly where are the corporate city limits located?" A satisfactory answer might be that, having arrived, we will consult one of the local inhabitants if such distinctions seem important to us. On the other hand, we would like to be sure that the sign on the outskirts of Sacramento refers to the city in California rather than the city of the same name in Argentina. It is likely that in most cases we will be reassured on this point from the circumstances of the locale - the context - in which the sign is found. These headings are meant for use with biomedical literature and in biomedical libraries. They usually designate narrow and specific biomedical concepts. In peripheral fields headings used are more general in import: Automobiles appears in the list but not Hot-Rods, or Trucks. And when Vehicles appears, it may be assumed that in context this means pharmaceutical vehicles rather than vehicles of the four-wheeled kind Also, since medical libraries, are bibliothecal as well as medical, by definition, bibliothecal terms such as 'Subject Headings', may also be found in this list.
Form of Headings
General: There is a preference for headings using the direct rather than the inverted form; thus Mitral Valve, rather than Valve, Mitral. In a number of instances, however, inverted forms are used. This is particularly in evidence in those cases where it is possible, in a series of inverted headings, to provide built-in annotations by virtue of the fact that surrounding headings tend to define the scope of a particular headings. Thus,
Psychoses, AlcoholicPsychoses, InvolutionalPsychoses, Senil
are self-annotating as a group. Similar groups appear under Reflex, under Carcinoma, under Chemistry, and so forth. But there are always exceptions to the exceptions. Though we say Chemistry, Analytic nevertheless we say Biochemistry rather than Chemistry, Physiological. In most cases. Cross-references are provided.
American rather than British spelling is preferred: thus Anesthesia rather than Anaesthesia. Occasionally this leads to difficulties. For example, Amoeba is used rather than Ameba because that form of the name is preferred in taxonomical nomenclature' Nevertheless, when we name the disease, we say Amebiasis rather than Amoebiasis.
If a hundred books and articles appear on Krebiozen, that is what they are about, and they will be so entered in the catalog. On the other hand, we try to avoid burdening the list with hundreds of terms for special practices.
There are some terms in the foreign literature which defy translation into English; they designate concepts which are unknown in the English literature. In some cases we have found no solution except to anglicize the foreign term. Thus radiesthesia, and reflexotherapy, describe concepts appearing in the French and German literature, and in terms similar to the original French and German. In other instances such a practice would be extremely misleading. For instance, German literature has material on a concept best translated as respiratory therapy. we do not use such a designation because of the ambiguity it would hold for the English-speaking (and no doubt the Spanish-speaking and the Italian-speaking) community. The 'solutions' of such unusual cases are a good deal less than perfect, as witness our use of Breathing Exercises to cover the concept of Respiratory Therapy, since this seems to be the predominant element in this type of treatment.
We avoid eponyms whenever and wherever possible. That so many eponymics remain in the list is merely an indication that in a great many instances satisfactory substitutes are unavailable.
Choice of terms:
We have tried to use terms in the form and with the definitions set forth in standard reference works, when these can be made to agree. Thus, the names of bacterial genera are from the latest edition of Bergey's Manual, the names of neoplasms are the terms preferred in the American Cancer Society's Manual of Tumor Nomenclature; the names of enzyme groups are those promulgated by a Committee of the National Research Council; the names of drugs are the generic terms proposed by the American Medical Association's Council on Drugs: and so forth. Exceptions do occur; excepting oversights, or those cases in which it has been necessary to accept a term prior to its normalization in a standard reference work, exceptions have been made only after due deliberation.
1Swanson, M. J Bowman Gray Sch Med. 1959 Jun; 17:45-9.