The importance of keywords in scientific articles in Pharmaceutical Sciences, submitted in Estudo Geral: a comparative study of terms assigned in MEDLINE

• Autor(a) principal: Miguéis, Ana
• Data de Publicação: 2013
• Outros Autores: Neves, Bruno, Silva, Ana Luísa, Trindade, Álvaro, Bernardes, José Augusto
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: InCID
• Texto Completo: https://www.revistas.usp.br/incid/article/view/69284
• topic: Recuperação da informação; Linguagens de indexação; Palavras-chave; Artigos científicos; Medical Subject Headings; MEDLINE; Information retrieval; Indexing languages; Keywords; Scientific papers; Medical Subject Headings; MEDLINE

Objectives: To analyze the keywords used by authors at the University of Coimbra in papers published in the field of the Pharmaceutical Sciences, comparing them to Medical Subject Headings (MeSH) through the use of content analysis. Keywords used in scientific papers show the scope of a subject and its main concepts and are useful in indexing search mechanisms and categorizing texts. Investigation into the importance and characteristics of keywords has focused on various aspects such as efficiency in retrieving information, automatic extraction using different methods and algorithms; the use of keywords by authors and editors; and more recently, their use in metatags and comparison with titles, abstracts and full texts, confirms the interest in the topic. Few papers, however, were found in the literature comparing keywords provided by authors of scientific papers and the descriptors used by indexers, in order to analyze the authors’ contribution in choosing indexing terms.Method: After an exploratory bibliographic review in specialized literature, the sample was composed of 182 scientific papers produced by the community of the School of Pharmacy, University of Coimbra, deposited in Estudo Geral and edited between 1996 and 2012 (up to December 2012) and eventually published in international journals. Access to these papers, stored in open-access repositories, was accomplished by analyzing metadata and the content of full texts using techniques that represent and expose semantic relationships among documents, under which data are related. This paper’s specific objectives were: (i) to identify and measure the number of keywords used in the versions deposited in Estudo Geral and in the versions published by editors; and the number of descriptors assigned by MEDLINE; and (ii) verify the keywords used by the authors that coincide with or relate to the MeSH terms. Some conditions were established prior to data collection, namely to first compare the keywords used in the version published by the journal’s editor first identifying the names of the periodicals to which they belonged and then confirm the number the keywords in both resources. Afterwards, having all the keywords in each of the papers together, we compared the MeSH descriptors assigned to the titles indexed in the MEDLINE database.  The keywords collected from the metadata of the records of papers deposited in the repository and the editor’s final versions were organized for each paper in a grid together with the MEDLINE’s terms. Blank spaces for title, author and year of publication were also added. When quantitatively determining the relationships identified among the key words and MeSH descriptors, we considered equivalence (identical) and associative (related) relationships.Results and discussion: The results confirm that the keywords collected from the metadata of the papers deposited in the Estudo Geral do not correspond to all the keywords present in the published versions, which determined that the analysis of results focused only on the keywords presented in the editors’ versions, comparing them with the MeSH descriptors assigned to the titles indexed in the MEDLINE database. This comparison demonstrated that the number of keywords is significantly lower than the number of terms used by MEDLINE. In regard to the number of keywords per paper, most papers do not present more than five (5) key words, while the number of MEDLINE descriptors is much higher. From a semantic point of view, regarding equivalence and association relationships among the keywords and MEDLINE descriptors, about half of the keywords presented equivalence or associative relationships in similar proportions. Additionally, the relationships established by the keywords mostly present from 1 to 2 equivalent concepts and from 1 to 2 associate concepts.Conclusions: The results show that keywords are an advantageous source of access to scientific papers, highlighting the relevance of keywords chosen by authors to represent information provided by papers indexed in the MEDLINE database. Therefore, the number of equivalent and associative relationships in the 182 papers analyzed is close and even though the percentages are not very high, they enable us to conclude that authors directly or indirectly influence the choice of descriptors. Such a result may encourage the use of keywords to teach indexing per subject and improve efforts toward the automation of the indexing process. The use of Estudo Geral to support this study is intended to call attention to the institutional repository at the University of Coimbra and to the importance of authors as active actors in the process of the representation and retrieval of information through the use of keywords they choose for their papers. Acknowledging the importance of keywords in the representation and retrieval of information is a way to promote self-depositing and to encourage the completion of metadata of the documents deposited; this would contribute to greater visibility, an increased number of citations, and greater impact on the part of scientific production by the institution’s authors and the institution itself.