A substantial number of scientific publications originate from non-university hospitals

Dan Med Bull. 2011 Nov;58(11):A4332.

Influence of study design on the citation patterns of Danish, medical research

Andersen, J. P., & Schneider, J. W. (2011). Influence of study design on the citation patterns of Danish, medical research. I Noyons, E., Ngulube, P., & Leta, J. (eds.), Proceedings of the 13th ISSI Conference (s. 46-51). Durban, South Africa.

Searching PubMed for studies on bacteremia, bloodstream infection, septicemia, or whatever the best term is: A note of caution

ePub

Background:
There is inconsistency in the terminology used to describe bacteremia. To demonstrate the impact on information retrieval, we compared the yield of articles from PubMed MEDLINE using the terms “bacteremia,” “bloodstream infection,” and “septicemia.”

Methods:
We searched for articles published between 1966 and 2009, and depicted the relationships among queries graphically. To examine the content of the retrieved articles, we extracted all Medical Subject Headings (MeSH) terms and compared topic similarity using a cosine measure.

Results:
The recovered articles differed greatly by term, and only 53 articles were captured by all terms. Of the articles retrieved by the “bacteremia” query, 21,438 (84.1%) were not captured when searching for “bloodstream infection” or “septicemia.” Likewise, only 2,243 of the 11,796 articles recovered by free-text query for “bloodstream infection” were retrieved by the “bacteremia” query (19%). Entering “bloodstream infection” as a phrase, 46.1% of the records overlapped with the “bacteremia” query. Similarity measures ranged from 0.52 to 0.78 and were lowest for “bloodstream infection” as a phrase compared with “septicemia.”

Conclusion:
Inconsistent terminology has a major impact on the yield of queries. Agreement on terminology should be sought and promoted by scientific journals. An immediate solution is to add “bloodstream infection” as entry term for bacteremia in the MeSH vocabulary.

Price revisited: on the growth of dissertations in eight research fields

epub, Scientometrics, Co-authored with Björn Hammarfelt

This paper studies the production of dissertations in eight research fields in the natural sciences, the social sciences and the humanities. In using doctoral dissertations it builds on De Solla Prices seminal study which used PhD dissertations as one of several indicators of scientific growth (Price, Little science, big science, 1963). Data from the ProQuest: Dissertations and Theses database covering the years 1950–2007 are used to depict historical trends, and the Gompertz function was used for analysing the data. A decline in the growth of dissertations can be seen in all fields in the mid-eighties and several fields show only a modest growth during the entire period. The growth profiles of specific disciplines could not be explained by traditional dichotomies such as pure/applied or soft/hard, but rather it seems that the age of the discipline appears to be an important factor. Thus, it is obvious that the growth of dissertations must be explained using several factors emerging both inside and outside academia. Consequently, we propose that the output of dissertations can be used as an indicator of growth, especially in fields like the humanities, where journal or article counts are less applicable.

Publikationsaktiviteten på Aalborg Sygehus [Publication activity at Aalborg Hospital]

Jens Peter Andersen, Conni Skrubbeltrang, Hans Gregersen. (2010). Publikationsaktiviteten på Aalborg Sygehus. Ugeskrift for Læger 172 (17) p. 1279-1284

Introduction: In 2003 Aalborg Hospital became part of Aarhus University Hospital and in that context focus on research activities was increased. This article investigates whether the increased focus has led to changes in the quantity and/or quality of research publications in the following period.

Material and methods: All scientific articles published by Aalborg Hospital in the period 2002-2008, as well as information about author affiliations comprise the data material for the analysis. Different levels of journal groups are created based on Journal Performance Indicators combined with peer-reviewing as a measure of publication quality, awarding publications in highly esteemed journals a higher score than those published in less recognized journals. Together with the number of publications, a measure of research quality and quantity is thereby achieved. This method is compared to the more traditional journal impact factor method.

Results: Data show an increase in total publications per year while the mean number of points per publication decreases during the period. Results also show a relation between the score level of publications and the number of collaborations for the publication, i.e. large collaborations are more frequently published in top journals.

Conclusion: The study shows that the increased focus on research has led to increased publication activity without loss of quality, as the decrease in points per publication is associated with the increased mean number of collaborators. The results indicate that the method would benefit from a revision to facilitate clearer conclusions.