Response to Schoonbaert

The writer points out some useful considerations for using the PubMed “Publication Date” field in bibliometric studies involving post-2000 publications, calling for careful construction of search queries to identify and exclude instances of duplicate counts across year boundaries in order to minimize the impact. The data that we reported were not affected by these factors. In our study, the date range search feature was used to obtain the aggregated volume of publication for each country over the years 1995 to 2004. Because PubMed does not report duplicate records in a retrieval set, the fact that a paper can indeed carry two different publication dates, if e-publication and print publication years are different, did not affect the counts.

The writer also notes different results when using the limit feature versus entering a publication date directly into the search box. This is documented in the PubMed Help, which explains that the electronic publication date is not searchable if it is later than the print date, except when range searching. The limit feature in fact employs a date range search, which will include the relatively few records for which the e-publication date is later than the print publication date, the situation referred to in the letter as “the reverse also occurs, but apparently to a far lesser extent.” To double-check our data, we reran the searches using the range feature versus including each year individually in the search strategy, and the results were the same for both methods.