Determining Student-Faculty Ratios and Faculty Scholarship Levels/Rates

To the Editor. The recent article by Benavides et al¹ explored the correlation of faculty members' publication productivity to student-faculty ratios in colleges of pharmacy. These investigators also evaluated the influence of other factors, such as research funding, public vs. private university status, and supportive faculty members on scholarship. While these areas are important topics of investigation where additional insight would be welcomed, I would like to point out 2 potentially serious methodological errors in the study. These authors attempted to compile publication rates of colleges of pharmacy by searching PubMed. While there is insufficient detail in the article about how these searches were performed, it appears that the authors utilized the affiliation field of the MEDLINE database to search for individual schools and colleges. Unfortunately, the problem with this approach is that the MEDLINE database lists only the address of the corresponding author, not all the authors of the paper. So in a multi-university collaborative paper with, for example, 6 authors, only the corresponding author's address will appear in the MEDLINE record. This is in contrast to a database such as Science Citation Index (Web of Science online) which captures the address of every author on a particular paper. This error would result in a significantly underestimated publication count for some colleges.

The second related error is that many authors do not list “college of pharmacy” or “school or pharmacy” in their addresses. If this were part of the search strategy, it also would contribute to a significant underestimation of the true publication count. As an example, a 2007 calendar year PubMed search for just 3 individual basic science faculty members (KM Giacomini, A Sali, and BK Shoichet), all of whom are listed in the AACP Roster for the 2006-2007 academic year from the University of California at San Francisco (UCSF), yielded 43 non-overlapping publications. Since bibliometric author searches on any database can be contaminated by homologues^2,3 (ie, authors with the same name who are in different disciplines or institutions), I was able to validate 33 of these papers by searching the Web site of these faculty members at UCSF. The Benavides study lists only 24 total publications for UCSF for this timeframe. A careful inspection of the MEDLINE records for these 3 faculty members reveals that many list “Department of Biopharmaceutical Sciences” at UCSF in the affiliation field without mentioning school of pharmacy. As a result of these possible errors, this study may have seriously underestimated the publication output of some colleges and schools of pharmacy.

Dennis F. Thompson, PharmD
Southwestern Oklahoma State University