Expert Research Tip: Go back...

What does the 2001 death of a Johns Hopkins research participant have to do with understanding best research practices?  Everything, according to reports of the tragedy that were published in The Lancet in 2001.  In the article (1), author Sarah Ramasay cites "shortcomings in oversight of the research" as an explanation for what may have happened during the trial.  Basically, potential complications for the Johns Hopkins clinical trial appeared in biomedical literature published prior to 1966, but the principal investigator failed to look back that far, thus failing to locate information that would have likely saved the life of the 24-year-old participant.

This leads us to this week's expert research tip:

Relying solely on recent information can be dangerous.  (And this is what you do when you rely on using PubMed or Google Scholar to perform your literature searches).  


In another Lancet article (2) then-President of the Medical Library Association Carol Jenkins asserts that the Hopkins case underscores "the complexity of finding and evaluating biomedical literature." This same article concludes by positing that this situation serves to "highlight the value of older medical literature," and summarizes by saying that new questions very well may be answered in old literature.

Often, we want to limit our searches in PubMed (and other databases) to the most recent 2 or 5 years.  However, situations like the 2001 Johns Hopkins incident highlight why this is so dangerous.  Interestingly, Medline (PubMed) doesn't go back *that far*.  Earlier this year, it extended its reach back to 1947, which is definitely an improvement (3).  When the JHU researcher did his literature search in 2001, Medline only went back to 1966, which was part of the issue.  Similarly, ISI Web of Science only goes back to 1975, and no one really knows what Google Scholar searches, so no one knows how far it goes back, either!

So, how do you make sure you cover your bases and find important information within older literature?  My advice includes several options:

• Always review the references/bibliographic information included in articles that are relevant to your topic. If an article is foundational or seminal in a particular area of research, then your article should be citing it.
• Use databases or even print indexes that go back as far as possible.  VT, for example, has access to Biological Abstracts on paper that goes back to 1926.  Other indexes/archival databases that you could check might include:
• JSTOR Health Sciences, which goes back as far as 1840 for some titles
• American Chemical Society publications, which go back as far as 1879
• Engineering Village, which may have some relevant applied science information going back as far as1884
• ProQuest Historical Periodicals, which offers information from 1740-1900
• And many others, available through University Libraries databases

Although I wouldn't say that it's always necessary to search through 300-year old literature, I want to mention this to remind you all that current information, while desirable, isn't always sufficient.  If you want more suggestions or help with searching older literature, simply contact me!

References
1.  Ramasay S.  Johns Hopkins takes responsibility for volunteer's death.  The Lancet 2001;358:(9277):213.

2.  McLellan, F. 1966 and all that--when is a literature search done?  The Lancet 2001;358:(9282):646.

3.  PubMed extends its reach--biomedical database moves back in time to 1947.  NLM Technical Bulletin  2010; 373:e8.