WHAT IS ARROWSMITH AND HOW IS IT USED?
ARROWSMITH is interactive software that extends the power of a MEDLINE search. It operates on the output of a conventional search in a way that helps the user see new relationships and form and assess novel scientific hypotheses. It is based on the premise that information developed in one area of research can be of value in another without anyone being aware of the fact.
Currently (Jan 2004) MEDLINE input can be accepted from either PubMed or OVID, only. Read instructions on format and downloading.
The user begins with a question concerning the connection between two entities (for example, a dietary substance,"A", and a disease,"C") for which conventional searching provides no answer. Thus we ask: "can A, or a deficit of A, influence the course or risk of C?" A conventional search of MEDLINE for the intersection "A AND C" should be conducted BEFORE using ARROWSMITH, to determine whether the published medical literature might DIRECTLY answer the question. If it does not, then Arrowsmith might help find an indirect connection as described below. If A influences some factor, X, not mentioned in the direct A-C literature, where X in turn influences C, the A-C implication cannot be discovered by a conventional search. Without knowing X, a conventional database search cannot determine whether such an X exists, even though the literature on A and the literature on C separately mention X. But insofar as these connections are reflected in title words, subject headings, or in the CONCLUSIONS section of the abstract, ARROWSMITH provides a straightforward solution to this problem.
PREPARATORY STEPS
First conduct a MEDLINE title-word search for the word or term denoted by C, and download all records found. Next, create a similar but separate computer file based on the title-word or term corresponding to A. Title-word searching may be enhanced by including subject-headings as well. See additional suggestions for file-size and search strategy, following pointers on the Welcome page.
SENDING INPUT FILES; EDITING OUTPUT
Next proceed to the main input page, called STAGE 1. Enter the names given to each of the above two files. The files are then transmitted (uploaded) one at a time to ARROWSMITH. The software will produce first a list (called the "B-LIST") of all the important words and phrases (and MeSH terms) that are common to the two sets of titles. If the input format includes Medical Subject Headings, these also participate in the matching process. Title terms are ranked according to the number of MeSH terms that are shared by the A and C titles corresponding to any given title-based B-term. Each of the title terms is a potential candidate for the mysterious "X" mentioned above. Some of the terms are plausible and some are not, so the title-based list in general should be edited by the user, provided that a title format is used. (With the ranked list provided by subject headings, editing is largely unnecessary, and, in any event, not permitted.)
TITLE BROWSING
If you return at a later time to examine or review results of
Stages 1 and 2, use STAGE 3.
Each B-LIST is always displayed as a series of links. Clicking on
any B-term "X" results in displaying the corresponding titles that
contain both A and "X" and, next to these, titles that contain "X" and
C. From there you can also go back to any B-list. The output of
ARROWSMITH in effect serves as a title-browsing guide to the
scientific literature.
ARROWSMITH extends but does not replace conventional database searching, for it requires, as input, the results of a conventional search.