Developing a Database for a Literature Search Service at an Academic Health Sciences Library
 
Ben Harnke, MLIS; Kristen DeSanto, MSLS, AHIP; Christi Piper, MLIS; Nina McHale, MSLS; Lilian Hoffecker, PhD, MLS; Lawrence McGaughy

University of Colorado, Strauss Health Sciences Library
In the summer of 2017, librarians at the University of Colorado, Strauss Health Sciences Library, who were engaged in systematic literature searching -- the Search Team -- began to develop a formal literature search database to track details of the literature search requests.

Although the library had developed a formal service for conducting literature searches, tracking specific details about the requests (for example, the purpose of the literature search, the role of the librarian, the project timeline, etc.) were handled on an individual basis, typically in Word documents. This made assembling search service statistics cumbersome, time consuming, and inconsistent. The Search Team recognized the need for a system to track agreed upon standardized statistics and search details, both for reporting and in the event that searchers needed to fill in for one another. Furthermore, many libraries faced with starting a literature search service are often uncertain about the resources necessary or the steps to take to get started. Our library had an opportunity to systematically gather data that could objectively reveal to our library and others information such as time expenditure required to run this service.

While the library had in place an electronic search request form (Springshare’s LibAnalytics platform), it simply collected patron-provided search information that could be used in an initial search consultation. The Search Team looked at the possibility of modifying the LibAnalytics form to allow librarian input on the back end in order to start collecting more robust data. We quickly realized that, due to the various kinds of literature searches we offer, we would need a rather complex and customized form but none of the preprogrammed applications met the depth and flexibility we were looking for.

In a case of serendipity, the newly hired web applications programmer (Lawrence McGaughy), upon reviewing the current LibAnalytics form, suggested we build a new form in the Drupal web-hosting platform. He was confident that he could construct a conditional input-based form that met the requirements we were hoping for. He was furthermore confident that he could generate statistical reports based on the data fields in the form.

We quickly realized that this was a prime opportunity to evaluate our search service from top to bottom. After several meetings, it became clear we needed the form to collect three kinds of information: 1) search details, 2) time and resources devoted to the search; and 3) the use of the search service at university level.

The first item, search details, would be the information needed to actually complete the search and would consist of data such as purpose of search (systematic review, book chapter, guideline, etc.), date limitations, animal vs. human studies, databases to be searched, protocol registration, grey literature sources, study types to be included, language, delivery format, etc. The second item would allow us to advocate for search service resources and would be captured through time logs, tracking of electronic resources (databases, review software) and which library faculty were assigned to the search. The third item, use by the university, would identify individual author affiliation, including collaborators outside of the university, and, primarily in the case of a systematic review, the publication status (in-progress, submitted for publication, published, suspended). We also needed the form to track our paid search service option including invoice numbers and payment status.

Over the course of a year and many meetings with the web developer, we were able to complete a prototype form. Search Team librarians have been entering data since mid-October 2018 when the form went live.

As of late November, the database includes 18 searches in progress, 12 of which will include the librarian as a co-author. The time log gives only an estimate, since some of the entered searches started before the database was in place, but we can tentatively state that about 110 hours, or 14 working days, have so far been expended on these 18 searches.

The patron portion of the form is available at: https://hslibrary.ucdenver.edu/eform/submit/prof-literature-search-request.
Copyright 2014 - Doody Enterprises, Inc. - All rights reserved