A Case of Clustered Clarity: UP Health Sciences Library System

Page 1 of 2

Article ImageCase Study
Vivisimo & The University of Pittsburgh Health Sciences Library System

Company: The University of Pittsburgh Health Sciences Library System
The Health Sciences Library System (HSLS) provides information services to and serves as the central repository of print and electronic resources for faculty, medical staff, students, and researchers in the School of Health Sciences and the Medical Center at the University of Pittsburgh.

Business Challenge
The HSLS offers access to more than 300 health and biomedical titles in ebook form. Federated search allows for the consolidation of these resources, but most solutions don't excel at drilling down to find specific answers to specific questions—a must for HSLS's mostly clinical clientele—and they often fail to compensate for the amorphous nature of medical terminology, which consists of a host of terms with varying meanings depending on the context. The challenge was to ensure that the best information could be found quickly and easily.

Vendor of Choice: Vivisimo, Inc.
Vivisimo offered the company a new way to facilitate effective searching through the use of its proprietary Clustering Engine, which organizes search results into a series of folders, enabling users to start with simple search queries and then drill down through these clusters to obtain the best information available. With Vivisimo Content Integrator, a search can be performed across multiple content sources; Vivisimo Velocity packages these two technologies into one product.

The Problem in Depth
In the realm of academic libraries, federated search has already established its usefulness in consolidating disparate resources in the pursuit of forming a holistic view of whatever topic is being researched. But finding as much information as possible is not nearly as effective when you're trying to find a specific answer to a specific question. "The types of digital books that we buy are overwhelmingly clinical, clinical support materials, etc.," says Deb Silverman, associate director for Resource Management for the HSLS. "They're not research tools; they're answer tools. Nobody actually reads these things. They just go in, find their info, and get out."

When trying to find the proverbial hypodermic needle in the haystack of available documents, many clinicians rely on the small subset of resources they know and are comfortable with to uncover the answers they're looking for. "Oftentimes, when faced with massive search results, people will simply go with an adequate answer," Silverman continues. "In medicine, we were concerned that people find the best information."

Finding the best information has not been an easy task for HSLS patrons up until now, especially considering the fact that the ebooks accessible through the HSLS come from a variety of different publishers and providers, and there's no clearly defined index of all the available resources. "Even with sophisticated users, there are so many great resources out there, but typically only a subset of those resources is being used," says Don Taylor, senior director of Life Sciences at Vivisimo. "And even if they were to find the proper sources, they would have to search them independently, demanding that they know how to manipulate each search engine individually."

Part of Silverman's duties includes managing and paying for these resources. For quite some time she'd been intrigued by the potential of federated search as a way to encourage a more complete utilization of the resources available to the HSLS's patrons. "It frustrated me for a long time that we weren't getting enough bang for our buck for what we were paying for ebooks," she laments, "but I didn't like federated search because of the management of really large search results," which, more often than not, didn't allow for any way to quickly drill down through these results to find small pieces of information.

Exacerbating this problem is the very nature of the English language in the medical arena. "In medicine, there are a lot of words that have a lot of different meanings in different contexts," says Silverman. As an example, she cites the word "blast," which, when using a full-text search of medical ebooks, can reference everything from the abnormal blood cells associated with leukemia to injuries caused by explosions.

Page 1 of 2