Conversion of an Academic Health Sciences Library to a Near-Total Electronic Library: Part 2

Abstract

In the last ten years, many health sciences libraries have gradually converted their collections from predominantly print journal subscriptions to electronic-only subscriptions. This is being driven by budget, space issues, and user preference. The desire to retain both print and electronic versions of journal titles has proven to be unsustainable for many health sciences libraries in the face of flat or shrinking budgets and increased demand for space at the institutional level. Due to demand for space for a clinical simulation facility at the Pennsylvania State University College of Medicine, the George T. Harrell Health Sciences Library was faced with the need to accelerate the shift to primarily an electronic collection when more than 80% of the print journals and more than 20% of the print book collection were removed from the library. Part one of this article discussed how a case had to be made to the college administration that the older literature was still utilized and had value—and that it would be worthwhile to selectively purchase the electronic “backfiles” or archival files to replace high-use print journals. This second part of the article provides a detailed discussion of the decisions that resulted from the data analysis, subsequent actions that were taken to remove the collection, why certain backfiles were selected for purchase, and resulting impacts on both library users and library staff workflow.

KEYWORDS:

• collection overlap
• culture shift
• digital libraries
• electronic journals
• health sciences libraries
• journal backfiles
• journal usage
• off-site storage
• space allocation
• stable electronic access
• staff workflow
• weeding

INTRODUCTION

Part one of this article appeared in the previous issue of this journal and discussed the issues leading to the decision to convert the majority of the print journal collection to an electronic format for the George T. Harrell Health Sciences Library (Harrell HSL) at the Hershey campus of the Pennsylvania State University. ¹ Harrell HSL is the primary research library for the Penn State University College of Medicine, Milton S. Hershey Medical Center, and The Penn State Hershey Cancer Institute. Approximately 7,000 employees work at the Hershey campus and the more than 15 affiliated clinics in the surrounding Central Pennsylvania region. About 550 medical students are enrolled at the College, as well as more than 250 graduate students in various basic science specialties. The college and Harrell HSL have been in existence for more than 40 years, and at its peak the library's print collection had approximately 2,000 active serials subscriptions with approximately 110,000 journal volumes shelved in the stacks. The University Libraries, located at the main campus of Penn State University, is a two-hour drive northwest of Hershey.

Harrell HSL is centrally located in the main building on the Hershey campus midway between the College of Medicine and the Milton S. Hershey Medical Center and has often come under consideration when the need for space by the institution arises. In 2005, several significant events had a profound impact on both the space and the strategic direction of the Harrell HSL. Part 1 of this article provided an in-depth discussion of the factors leading up to the decision to repurpose 10,000 square feet of library space for a clinical simulation laboratory and learning complex. Some of the factors influencing this decision included the ongoing evolution of the open-access movement, availability of more e-journal “backfile” content, the unsustainability of maintaining parallel print and electronic journal collections, unexpected access to off-site storage space at the main campus, user preference, and, of course, the cost of space. As Montgomery et al. stated, “Storage space for low use bound journals is a major expense.” ² Once the decision to annex library space had been made by senior management, the Harrell HSL faculty and staff undertook the massive collection analysis that was needed to prove the value of the content of the print collection by showing which resources needed to be preserved, methodically redistributed to other libraries, or replaced with the purchase of electronic backfile packages, rather than simply discarding print volumes with no further consideration.

Part 1 of this article discussed various assessment methodologies, including usage analysis, levels of stable electronic content, overlap of collection holdings with other Penn State campuses, and the availability of electronic backfiles or archives to replace the print volumes. Part 2 will discuss how the various assessment factors were combined to create a “decision hierarchy” for retention or deletion; give a broad overview of the resulting decisions, including how interlibrary loan use of the collection was a factor; discuss the actions taken to ensure the print volumes were removed in a logical and responsible manner; and explain the decision process used to determine which backfiles to purchase. The authors will also discuss the resulting impact on library users and staff workflow since the print collection was downsized and the new challenges faced by the Harrell HSL.

MAKING THE DECISIONS

Creating a “Decision Hierarchy”

When it was determined that 10,000 square feet of the Harrell HSL's second floor would be repurposed for a clinical simulation lab, a false assumption prevailed that most of the print journal collection was already available online and that it would not be a problem to simply discard more than 90,000 bound volumes. Others assumed that most journals were duplicated at other Penn State campuses and that it would be easy to obtain articles through intercampus document delivery or through interlibrary loan from other institutions. Even if all journals had been available at other campuses, the impact on the interlibrary loan staff would have been overwhelming. The library faculty and staff knew, of course, that not all of the collection was available electronically and that it was not completely duplicated at other campuses, nor was it available “for free” on the Internet; however, an analysis of the collection had to be done to generate hard data to disprove these false assumptions.

Given the finite nature of the remaining library footprint, the clear need to preserve student study space, and the need to house the print monograph collection, the decision was made early in the process to retain, at a maximum, up to 3,500 linear feet of print journal holdings on site. Because of this limited space, a two-tier “decision hierarchy” was developed to methodically rank the print journals according to an identified set of criteria. The ranking would help to quantify the decisions about on-site retention, off-site storage (two hours away at the main campus), or removal for each of the 2,500 titles that were housed in the Harrell HSL. The following criteria were used to develop the decision hierarchy.

PRIMARY CRITERIA

	Print holdings duplicated by stable electronic access (current and backfiles)—all volumes of a title; some volumes; no volumes.
	Duplication of holdings within the University Libraries—all volumes; some volumes; no volumes.
	Usage data for the print holdings—most recent four years' circulation and in-house use.

SECONDARY CRITERIA

	Amount of shelf space (linear feet) required per title.
	Interlibrary loan use of the print holdings as a factor for retention.
	Availability of titles/runs at regional or national levels.

The primary criteria were used to rank the titles and the secondary criteria helped to “fine tune” the decisions. Electronic duplication, print holdings duplication at other Penn State campuses, and usage data were the top three criteria used to create ranking levels. The remaining factors of interlibrary loan use, shelf space, and regional availability were then used to make further determinations. Although this proved to be a very labor-intensive and time-consuming process, the resulting analysis supported a high level of confidence in the decisions. Table 1 shows the combinations of criteria and the decision ranks that were used in the assessment.

TABLE 1 Decision Hierarchy

Rank	Print duplicated by electronic (current and backfiles)	Print at other campuses	Use of print volumes in four years	Decision notes
1	All volumes	All volumes	Not needed	Discard
2	All volumes	Some volumes	Not needed	Some volumes to offsite; discard duplicates
3	All volumes	No volumes	High use at Hershey	Keep some at Hershey
4	All volumes	No volumes	Low use at Hershey	Send volumes to offsite
5	Some volumes	All volumes	High use at Hershey	Keep some at Hershey
6	Some volumes	All volumes	Low use at Hershey	Discard
7	Some volumes	Some volumes	High use at Hershey	Some volumes to offsite; keep some at Hershey
8	Some volumes	Some volumes	Low use at Hershey	Some volumes to offsite; discard duplicates
9	Some volumes	No volumes	High use at Hershey	Keep all at Hershey
10	Some volumes	No volumes	Low use at Hershey	Send volumes to offsite
11	No electronic	All volumes	High use at Hershey	Keep all at Hershey
12	No electronic	All volumes	Low use at Hershey	Discard
13	No electronic	Some volumes	High use at Hershey	Keep all at Hershey
14	No electronic	Some volumes	Low use at Hershey	Some volumes to offsite
15	No electronic	No volumes	High use at Hershey	Keep all at Hershey
16	No electronic	No volumes	Low use at Hershey	Send volumes to offsite or discard

Overview of Resulting Decisions

The analysis of the print journal collection allowed Harrell HSL to make the most logical use of the available shelf space. The library went from approximately 12,000 linear feet of standard shelving in roughly 32,000 square feet of library program space to 7,000 linear feet of compact shelving (3,500 feet for journals and 3,500 feet for books) housed in the remaining 22,000 square feet of library program space. It simply was not an option to keep all of the print journal collection or even a substantial portion on site. Off-site storage at the main campus did not become available until late in the analysis, but when it was determined 3,000–4,000 linear feet of space would be available for the Harrell HSL journals, more than 12,000 volumes were shipped to that facility. Approximately 22,000 volumes were retained at the Harrell HSL, and the rest was discarded. What were the deciding factors for retention? Some of the titles that were kept on site were considered “the classics” such as Science, JAMA, and Nature, all with holdings back to the first volumes; some were journals that were not available electronically but still saw significant use; others were titles that were used heavily both locally and for interlibrary loan lending; and others were titles that were kept because of a local obligation such as those published by Pennsylvania or regional health organizations.

Decisions were modified up until the day the volumes were physically removed from the shelf. Data were updated if electronic backfiles became available and were purchased or if in-house or interlibrary loan use increased significantly. No holdings information was changed in the library catalog or elsewhere until the volumes were physically removed from the library, but once the volumes were removed there was no “turning back.” Users and library faculty and staff were reminded that the holdings information in library systems would take a significant amount of time to update.

Users were also informed that Harrell HSL staff would identify and acquire materials that were needed locally, whether through the purchase of electronic backfiles when feasible; through document delivery from the off-site storage facility and the other Penn State campuses; or through interlibrary loan. Commercial purchase options were considered and utilized if a document was needed for a patient or administrative emergency and could not be reasonably obtained otherwise.

The Interlibrary Loan Factor

Interlibrary loan lending activity was considered an important criterion when making decisions about the final disposition of the print journal collection. Harrell HSL participates in both DOCLINE and OCLC, and is a member of the CIC Resource Sharing Agreement. In addition, it serves as a Resource Library for the Middle Atlantic Region of the National Network of Libraries of Medicine. Because of these relationships, it was important to understand the impact that becoming a digital library would have on the Harrell HSL interlibrary loan lending operation.

Much recent library literature that has discussed interlibrary loan implications of converting journal collections from print to online-only subscriptions has indicated that interlibrary loan levels of activity decreased overall. In 2004, Lynn Wiley, from the University of Illinois Urbana-Champaign in Urbana, published an excellent discussion about interlibrary lending and e-journal licenses. ³ It was interesting to note that not all of the limitations on interlibrary library loan were because of the license agreement. Libraries responding to Wiley's survey indicated other reasons for not filling requests from electronic versions of journals, including difficulty accessing the license information easily and as a result, making decisions to fill requests only from print versions of journals even if the electronic version was available. In their article “Changing Patterns in Interlibrary Loan and Document Supply,” Goodier and Dean also voiced this same difficulty. ⁴ The Harrell HSL clearly experienced a similar decline in lending activity as illustrated in Table 2.

TABLE 2 Lending Requests Received/ Filled at Harrell Health Sciences Library (from AAHSL Statistics)

Year of activity	Number of lending requests received	Number of lending requests filled
2000/2001	12,792	10,005
2001/2002	10,305	7,670
2002/2003	9,463	8,496
2003/2004	10,993	8,608
2004/2005	8,041	7,110
2005/2006	6,986	6,469
2006/2007	5,258	4,877
2007/2008	10,791	9,107 (purchased electronic backfiles, began adding holdings to SERHOLD)
2008/2009	15,279	10,110 (purchased electronic backfiles, continued adding holdings to SERHOLD)

Harrell HSL staff conducted a thorough analysis of existing licensing agreements and contracts to determine which electronic resources would allow interlibrary loan lending in some format. These titles were added to SERHOLD, the National Library of Medicine's (NLM) database of serial holdings used by DOCLINE members. As additional electronic journals are added to the collection, the interlibrary lending clause of each license is carefully reviewed to ensure access. These actions have reversed the decline in interlibrary loan activity. The recent dramatic increase in lending requests approaches levels similar to when the journal collection was mostly in print format. In addition, a subset of print titles that had low to moderate use in-house and that would have been discarded or shipped to off-site storage, was retained on site because interlibrary loan lending from those volumes was significant. However, because the off-site storage facility was two hours away at the main campus, it was decided that the 12,000 volumes that were transferred would become part of the collection at the main campus and would not be practical for interlibrary loan lending by Hershey.

LOGISTICS FOR REMOVAL OF THE PRINT VOLUMES

After the assessment was completed using the decision hierarchy, a distribution plan was put in place. The logistics of relocating, discarding, or moving more than 90,000 bound journal volumes was daunting. Early in the project, a faculty member had identified a university in Serbia that had expressed an interest in acquiring any volumes that Harrell HSL decided to discard. It was hoped that the discarded volumes could be shipped overseas; unfortunately sufficient funding for the packing and shipping costs (about $100,000) could not be found. Instead, arrangements were made with a local recycling company to remove the 55,000 volumes at no cost to the Harrell HSL. Approximately 12,000 volumes were moved to the off-site storage facility in State College, Pa., with another 1,000 volumes also being moved to the main campus in order to fill gaps in the Life Sciences Library collection; the National Library of Medicine was contacted and approximately 200 volumes were shipped to Bethesda, Md., to fill gaps in their collection; and 22,000 volumes were retained by the Harrell HSL. A buyer was also found for the 12,000 linear feet of both metal and wooden shelving. The money received for the shelving was not significant, but the company supplied the labor and transportation to remove the shelving, which also saved the project that expense.

Professional movers were hired to do a multistaged shift of the volumes that were retained—from existing stacks to a “holding” area, then to the newly installed compact shelving. With all of the moves that occurred, it was a challenge to ensure that each volume was handled correctly – particularly when only a few volumes were to be pulled from a title to fill in gaps at the National Library of Medicine or at the main campus library. To help minimize possible errors, several actions were taken:

	First, volumes to be sent to the main campus library were pulled and shipped.
	Next, volumes to be sent to the National Library of Medicine were pulled and shipped.
	Volumes to be discarded were “turned down” on the shelf and marked with a black “slash” across the top of the volume.
	Volumes that were not to be recycled were “roped off” on the shelf with strips of masking tape. There were a few problems in that the tape would not stick to some volumes or became loose when users pulled volumes from the shelves. The stacks were still accessible and in use during all of the processing.
	Once the 55,000 volumes to be discarded were removed from the shelves, then the 12,000 volumes that were to be shipped to off-site storage were marked with another color across the top of the journal and were “turned down” on the shelf until they could be pulled by the professional movers. Costs associated with moving the collection were part of the overall simulation laboratory/learning center project.
	Finally, about 22,000 volumes were moved from stacks on the second floor of the library to about 3,500 linear feet of compact shelving installed on the main floor of the library. Staff had to remove the masking tape strips before the volumes could be moved—not a small undertaking.

Overall, out of the more than 90,000 volumes that were managed, only about 300 volumes were marked incorrectly, moved to off-site storage in error, or discarded by mistake by the recycling company, a result that was well worth the planning and effort. Table 3 gives a synopsis of the numbers of titles, volumes, and shelf space that were managed with the project.

TABLE 3 Synopsis of Conversion from Print to Mostly Electronic Journal Collection

Date	Distribution	Number of titles	Number of volumes	Shelf space in linear ft.	Access to content available through
Before Summer 2008	Existing print journal collection	2,162 titles (but not all current subscriptions)	91,632 volumes	About 12,175 linear feet	In the Harrell HSL and many also available electronically
Summer 2008	Volumes sent to National Library of Medicine to fill gaps	Selected volumes from about 45 titles	200 volumes	15 feet	Most are free through PubMed Central
October 2008	Discarded volumes removed	1,338 titles	55,261 volumes	6,600 feet	Electronic full-text access, document delivery, ILL, or pay-per-view
November 2008	Volumes moved to Life Sciences Library at University Park	Selected volumes from about 65 titles	976 volumes	60 feet	Available free through internal PSU document delivery systems
November 2008	Volumes moved to annex space at University Park	Selected volumes from about 437 titles	12,624 volumes	2,000 feet	Available free through internal PSU document delivery systems
December 2008	Journals retained at Hershey	277 titles (some newspapers and some book series)	22,571 volumes	3,500 feet	In the Harrell HSL

BACKFILE SELECTION AND PURCHASE

The journal backfile packages that were considered and the resulting decisions to purchase are listed in Part 1 of this article. Packages were selected based on:

	How many print titles were going to be “replaced” by the backfile purchase?
	The levels of usage that were recorded for the print titles to be replaced.
	The amount of shelf space that would be “cleared” with the removal of the print volumes duplicated by the backfile package.
	The terms of the license agreements. Were proxy access, interlibrary loan use, etc., allowed?
	The cost of the package including a subsequent annual maintenance fee.

Some packages were not selected for purchase because the content did not meet the intended purpose of replacing the print volumes in the Harrell HSL collection. At that time, the goal was not to add new content, even though older years of some additional titles were part of the content of many packages. Entering this information into the library's A–Z list sometimes led to questions from users as to why the library “owned” older content and not the current years of a title. In some cases, subscriptions to the current years of a title were added to the library's collection because of first having access to the backfile, a slightly “reversed” collection development process.

DATA MAINTENANCE AND “CLEANUP”

As decisions were being made about what to do with the physical journal collection, it became obvious there would be substantial maintenance and cleanup work associated with the transition from print to electronic. Work forms were created for each title to better track the work that was to be completed in all of the relevant systems, including:

	The ILS System—the catalog, serials control records, order records, MARC records, MARC holdings records, and barcode/item records for individual volumes.
	DOCLINE—delete print holdings; add electronic holdings for e-journals that would allow interlibrary loan lending.
	LinkOut—delete print holdings, add electronic holdings, and add new titles.
	Databases such as Ovid, Web of Knowledge, SciFinder (Chemical Abstracts), and others—add links and holdings information for electronic journals.
	A–Z list—make changes in holdings, activate links, delete print holdings, change location, and make other notes.
	EZProxy—updated to reflect current access.

There were also unexpected “pockets” of work that were discovered as the cleanup progressed, such as the removal of analytic records in the catalog for older journal volumes and decisions about what to do with the print issues that came with a “combination” print/electronic subscription. Also, when the journals were shifted to the new compact shelving, it was decided to alphabetize the titles exactly word-by-word to match computer alphabetization, rather than the former arrangement of alphabetization by only the significant words in the title. This caused some additional work to shift the volumes by an alphabetical order that included “stop words” such as “the,” “for,” or “of” in the journal titles. (Example: Journal of the American Medical Association was shelved as “Journal American”; but in the new arrangement it is shelved as “Journal of … ”—a much different location in the order of titles.)

THE “POST-PRINT” IMPACT

Resulting Impact on Library Users

When the project began in 2005–2006, it was feared that there would be a tremendous objection from Hershey faculty and staff about the transition from a print-based journal collection to mostly an electronic journal collection. However, when the print volumes were finally removed three years later, this was not the case. Library users have come to expect easy access at the point of need. There has been the occasional question, more out of curiosity, about what happened to the print journals, but basically the response from faculty, staff, and students has been very positive. It is believed this is the result of ongoing communication with faculty, staff, and students as the project unfolded, as well as the changing needs and expectations of users. Harrell HSL's experience mirrors that of the Cushing/Whitney Medical Library at Yale University. Dollar et al. observed that “patrons showed little reaction to the idea of canceling print journals,“ ⁵ and “except for a few faculty asking about preserving access to online content, the primary response was silence—a clear indication that constituents had already moved to a primarily online environment for their information needs, while the library was still catching up.” ⁵

On occasion, Harrell HSL users have encountered access problems or incomplete content in a journal, but overall the number of complaints about the print version no longer being available on site has been very few. People clearly prefer the convenience of full-text access to the journal literature from their office or remotely from anywhere with an Internet connection. Some of the user feedback that is being addressed includes:

	Users still want to see a “current issues” section or lounge area to browse newspapers and casual literature. This area was temporarily removed with the reorganization of furniture and compact shelving in the library.
	There is a need for more computer workstations and for more “wet carrel” space for laptops since the collection is mostly accessible online.
	Users tend to think that if a resource is not “hot linked” in a database, then the subscription is not owned by the library. It is a constant challenge to maintain correct data in the library's A–Z list and database services such as Ovid and Web of Knowledge and to effectively communicate the breadth and scope of the current collection.

Another interesting outcome of the transition to mostly a digital library has been the impact on photocopying. Aggregate numbers for total copies remained high through 2004–2005 and then began to decline precipitously. At the same time, the demand for public printing increased significantly. When the decision was made to annex 10,000 square feet of library space, the Harrell HSL had already made a significant investment in electronic resources as discussed in Part 1 of this article. These e-resources already had an impact on the utilization of the print journal collection. The need to manage the public printing service became acute. The Harrell HSL, in partnership with the Information Technology Department, introduced a pay-to-print service in 2005–2006. As the photocopying declined, pay-to-print increased. Table 4 highlights this shift. Regardless of format, Harrell HSL users still prefer to print articles, although the initial increases in printing total numbers have dropped off. It is unclear if this trend will continue with pay-to-print.

TABLE 4 Transition from Photocopying to Pay-to-Print (Number of Pages)

Service	FY 2002/2003	FY 2003/2004	FY 2004/2005	FY 2005/2006	FY 2006/2007	FY 2007/2008	FY 2008/2009
Photocopy	450,571	390,244	300,185	220,062	186,577	129,317	88,476
Pay-to-print				427,695	501,530	281,235	346,844

Resulting Impact on Staff Workflow

Since the print volumes have “left the building,” a much greater impact has been experienced by the library faculty and staff. As stated by Carol Hansen Montgomery and JoAnne L. Sparks from Drexel University:

Some of these changes were obvious and predictable; others, such as the substantial increased need for skilled professional staff, were not. In brief, we set out to change the format of the journals from print to electronic, and it quickly became apparent that we were forcing fundamental changes in library operations. Almost no area of the library has been left untouched. ⁶

Significant work remains to address the profound changes in workflow dictated by the transition to electronic delivery of information. With many of the critical maintenance and cleanup projects nearing completion, Harrell HSL is embarking on an analysis and redesign of workflows throughout Collection Access & Management and Technical Services divisions. This scrutiny is in response to the “fundamental changes occurring in library operations.” ⁶ These changes will no doubt impact all areas of the library as administration begins reengineering processes and reorganizing staff to better serve user needs. There are a number of excellent articles published in the library literature that address the changes that libraries face with the conversion to e-journals. Both Graves and Arthur ⁷ and Yue and Anderson ⁸ in their respective articles discuss the use of flowcharting to visualize workflows associated with managing e-journals. Daryl Youngman in his article, “Library Staffing Considerations in the Age of Technology: Basic Elements for Managing Change,” addresses change from a human resource perspective. ⁹ Duranceau and Hepfer discuss staffing changes needed to support the significant changes in workflow and the increased complexity of managing e-resources. ¹⁰ Dollar et al. describe their experience “of redesigning the workflow of the Cushing Whitney Medical Library's technical services area to focus on directing scare resources to best meet the needs of its users.” ⁵ Harrell HSL faculty and staff will rely on these and many other examples to help guide them through the continuing process of change and transformation. Some of the specific issues/scenarios that Harrell HSL plans to address include:

	Obtaining or adopting electronic resource management software/systems to better manage the myriad of details related to electronic resources.
	Negotiating licenses to better meet the needs of changing user groups and network configurations.
	Working more closely with the Information Technology Department for more streamlined resource authentication and access.
	Having a better understanding of issues related to the cataloging and creation of metadata records so that users can more easily identify resources that are available.
	Training reference and other front-line staff to be more knowledgeable about the technology being used to access information such as wireless settings, laptops, handhelds, and Web-based tools that need to be developed.
	Expanding and evolving outreach services such as liaison programs, library instruction, and curriculum integration efforts.
	Exploring options for new ways to assess the library's impact and value of particularly staff and space.
	Repurposing staff from managing print journals to new responsibilities with electronic resources management, digitization, and metadata projects.

CONCLUSION

Parts 1 and 2 of this article describe Harrell HSL experiences with transitioning from a primarily print collection to a predominantly digital collection. After senior administration made the decision in 2005 to annex and repurpose 10,000 square feet of library space, the die was cast. It was unrealistic to retain the entire print collections within the confines of the remaining space. Clearly, forces beyond the library also influenced the direction of the library. ¹¹ These forces included changes in the reporting structure described in Part 1, ongoing inflationary increases in journal prices, user preference for online access, the open-access movement, and consolidation within the publishing industry.

Much work remains; the massive print journal review project described in these two papers is only the first step. There are many issues that must be addressed as the Harrell HSL completes its conversion into a digital library. These issues include the need to:

	Analyze and redesign workflows.
	Recognize that librarians and staff alike require new skill sets.
	Envision the “library of the future” scenarios.
	Better manage electronic resources through the use of next-generation systems.
	Above all, be agile in the face of great change.

In truth, the Harrell Health Sciences Library is only at the beginning of this journey.

ABOUT THE AUTHORS

Virginia A. Lingle, MSLS, AHIP ([email protected]) is the Librarian for Collection Access and Development and Cynthia K. Robinson, MA, AHIP ([email protected]) is Director; both at The George T. Harrell Health Sciences Library – H127, Penn State Hershey, 500 University Drive, P.O. Box 850, Hershey, PA 17033.