National Network of Libraries of Medicine

nnlm.gov

LESSONS LEARNED

The question of Internet access for hospitals has appeared and reappeared on MEDLIB-L (the medical library electronic discussion group) throughout the course of this project. The findings of the pilot connections project may provide a useful framework for the discussion of this issue and may suggest answers to some of these questions. We have organized this examination into categories of administrative, technical, financial, and organizational issues. Each category is discussed in turn and conclusions, or lessons in each are, are drawn. We conclude this paper with a brief discussion of what we think the implications of these lessons are.

Administrative Issues

From the beginning of the project it was apparent that hospital administrators were attracted to the idea of the hospital as a node on the Information Superhighway. Some were excited by the prospect of being able to send and receive documents and diagnostic images at high speeds. None appreciated what was required to do this. It was not well understood initially that full service, institution-wide Internet access (a direct connection) was not something purchased from a vendor, who installed it and trained key staff how to use it. Librarians in even small organizations have been implementing new information technologies using this model for decades. In contrast, establishing a direct Internet connection affects an entire organization and requires that administration be knowledgeable, committed, and supportive. Collaborative relationships among administrators, information systems staff, and the librarian are essential because the issues are complex and any single decision can affect many areas in the organization.

Dial-up access, however, is possible without a strong commitment from senior managers or administrators, and still opens up a variety of resources not available through the usual value- added networks familiar to librarians,. It is a logical extension of traditional library resources and services. It is a matter of selecting a provider, learning to budget for the service as with any new electronic service, learning to use the resource effectively, and training others to use it. In short, it's using a new technology to do what we have always done.

Dial-up Internet access is a low-cost, low-risk endeavor suitable for the library or departmental level. Whereas, a direct connection involves the entire organization and requires high level knowledge, commitment, and support.

Technical Issues

The lack of technical expertise available within the community hospital computing environment is probably the biggest barrier to be overcome in implementing a direct Internet connection. When the project began, none of the sites had staff who knew about the Internet, or had any experience with the UNIX operating system or the TCP/IP networking software protocols. NorthWestNet was accustomed to working with academic organizations with ample onsite networking expertise and technical support. The service provider staff were cautious about entering into this joint endeavor, knowing how necessary the onsite expertise was. They agreed to participate with the general understanding that they would probably have to provide more technical support than they would ordinarily, and the RML would provide the required user support, principally in training and applications.

No one fully appreciated how difficult the lack of onsite technical support was going to be, or how much re-thinking and retrofitting of existing network structures would be required. Librarians are faced with a steep learning curve in order to understand the components of the Internet connection and related issues. At every turn, issues of compatibility with one configuration or another grew into difficult questions and created delays. Aggressive technical assistance was never envisioned as being a part of this project, and yet it was clearly needed. The Internet service provider was not prepared to offer the support that was needed, in keeping with their standard practice and with their understanding of their role in this project.

Hospital system vendors were often of surprisingly little help. One site's primary hardware vendor was openly hostile to the site's participation in this project and to the idea of connecting to the Internet. Many hospitals have made decisions over time to purchase only turnkey computer systems and applications and to rely exclusively on the vendor for support. They therefore have no systems staff onsite who are capable of assisting with complex networking problems. Another site's networking software developer was sympathetic with the compatibility problems they were experiencing, but ultimately not helpful because of the mix of hardware and software in question and the use of untested combinations and first releases of products. Often it was a case of "no one knows how to do it."

All sites faced some early telecommunications problems with the Internet service provider's host computer and occasionally with other Internet services. But, the dial-up sites were spared any other technical problems. The four sites were provided with high speed modems (14.4k bps) in anticipation of taking advantage of SLIP connections, providing virtual access to full Internet services.

The project planners thought the Internet service provider was going to make SLIP service available, possibly on a test basis through this pilot project, but this did not happen. The dial-up sites were limited to low speed (2400 bps) access throughout the project. This has been a disappointment and has hindered the use some of these sites have made of the access. The reason given for not providing this enhanced dial-up service was that too much technical support would be required by users of SLIP. SLIP is more difficult to use but we note that many other service providers are now offering this service at relatively modest fees.

The lack of technical resources available locally often led the project librarians, and occasionally other staff from the hospital sites, to turn to the RML for assistance. The RML staff did not always have the answers, but they were willing to assist and had a helpful perspective. They often knew, particularly the Systems Coordinator, where an answer might lie or what direction to take. The project participants valued this assistance. Other librarians, and some health professionals in the region, are also beginning to consult the RML on Internet connection and use issues.

Occasionally, the Internet itself provided the means of acquiring assistance from experts at other locations. But, as earnest as this help often was, it does not take the place of onsite or local expertise. Community hospitals attempting to establish a direct Internet connection need the extensive, close involvement of knowledgeable and trained systems engineers.

Financial Issues

Cost is a major issue in establishing a direct Internet connection at a community hospital. It is disconcerting to participants in this project to see the occasional comment from someone safely ensconced in an academic setting that Internet access is free. It is not.

The low speed dial-up access provided to all seven sites (enabling them to reach their Internet accounts on the host computer in the Seattle area) was not provided through a toll- free service, but the long distance charges were covered by project funds. Two sites had access to local Internet hosts, thus avoiding long distance, and one was in the local calling area. The four others incurred long distance charges and some of those had average monthly line charges exceeding $250. Overall, long distance charges were seen to be a clear inhibitor of use.

Many substantial costs are involved in establishing a direct Internet connection. A high speed (56k bps) data circuit needs to be installed by the telephone company ($1,500). Monthly charges for the data circuit range from $190 to $600 per month depending on location -- the more remote, the more expensive. The Internet service provider charged a one-time fee ($10,000) for a connection service package. This service included installing and maintaining the network router which is a specialized computer to direct traffic between the local area network and the Internet. First year costs, then, range from $14,000 to $19,000, including $2,280 to $7,200 for circuit line charges.

These external networking costs were covered by project funding. Other, internal costs -- for example, hardware and software required to bring a network into TCP/IP compliance -- had to be met by the site.

In addition to the cost for tangibles such as circuits and hardware and software, there are also annual membership fees levied by the Internet service provider to support the value-added services and other benefits enjoyed by the user community. Institutional fees are based on a combination of the bandwidth used by the institution (i.e., how "large," or fast, a data circuit) and the expected use made of the service, as measured by the annual operating budget of the institution. Corporate entities and government agencies are levied the standard fee -- $15,000 for a 56k bps circuit. Health care organizations receive a 20% discount from this standard fee -- $12,000. Education and research institutions receive a further 20% discount -- $9,000.

At the beginning of this project there was a simpler fee structure in use based only on institution size, as measured by total FTE count. This was fair enough when all member organizations were of the education and research mold. The project planners argued, however, that this structure was biased toward academic institutions and was not reasonable when applied to multiple types of organizations. We argued against applying the size model to community hospitals, contending that a far smaller base of community hospital staff will make use of network resources than is the case in the typical academic or research setting. Project experience bears this out. We recognize that this might not be true in a number of years, when Internet use is more pervasive, even in community hospitals, and there are more clinical applications.

Special pilot project fees were negotiated for the sites. The direct connection sites were assessed $8,000 for the first year, which was mostly, but unfortunately not entirely, covered by project funds, and $12,000 thereafter. The dial access sites were assessed a fee of $2,500 each, which included seven online accounts (one for the project librarian and six others to be assigned to key staff at the librarian's discretion). As noted earlier, this dial-up access was not toll-free, or high speed.

The Internet service provider did not intend to offer dial-up access as a standard service and membership option. The provider's approach emphasizes value-added services that are more attractive to those academic institutions which are widely networked and can take full-scale advantage of the economic value of these services. This value is effectively lost through a limited dial-up access.

Cost is always a major issue in health care and becoming more so. It is highly visible to administrators, especially when introducing a new technology, and even more so when the need for that technology is not well understood. Internet service providers have to realize that the community hospital market will require a different strategy and fee structure than that used for academic and research institutions. With affordable, entry level opportunities (e.g., a low cost SLIP connection over a local area network), community hospitals can build a base of support for a full service, direct Internet connection.

While this is potentially a large market, Internet service providers will have to develop new approaches to take advantage of it. It is probably not economically feasible for a provider to offer the intensive but affordable service this market requires, at least for the foreseeable future. However, many factors -- economic and regulatory, to name two possibilities -- could cause this to change quickly in the next few years.

Organizational Issues

Hospital librarians have played the central role in introducing the Internet to these clinical settings. This is a logical extension of the librarian's role as the resident expert in managing external information resources. There is no controversy in this role extension as long it does not raise potentially difficult questions about the hospital's information systems strategy, structure, and accountability.

Dial-up Internet access does not raise these questions. Hence, it's organizational risk is low. Attempting to establish a direct Internet connection, in this case promoted by the librarian, often does raise these questions. Although the librarian may often be viewed as a neutral (or merely non-threatening) member of hospital management, this questioning of hospital computer networking can be seen as threatening by the information systems staff. The perceived threat and a defensive response may lead to a territorial stand off.

One of the sites has had a particularly difficult and protracted turf battle over the question of Internet access. In this case, the information systems staff did not think the librarian had any business making recommendations about the direction the hospital's computer networking structure should take. This has led to a good deal of professional distrust on both sides of the argument and to delays. This will surely be faced by other hospitals.

The Joint Commission on the Accreditation of Healthcare Organization's newly revised standards for information management implicitly call for information systems staff, medical records staff, and hospital librarians, to work together; to collaborate in the information management of the hospital. The librarian's potential contribution to the hospital's response to these performance-based standards can lead to eventual enlightenment on the value of networked information in the community hospital setting. The standards may be used as an argument for resolving, or at least ameliorating, potential turf battles.

Turf battles notwithstanding, each group has concerns that are legitimately distinct, but these cannot be considered in complete isolation from each other. Clinical staff are focused on inpatient care. Information systems staff are focused on keeping the hospital's systems running. Medical records staff are focused on patient information and billing procedures. Librarians are rightly focused on the environment and on external information resources -- beyond the hospital walls -- in order to fulfill the information needs of clinicians and administrators alike.

These distinct professional outlooks can contribute to an overall view of information management that will serve the needs of the organization. Just as many academic medical centers have developed IAIMS models, so too, must community hospitals develop an integrated approach to information management. It will be serious disservice to the community if the promise of Internet connectivity is held hostage in fruitless battles over departmental control.

Go to Abstract / Introduction & Project Summary / Lessons Learned / Conclusion, Acknowledgements & Resources / References