INABIS '98 Home Page Your Symposium Related Symposia & Posters Scientific Program Exhibitors' Foyer Personal Itinerary New Search

Providing Electronic Information
Resources to Rural Hospitals


Background and Introduction

 

Located in Iowa City, Iowa, The University of Iowa Hospitals and Clinics (UIHC) is the state’s only tertiary health care center. One of the largest university-owned teaching hospitals in the nation, the UIHC has 873 inpatient beds and serves patients and families from all of Iowa’s 99 counties. From July 1, 1996 – June 30, 1997, there were 41,818 patient admissions and 581,446 ambulatory clinic visits to the UIHC. The UIHC complex includes a main building and six interconnected buildings, covering approximately 4.0 million square feet.

Many of Iowa’s rural and community health care workers are geographically and professionally isolated from the information resources and specialty consultation services to which academic health professionals have easy, daily access. Of the state’s 117 hospitals, 96 are classified as either rural or rural-referral and 71.1% have 99 or less beds.

Recruitment and retention of health care workers in rural areas is difficult in Iowa as it is in other areas with geographically dispersed populations. To remain current in their respective fields, rural health providers must take time away from their patients to travel to sites where continuing education opportunities are available.

Thanks to a 3-year telemedicine contract from the National Library of Medicine (NLM) that began in 1994, The UI has developed a high-speed, point-to-point telecommunications network to deliver clinical and educational services to ten rural and community Iowa hospitals. This contract created the National Laboratory for the Study of Rural Telemedicine administered by the Telemedicine Resource Center and directed by Michael G. Kienzle, M.D, Associate Dean for Clinical Affairs and Biomedical Communications in the UI’s College of Medicine.

 UIHC’s Telecommunications Service Department supplied each of the hospitals in the network with the requisite electronics, equipment, and hardware needed to support the delivery of telemedicine applications. Circuits were provided by the state’s fiberoptic infrastructure, the Iowa Communication Network (ICN), and by local and long distance carriers in the areas served (for example, U.S. West and GTE).

 Staff from the Resource Center, Healthnet Project, and Virtual Hospitalâ Laboratory provided the ongoing training, on-line and print documentation, and in-service (hands-on) instruction necessary to ensure that utilization of telemedicine services was maximized at the remote hospitals. To further strengthen end-user support, the Resource Center installed a 1-800 help line and established e-mail capabilities between the rural hospitals and the UIHC.

Overall project goals for the Healthnet project were to:

 

Methodology

 

Equipment

 A total of thirty-nine Macintosh 7100 computers and Apple Laser Printers were configured and deployed to the networked hospitals. These computers served as multi-user workstations from which health professionals at the remote hospitals could access a variety of on-line resources. All computers were equipped with Netscapeâ browsers and a software program called At Easeâ which was loaded on each of the computers to protect the hard drives from overzealous downloaders and potential hackers.

A document delivery workstation (486 Intel/Windows 3.11) was configured and installed in each of the ten hospitals. These workstations were equipped with Arielâ software and were used almost exclusively as a central site repository for receiving and printing documents received via the Arielâ document delivery system. The document delivery service was enabled in late January of 1996 through the creation of a special web site which allowed individual users to request articles from the Hardin Library where interlibrary loan (ILL) staff retrieved, scanned, and sent the articles via the Internet to a central printer at each site. In some cases, these workstations were equipped with Netscape software and did double-duty as search stations.

Databases

Healthnet is mounted on a Unix server at the Hardin Library for the Health Sciences. Healthnet uses the OVID search interface and includes the following database:

 

All the databases include abstracts and easy access (a single mouse click) to a web-based document request form. Appropriate licensing agreements with OVID to provide searching software to the remote sites were arranged before the contract was finalized. Additionally, users could telnet to Micromedexâ, a computerized clinical information system which provided information in the areas of toxicology, poisons, drugs, and emergency medicine. The University of Iowa’s online book catalog, OASIS, was also available to rural end-users.


Local Support

Hospital CEOs were asked to appoint a 'telemedicine liaison' from their hospital personnel to facilitate communication with UI Telemedicine staff. These individuals included four directors of education/information services, two librarians, two administrative assistants, one radiology director, and one emergency room supervisor.

It should be noted that only three of the ten hospitals had a medical librarian in their facility. One librarian (not represented in the group described above) was a backup telemedicine contact for her hospital. Telemedicine liaisons at each hospital then decided where the workstations would be located in their facility to optimize utilization by their staff. Typically, locations selected were physicians’ lounges, nurses’ stations, administrative offices, and emergency rooms.

User interface

End-users first logged in to the project’s default home page. A successful login brought up the Virtual Hospitalâ which served as the ‘front end’ and allowed health professionals to search The Virtual Hospitalâ and access the Healthnet databases, the online document delivery request forms, Micromedexâ , and other health resources on the World Wide Web (WWW). All applications were timed-out to return to the startup login screen after ten minutes of non-use.

 

Results

Project Registrants

 As of October, 1997, 1,339 health and allied health professionals from the ten networked hospitals have registered to use the Virtual Hospitalâ workstations. To gather usage statistics based on site and profession, it was necessary to register each user individually. Table 1 shows the number of staff and registrants at all sites across professions.


Table 1

Telemedicine Registrants By Profession

Profession Number of Staff Number of Registrants Percentage
Administration

202

159

78.71%

Allied Health

1370

321

23.43%

Information Systems

41

22

53.66%

Librarian

3

3

100.00%

Nurse

1828

588

32.17%

Physician

481

246

51.14%

Total

3925

1339

34.11%

 

Healthnet Database Searching

Table 2 shows the use of the Healthnet databases by each profession for the eleven month period from July, 1996 through May, 1997.


Table 2

Healthnet Sessions by User Category (excluding librarians)

July 1, 1996 - May 31, 1997

User Category Medline CINAHL HealthStar CurCont CancerLit Total
Administration

43

2

12

5

1

63

Allied Health

74

1

3

3

0

81

Nurse

399

146

53

37

18

653

Physician

380

0

6

1

4

391

Total

896

149

74

46

23

1188

Electronic Document Delivery 

Electronic document delivery was enabled in late January of 1996 through the creation of a special web page which allowed individual users to request articles from the Hardin Library where InterLibrary Loan (ILL) staff retrieved, scanned, and sent the article via the Internet to a central printer at each site. The Arielâ software was used for this function.

The Hardin Library interlibrary loan staff achieved a "fill rate" for requested articles of approximately 90%. For purposes of this study, requests which could not be filled through the Hardin collection were not referred to other libraries. Since the Healthnet databases are coded by Hardin Library staff members to indicate which articles are held at the Hardin Library, some filtering was performed by the requesters who also had the option of limiting their Healthnet database retrieval to only those journals held at the Hardin Library.

A total of 4,620 requests were made by personnel at all of the sites from January, 1996 through August, 1997, resulting in 4,150 document transmissions. Although Davenport (712 requests) is the largest of the ten institutions, the highest number of requests were generated by Ottumwa (2,161 requests) where document delivery was heavily promoted and supported by the site librarian. It should also be noted that Davenport used the electronic document delivery program primarily to supplement its established interlibrary loan program: therefore, a much smaller fraction of their requests were generated via the electronic document delivery program.

 Figure 1 shows the total number of requests generated by each of the user categories. A total of 249 registered requesters generated 4,620 requests (or an average of 15.75 requests per actual requester). Also shown is the number of requests from unregistered users (i.e., those who did not formally register as telemedicine users but who used the web site to request articles). Since no registration data were available for these users, it was not possible to categorize them by profession. Nurses accounted for the greatest number of document delivery requests with over half (56.8%) of all requests coming from this end-user category.

 

Benefit of Services

 Site coordinators (n=10) were surveyed on a number of parameters including their perceptions of who received the greatest benefit from the programs available through the Virtual Hospitalâ workstations. For the end-user categories in Figure 2, site coordinators were asked to rate the benefits from 1-6 with 6 receiving the most benefit, and 1 receiving the least benefit. Average scores for each group are listed below and take into account that not all hospitals have a librarian or computer support personnel.

Figure 2

 

Information Workstations : Site Coordinators’ Perception of Benefits

6=Most Beneficial; 1=Least Beneficial

 
Group Score Minimum Maximum
Library Staff 5.7 4 6
Physicians 4.9 3 6
Nurses 4.0 1 6
Allied Health 3.6 3 6
Administrators 2.4 2 4
Computer Support Staff 1.8 1 5

 

Troubleshooting. A one-year analysis of the troubleshooting log (February 2, 1996 – February 26, 1997) revealed that the 220 incoming calls to the 1/800 help line during that timeframe could be divided into 13 categories based on the reason for the call. The top three problem categories related to 1) Informm, the mainframe-based online patient record system, which has a relatively complex security access structure (n=69); 2) system and server errors (n=26); and 3) technical problems relating to hardware and networking issues (n=23). Specific questions about Healthnet were fourth out of thirteen in frequency, with 21 calls being placed via the Help Line during the one-year troubleshooting study period.

Document Delivery. The special Web form used by participants to request article delivery appears to have worked quite successfully. The form did require, however, that the user click on a different web page to complete the form. Ideally, article delivery request capability would be linked directly to database searching so that articles could be selected from within the Healthnet database. We had hoped to incorporate this feature into the Healthnet software but were unable to do so in the time allotted to the contract period.

On the whole, the document delivery component worked very well. Thanks to an extremely efficient Interlibrary Loan Department at the Hardin Library for the Health Sciences, nearly all articles were delivered within one working day of the request. Six months into the service, it was necessary to place an upper limit on the number of articles which could be requested from a single user on a given day to avoid bottlenecks. This restriction was met with understanding by the site coordinators and end-users at each of the remote hospitals.

Healthnet Because the Virtual Hospitalâ workstations were Macintosh based, and the OVID search software was not initially available on this platform, database searching for the first six hospitals connected to the network took place in a telnet environment. This non-Graphical User Interface (GUI), while more difficult to navigate than the Web or Windows Client version, did not seem to pose as great an obstacle as was originally supposed. We do know, however, that toward the end of the study (June, 1997) when OVID Web software was made available at the sites, we received favorable feedback from the users with regard to this "friendlier" interface. Unfortunately, the statistical package which was put into place for the telnet version was not compatible with the Web version, so we were not able to compare usage figures before and after this change.



Discussion Board | Previous Page | Your Symposium |