RadioGraphics
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
 QUICK SEARCH:   [advanced]


     


This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Willing, S. J.
Right arrow Articles by Berland, L. L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Willing, S. J.
Right arrow Articles by Berland, L. L.
Related Collections
Right arrow Informatics
(Radiographics. 1999;19:169-182.)
© RSNA, 1999


infoRAD

A Radiology Department Intranet: Development and Applications

Steven J. Willing, MD, MBA1 and Lincoln L. Berland, MD1

1 Department of Radiology, University of Alabama, 619 S 19th St, Birmingham, AL 35233.


    Abstract
 Top
 Abstract
 INTRODUCTION
 WHAT IS AN INTRANET?
 DEPARTMENT OF RADIOLOGY PROFILE
 HARDWARE REQUIREMENTS
 SOFTWARE REQUIREMENTS
 KNOWLEDGE REQUIREMENTS AND...
 CONTENT DEVELOPMENT
 DEVELOPMENT PROCESS
 CONCLUSIONS
 References
 
An intranet is a "private Internet" that uses the protocols of the World Wide Web to share information resources within a company or with the company's business partners and clients. The hardware requirements for an intranet begin with a dedicated Web server permanently connected to the departmental network. The heart of a Web server is the hypertext transfer protocol (HTTP) service, which receives a page request from a client's browser and transmits the page back to the client. Although knowledge of hypertext markup language (HTML) is not essential for authoring a Web page, a working familiarity with HTML is useful, as is knowledge of programming and database management. Security can be ensured by using scripts to write information in hidden fields or by means of "cookies." Interfacing databases and database management systems with the Web server and conforming the user interface to HTML syntax can be achieved by means of the common gateway interface (CGI), Active Server Pages (ASP), or other methods. An intranet in a radiology department could include the following types of content: on-call schedules, work schedules and a calendar, a personnel directory, resident resources, memorandums and discussion groups, software for a radiology information system, and databases.

Index Terms: Computers • Internet • Intranet


    INTRODUCTION
 Top
 Abstract
 INTRODUCTION
 WHAT IS AN INTRANET?
 DEPARTMENT OF RADIOLOGY PROFILE
 HARDWARE REQUIREMENTS
 SOFTWARE REQUIREMENTS
 KNOWLEDGE REQUIREMENTS AND...
 CONTENT DEVELOPMENT
 DEVELOPMENT PROCESS
 CONCLUSIONS
 References
 
With the latest desktop computers and World Wide Web software, any organization can now develop its own intranet quickly and with relative ease. Web technology is evolving rapidly, bringing a steady stream of exciting new applications and capabilities to both the developer and the end user.

This article describes the implementation of an intranet in a large academic radiology department and discusses the advantages of an intranet for information management within an organization. First, the term intranet is defined and the radiology department at our institution is profiled. The reader is then introduced to the process of intranet development, which begins with the basic hardware and software requirements, progresses through knowledge requirements and programming, and culminates in content development. The roles of server- and client-side scripting are discussed, and emphasis is given to the use of intranets to access and maintain databases. Finally, the development process for the intranet implemented in our department is described. Human and organizational issues pertinent to acceptance and maintenance are also addressed.


    WHAT IS AN INTRANET?
 Top
 Abstract
 INTRODUCTION
 WHAT IS AN INTRANET?
 DEPARTMENT OF RADIOLOGY PROFILE
 HARDWARE REQUIREMENTS
 SOFTWARE REQUIREMENTS
 KNOWLEDGE REQUIREMENTS AND...
 CONTENT DEVELOPMENT
 DEVELOPMENT PROCESS
 CONCLUSIONS
 References
 
An intranet is a "private Internet" that uses the protocols of the Web to share information resources within a company or with the company's business partners and clients (1). The Web protocols used primarily involve Web pages (hypertext markup language [HTML] documents) transmitted by means of the hypertext transfer protocol (HTTP); however, other Internet protocols, of which there are many, can be incorporated according to need. With traditional forms of information management, information resides in printed form—in bulging file cabinets, stacks of computer printouts, guides, logs, ledgers, and corporate manuals—or, if stored electronically, remains dispersed on hard drives and floppy disks throughout the company. Information so stored may be difficult to locate and retrieve, is available only to those with physical access, and cannot be readily updated. Information that must be widely and quickly accessible (such as schedules or telephone numbers) is usually distributed on paper, which is tacked up or taped up, sometimes torn, and often left hanging long after its useful life has ended. These pieces of paper soon grow out of date and usually end up as unsightly "scratch pads." The process of creating, duplicating, distributing, and replacing paper postings is time-consuming and wasteful and can never keep up with changes or with the individual needs of diverse areas within a radiology department.

With networked electronic data storage, given proper planning and management, information may be retrieved in seconds from anywhere in the world by using a simple personal computer and a link to the Internet. Although sharing information on a network is not a new concept, the intranet phenomenon has driven networked data management to new levels of cross-platform compatibility and simplicity for the end user. The ability to access many databases and applications by means of a single user interface—the Web browser—provides an appealing alternative to the installation and mastery of multiple proprietary client interfaces. The Web browser (eg, Netscape) has become a ubiquitous interface by now familiar to most computer users (Table 1).


View this table:
[in this window]
[in a new window]

 
Table 1. Popular Web Browsers
 
The Internet, and thus most intranets, are based on the transmission control protocol (TCP)/Internet protocol (IP) stack, a standardized set of protocols for data communication that allows computers built of different hardware and running different software on different operating systems to communicate with one another in a common language. This property, known as platform independence, allows deployment of an intranet as a company information resource while enabling access from all existing desktop computer systems.

Intranets follow the client-server model of networking, in which most of the data and Web content reside on dedicated computers called servers. The term server describes a system's function, not its hardware type, because even the humblest desktop computer can function as a server so long as it resides on a network. The client is any remote system accessing that server's content (Fig 1). The principal difference between an intranet and the Internet is that an intranet is not intended for public access. An intranet may carry sensitive employee or financial information and thus may require a very high level of privacy and security. Access to the server can be restricted by means of password protection or the client's IP address (a unique 4-byte numeric code [eg, 123.45.67.89] that identifies a machine connected to the network), and transmitted data can be encrypted (usually by means of the Secure Sockets Layer [SSL] protocols) to prevent interception in transit.



View larger version (23K):
[in this window]
[in a new window]
[Download PPT slide]
 
Figure 1.  Basic structure of an intranet. The server receives HTTP requests from the client and returns HTML-formatted pages to the client.

 
It is the fact of easy, instantaneous access to information that has driven the proliferation of corporate intranets. However, the advantages of the intranet paradigm extend to management of the information as well. With an intranet, information can be added and updated from anywhere with the same ease with which it is accessed. Once updated, the correct information is immediately accessible to anyone anywhere without the delays of reprinting and redistribution. Nor are the benefits restricted to data and documents. Intranets can be used to distribute a wide variety of information media including graphics, audio, video, and software. Basic Web technology can facilitate intradepartmental research collaboration by means of discussion groups, group scheduling, and document sharing. Emerging new technology enables real-time collaboration by means of "white boards," chat rooms, Internet telephony, or video conferencing.


    DEPARTMENT OF RADIOLOGY PROFILE
 Top
 Abstract
 INTRODUCTION
 WHAT IS AN INTRANET?
 DEPARTMENT OF RADIOLOGY PROFILE
 HARDWARE REQUIREMENTS
 SOFTWARE REQUIREMENTS
 KNOWLEDGE REQUIREMENTS AND...
 CONTENT DEVELOPMENT
 DEVELOPMENT PROCESS
 CONCLUSIONS
 References
 
The core facilities of the University of Alabama at Birmingham Health System include a tertiary-care, 750-bed university hospital and the Kirklin Clinic, a 450,000–sq ft (40,500-m2) outpatient facility. There are about 35,000 hospital admissions and more than 330,000 outpatient visits annually to the approximately 650 faculty physicians employed within the medical center.

Over 300,000 examinations are performed per year within the Department of Radiology, and the results of over two-thirds of these studies are processed, distributed, and archived digitally. A full-time faculty of about 40 radiologists and six physicists encompasses every major radiologic specialty; there is a training complement of 20 residents and about 10 fellows. Besides the intranet described herein, we have a public Web site at http://www.rad.uab.edu that describes our medical center, department, faculty, and programs in detail and contains one of the largest online radiology teaching files available.


    HARDWARE REQUIREMENTS
 Top
 Abstract
 INTRODUCTION
 WHAT IS AN INTRANET?
 DEPARTMENT OF RADIOLOGY PROFILE
 HARDWARE REQUIREMENTS
 SOFTWARE REQUIREMENTS
 KNOWLEDGE REQUIREMENTS AND...
 CONTENT DEVELOPMENT
 DEVELOPMENT PROCESS
 CONCLUSIONS
 References
 
An established local area network of interconnected desktop computers and servers is a prerequisite for intranet development. In a typical large medical center, many operating systems and hardware types are represented. Workstations may include PC-compatible machines, Macintoshes (Apple Computer, Cupertino, Calif), UNIX workstations, and "network computers." Workstations and servers can be connected with an assortment of hardware, which might include Ethernet (10 Mbit/sec), fast Ethernet (100 Mbit/sec), token ring, and asynchronous transfer mode (ATM). The connection hardware offers a common set of services to the Internet protocols, which in turn are designed to function independently of the hardware and software servicing them.

The hardware requirements for an intranet begin with a dedicated Web server permanently connected to the departmental network. The network should offer domain name service (DNS), which converts easily remembered alphabetic addresses to the numeric IP address required to actually find another computer on the network and establish a connection. Hubs and routers are required for interconnectivity and traffic control. Although a Web server can be run on almost any modern desktop computer, certain features help optimize server performance: (a) The central processing unit (CPU) should be as fast as possible. (b) System memory (random-access memory [RAM]) is inexpensive, and the more the better. (c) Large hard drives may be required for graphics storage and bulky operating systems and should be as fast as possible for rapid multiuser access. A tape drive or redundant disk array can provide data backup and security. (d) A network adapter is an absolute requirement; 10-Mbit/sec Ethernet is the de facto minimum standard. (e) The server must have a fixed IP address so that clients can find it on the network and preferably a registered alphabetic name, such as www.intranet.web, for the convenience of users. (f) Although a compact disk, read-only memory (CD-ROM) drive is not required for operation, it is almost invariably necessary for installation of most modern software applications.

On the other hand, large monitors, advanced video cards, and multimedia enhancements contribute nothing to server performance. The desktop computer originally used as our server ran on a Cyrix 586-P166 processor (National Semiconductor, Santa Clara, Calif) with 32 Mbytes of RAM, a 1.2-Gbyte hard drive, and a 10-Mbit/sec Ethernet adapter. (We have since upgraded to a 300-MHz Pentium II system [Intel, Santa Clara, Calif] with 128 Mbytes of RAM, a 6.0-Gbyte hard drive, and a 100-Mbit/sec Ethernet adapter.) A digital camera is a useful accessory, which we used to acquire portraits and import them into a set of pictorial directories, bypassing the steps of processing film and scanning in the results.


    SOFTWARE REQUIREMENTS
 Top
 Abstract
 INTRODUCTION
 WHAT IS AN INTRANET?
 DEPARTMENT OF RADIOLOGY PROFILE
 HARDWARE REQUIREMENTS
 SOFTWARE REQUIREMENTS
 KNOWLEDGE REQUIREMENTS AND...
 CONTENT DEVELOPMENT
 DEVELOPMENT PROCESS
 CONCLUSIONS
 References
 
The software requirements for an intranet are straightforward. The operating system (eg, MacOS [Apple Computer], Windows [Microsoft], IntranetWare [Novell, San Jose, Calif], UNIX) should be in its most current release with all of the required networking support. The heart of a Web server is the HTTP service, which receives a page request from a remote client's browser and transmits the page back to the correct client. Individual applications are required to run the HTTP server (Table 2), to author Web content, to create and maintain databases, and to create graphics. Some integrated applications, such as Microsoft FrontPage, incorporate both page editors and the HTTP server into a package that can have a new Web site up and running in less than an hour. Once the intranet becomes "mission critical," development work should be relocated to a peripheral system and new content uploaded to the server only after testing and debugging. The reason behind this strategy is that development activities consume processor time, and many of the applications used for development may precipitate system "crashes" or unexpectedly alter the network configuration. The use of "beta" software allows development at the cutting edge of Web technology, but one is advised to avoid beta software for deployment of critical services.


View this table:
[in this window]
[in a new window]

 
Table 2. Web Server Applications
 
For our intranet, we use the Windows NT 4.0 operating system. Windows NT Server comes bundled with Internet Information Server, a powerful and relatively easy-to-use application that can run file transfer protocol (FTP) service as well as HTTP service. Internet Information Server version 4.0 or higher has the capability to generate pages from a database "on the fly" in response to user queries (see "Programming: Data Access").

High-quality Web publishing can also be run on the MacOS platform. A majority of the high-end commercial Web servers are running on UNIX platforms, and a free UNIX application named Apache is by far the most common Web server application deployed on the Internet. A Web server written completely in Java has been released by Sun Microsystems, and a Java operating system was recently developed by a joint partnership between International Business Machines and Sun Microsystems.


    KNOWLEDGE REQUIREMENTS AND PROGRAMMING
 Top
 Abstract
 INTRODUCTION
 WHAT IS AN INTRANET?
 DEPARTMENT OF RADIOLOGY PROFILE
 HARDWARE REQUIREMENTS
 SOFTWARE REQUIREMENTS
 KNOWLEDGE REQUIREMENTS AND...
 CONTENT DEVELOPMENT
 DEVELOPMENT PROCESS
 CONCLUSIONS
 References
 
Until recently, authoring Web pages required a working familiarity with the HTML syntax, which, although not difficult to learn, can be quite tedious. If one does not object to typing in all of the tag information, Web pages can be authored by using any text editor. Fortunately, nowadays there are several high-quality WYSIWYG (what you see is what you get) editors for page authoring that do not require knowledge of HTML (Table 3). (With WYSIWYG, the page theoretically appears the same in the editor as it will when viewed in the browser.) Some of these editors are compatible with other office products and allow easy cutting and pasting of documents, images, and spreadsheets from other applications. Still, we occasionally find it necessary to edit the HTML directly to perfect the final appearance or to correct various errors.


View this table:
[in this window]
[in a new window]

 
Table 3. WYSIWYG HTML Authoring Applications
 
With modern HTML editors, anyone who can use a word processor can create static Web pages. To take advantage of the many advances in database access, live content, and customization of sites, more advanced skills are required. The first objective should be to become familiar with the HTML syntax. Managing more complex Web sites requires an assortment of skills, including programming (scripting languages and Java) and database management (especially Structured Query Language [SQL]). A working familiarity with at least one Web-enabled database application such as Access (Microsoft) or FileMaker Pro (FileMaker, Santa Clara, Calif) is particularly useful.

Programming: Scripting, Java, and SQL
Programming skills are required for many advanced intranet functions. Scripts are small programs contained within the HTML document but invisible to the user. Scripts can be designed to run on the client or the server. (Client-side scripts [scripts run inside the client's browser] can be seen by choosing "view source" options from the browser menu.) JavaScript (Fig 2) was developed by Netscape Communications and will run on all major browsers. VBScript (Fig 3) is another Microsoft product and is closely related to Visual Basic. VBScript will run only in Internet Explorer but has the advantage that it can employ a wide variety of "ActiveX" controls (Figs 46) that enhance page appearance and functionality. Sites that make use of ActiveX controls, including our own, must provide alternative means of delivering content to non-Windows systems to maintain platform independence. It is a simple matter for the Web server to recognize which browser a client is using and return the proper content according to the browser and operating system. One of the more common and practical examples of scripting is form-field validation. Form-field validation prevents the client from submitting a form when required fields are blank or when text is entered where numbers are required. Incorporating validation onto the page reduces unnecessary network traffic as well as the load on the server.



View larger version (22K):
[in this window]
[in a new window]
[Download PPT slide]
 
Figure 2. Figures 2, 3. (2) Sample of JavaScript from a calendar-generating program. (3) Sample of VBScript from a calendar-generating program.

 


View larger version (34K):
[in this window]
[in a new window]
[Download PPT slide]
 
Figure 3. Figures 2, 3. (2) Sample of JavaScript from a calendar-generating program. (3) Sample of VBScript from a calendar-generating program.

 


View larger version (45K):
[in this window]
[in a new window]
[Download PPT slide]
 
Figure 4. Figures 4–6. (4) ActiveX Popup Menu control. RIS = radiology information system, SSA = Systems Administration. (5) ActiveX Spinner controls. (6) ActiveX Calendar control. This control was programmed so that the user can obtain personnel on call for any date—past, present, or future—by simply clicking on the date.

 


View larger version (100K):
[in this window]
[in a new window]
[Download PPT slide]
 
Figure 6. Figures 4–6. (4) ActiveX Popup Menu control. RIS = radiology information system, SSA = Systems Administration. (5) ActiveX Spinner controls. (6) ActiveX Calendar control. This control was programmed so that the user can obtain personnel on call for any date—past, present, or future—by simply clicking on the date.

 
Java is a complete programming language similar to C++. The vision of its developers is that Java programs will be platform independent so that a single version will run on all systems. This platform independence is accomplished by means of the Java Virtual Machine (JVM): software written for each specific combination of browser and operating system and able to execute Java programs on any client so equipped. The JVM is installed automatically with current versions of Netscape browsers and Internet Explorer. Java programs can be stored on the server and downloaded via the Web to run on the client as needed.

SQL is a database management language and the most popular standardized tool for querying and managing databases.

Programming: Data Access
Most of the major database developers have integrated Web server connectivity into their newer releases to allow data to be displayed and managed remotely via a Web browser. Several integrated development tools for creating Web-database connectivity are on the market (Table 4). The key features to look for are cost, ease of use, suitability to scale, and compatibility with the user's existing applications.


View this table:
[in this window]
[in a new window]

 
Table 4. Web-Database Development Applications
 
Web-based database management presents particular challenges to the basic HTTP. The first problem relates to security and the fact that HTTP transactions are stateless. In this context, the term stateless means that an active TCP connection exists between the client and the server only for the brief time required for the client to transmit an HTTP request and the server to return its response. (In the hierarchical models of networking, the TCP occupies the transport layer, one level above the network layer of the IP. Whereas the IP is used to move data from one machine to another, TCP connections are between applications and enable data to get to the right application on the client and the host.) The TCP connection closes immediately on completion of the transaction. Brief connection times mean that Web servers can service vast numbers of clients in a given period. The disadvantage is that the server has no memory of the transaction when a new request comes in from the same client. Typically, in an intranet as well as many commercial Internet sites, the host would like to restrict data access to authorized users, who must pass security by sending a user name and password with the HTTP request. However, it would be inconvenient for the client to have to type in log-on information repeatedly for each page request.

Several solutions have been developed to enable a Web server to maintain a "memory" of client connections so as to simulate persistent connections in a stateless environment. The first method employs scripts on the server to write information (eg, the user's identification and password) to hidden fields in the HTML page returned to the client. By appending these hidden fields to forms on the page, this information can be returned to the server in subsequent client requests in a process invisible to the user (2).

A second and more recent method employs "cookies." Cookies are small text files saved to the client's hard drive by the server. They are typically stored in the same default directory and can include fields indicating the user's name, identification, password, preferences, and expiration date and time (Fig 7). When a cookie is present on the client's hard drive, a Web page loaded in the client's browser can retrieve the cookie's content by means of scripting and return to the server the data needed to allow a log-on or to return a personalized page.



View larger version (20K):
[in this window]
[in a new window]
[Download PPT slide]
 
Figure 7.  Example of a cookie opened in text editor.

 
The second major challenge in Web-enabled database management relates to interfacing databases and database management systems with the Web server and conforming the user interface to HTML syntax. Integration of databases with the Web server has been accomplished by several means. The first, and still the most prevalent, is the common gateway interface (CGI). The CGI provides a standard means by which programs other than simple HTTP transactions can be run by the Web server. CGI programs can be written in a variety of languages, of which Practical Extraction and Reporting Language (PERL) has been the most popular. However, the CGI has a steep learning curve, and fairly simple programs may require hundreds of lines of code. Another disadvantage is that each new CGI request creates ("spawns") a new process on the server. The cumulative effect of many CGI requests in a short time places significantly greater load on the host system than do newer alternatives.

One of several newer approaches to database integration, known as Active Server Pages (ASP), was developed by Microsoft to run on Internet Information Server but can be enabled on Netscape servers as well. (Technically, ASP is made possible by the Internet Server Application Programming Interface [ISAPI]. This is a set of programmable objects and interfaces exposed or recognized by Internet Information Server. Counterparts of the ISAPI are the Netscape Server Application Programming Interface [NSAPI] for Netscape servers and the Oracle Web Request Broker [OWRB], which runs with a variety of Web servers.) ASP programs are scripts written in JavaScript or VBScript. The program is saved as a plain text file with the extension ".asp" (eg, "page.asp"). The "asp" extension indicates to the server that a script is to be processed rather than an HTML document sent. The ASP script generates HTML dynamically in response to parameters passed from the client, such as query terms or personal profiles stored in cookies. Integrated development environments such as Visual InterDev (Microsoft) enable generation of highly sophisticated ASP programs rapidly and with relative ease. The final integrated process is summarized in Figure 8. Most of our intranet-based data functions use the ASP method.



View larger version (23K):
[in this window]
[in a new window]
[Download PPT slide]
 
Figure 8.  Overview of Web-based data management. The HTTP server and server script application run on the server operating system. The database can reside on the Web server or any other server on the network. The HTTP server communicates with the script application via the application programming interface (API); the server passes on query parameters and receives back data formatted within an HTML document. The script application communicates with the database (eg, Access, Oracle) via Object DataBase Connectivity (ODBC) or Java DataBase Connectivity (JDBC). The script application receives search parameters from the HTTP server, converts them into a proper SQL query, and passes the query to the database. The database returns the query results in SQL, which the script application must format into HTML by means of some predefined template or rules.

 
Many other server-side programming and database solutions are available. One that has generated significant interest among health care application programmers involves the use of Java and the common object request broker architecture (CORBA). In the words of the Object Management Group (OMG), "CORBA allows applications to communicate with one another no matter where they are located or who has designed them" (3). By using the Internet Inter-ORB [object request broker] protocol (IIOP), distributed applications using CORBA can communicate across the Internet to share both data and applications. CORBA-based data delivery may be within the framework of a Web browser but is not limited to this interface. (Remember, the IP supports a number of higher-level protocols of which the familiar HTTP of the Web is only one.)

At the opposite extreme to the server-side models we have discussed, it is possible to download to the client's browser a dedicated program for direct database access by using Java or an ActiveX plug-in. (A plug-in is a small application that is stored on the client computer and adds functionality to the Web browser.) This approach is termed the fat client approach because a sizable program may be downloaded rather than a "lean" HTML document. Once running, the client application can use any networking protocol and data format the programmers desire. For example, on our intranet we have tested a Java-based terminal emulator for accessing the RIS. One disadvantage to such an approach is that it requires users to learn to use a nonstandard interface. In our RIS implementation, learning to use a nonstandard interface is not such a problem because users must use the same interface for routine RIS access "outside" the intranet.


    CONTENT DEVELOPMENT
 Top
 Abstract
 INTRODUCTION
 WHAT IS AN INTRANET?
 DEPARTMENT OF RADIOLOGY PROFILE
 HARDWARE REQUIREMENTS
 SOFTWARE REQUIREMENTS
 KNOWLEDGE REQUIREMENTS AND...
 CONTENT DEVELOPMENT
 DEVELOPMENT PROCESS
 CONCLUSIONS
 References
 
Once the hardware and operating systems have been installed and a Web server is operational, the development effort shifts to the more creative side of intranet implementation: developing content. The possibilities are limited more by time and imagination than by technology. Suggestions for content are listed in Table 5. Our intranet incorporates most of the items listed.


View this table:
[in this window]
[in a new window]

 
Table 5. Potential Content
 
Platform Independence: Limits and Trade-offs
Although the ideal of platform independence is a valued benefit of Web technology, in actual deployment there are many exceptions. Certainly, the core networking technologies are platform independent because the various Internet protocols have become codified by international standards organizations. An example is the W3 Consortium (www.w3.org), which formalizes HTML coding conventions. However, the fact is that incompatibilities exist between the major browsers, client platforms, and available Web technology. Java applets and programs may be touted as "write once, run anywhere," but to date Java programs have exhibited significant performance disparities on differing platforms and occasionally fail to run at all. A case of platform "dependence" is the ActiveX technology, which is based on the Windows platform. The ActiveX technology permits programs to be downloaded and run through a browser, similar to Java, but with the added ability to access built-in Windows functions and without the security restrictions against disk access of Java. Developers may be forced to choose between true platform independence and enhanced functionality. Developers of public Web sites must be sure to provide content independent of the client's browser or develop duplicate resources customized to the two leading Web browsers. A private intranet is not so constrained, and some organizations (not ours) simply choose to abandon platform independence and decree a company-wide platform so that they can standardize the interface or deploy ActiveX and other proprietary technologies.

Resources Offered on Our Intranet
On-Call Schedules.—Accurate and timely on-call schedules are crucial because such information is frequently needed urgently and is constantly revised as individuals trade on-call days. Our traditional system used various lists posted in the radiology library and given to the hospital paging operators. Whenever a change was made, someone was responsible for crossing through and replacing the old name. The obvious limitations of this system in a large center such as ours are that it is often difficult to obtain updated information and considerable time might elapse before the change is recorded. With an intranet, this information can be accessed or updated immediately at any computer. In addition, the hospital call center, which handles most pages of on-call staff, can use this system. Our intranet displays on-call schedules for faculty (general radiology, imaging, neuroradiology, and interventional), residents, angiography nurses and technicians, and support personnel (administrators, service engineers, RIS support staff, and picture archiving and communication system [PACS] support staff). Some on-call schedules are saved as static tables, which are periodically edited; however, most schedules are in a database, which can be maintained via any browser.

For various reasons, we chose to make the resident on-call schedule the first to go completely online. The resident on-call schedule for the entire academic year is entered into a database via a browser. The schedule can be queried by day, person, or type of call. Any user can enter a change simply by clicking the "update" link.

Work Schedules and Calendar.—Work assignments for faculty, fellows, and residents are prepared in different offices and updated at different times. These schedules include the neuroradiology, general radiology/body imaging, and resident rotation schedules. The general radiology schedule is generated by means of a proprietary commercial scheduling program that uses a Microsoft Access database and prints out a large, complex table. By carefully analyzing the database structure and employing ASP technology, we were able to make the schedule browser accessible, which allows customized views by name, date, or assignment.

The calendar includes events of interest to the department such as conferences, meetings, and social events, which are linked to online maps when appropriate. Using a scripted program, the user can select any month of any year to be displayed alongside the list of events.

Personnel Directory.—As our department has grown and members have come and gone, keeping track of frequent changes has become daunting. Previously, multiple printed lists of telephone numbers, fax numbers, e-mail addresses, office locations, status changes, and so on were maintained in separate administrative offices. We developed a customized, Web-accessible directory database that unifies all of these functions and eliminates the need to periodically reprint and distribute all of these lists. The intranet always returns the most up-to-date information and allows a user to target only the desired information without sifting through numerous directories. Updates are made by the intranet project manager via a customized browser interface. Private information such as home addresses or telephone numbers can be accessed for administrative use but remains secure from unauthorized users. The database also includes hospital telephone numbers useful to departmental employees, such as those of nursing units, laboratories, and clinics.

The directory can be searched by entering any portion of a name. A list of names containing the entered text string is displayed, and the user can click on the correct name to view basic information such as position, office location, telephone numbers, pager number, e-mail address, and photograph. To send an e-mail message, the user need only click on the e-mail address to automatically open his or her mail editor, which is already addressed to the recipient.

Resident Resources.—The resident section includes digital pictorial directories of residents and fellows, an online version of the resident manual, specific guides to clinical rotations, and other items of interest. Some faculty have used this site to place conference handouts online rather than printing and distributing multiple paper copies.

Memorandums and Discussion Groups.—In another section, memorandums of general interest could be entered into a form and posted to the intranet. This option was rarely used because distribution by means of e-mail provides the added benefit of recipient notification. In departments without universal e-mail, such a function might be more widely used. This site was recently redesigned as a "What's New?" site, which allows individuals to easily post announcements of general interest.

Another area provides for discussion groups on user-determined topics. So far, this site has been little used because of the ease of e-mail or direct face-to-face communication.

RIS Support.—In 1996, we developed in-house a customized application to enable report editing and approval on Windows 95 clients, a move necessitated by the hospital-wide replacement of old MS-DOS (Microsoft) systems with new Windows 95 workstations. Rather than trying to copy and circulate the application on multiple sets of floppy disks, we placed the RIS program (~4 Mbytes) and instructions for installation and use on the Web site. The software could then be easily and reliably downloaded and installed on any computer on the network or even at home.

Databases.—Using the methods described in the section titled "Programming: Data Access," we have placed many databases online. The personnel directory has already been described. We developed another database for maintaining on-call schedules. A click on the "Who's on Call?" link causes the server to query this database on request and display all on-call personnel for the current date and time. On-call schedules for faculty, residents, technologists, administrators, and nurses are included. Designated personnel maintain these databases over the intranet via customized Web interfaces.

By means of our intranet, a wealth of clinical data has been made available to "mine" for research purposes. In a recent project, we imported over 1 million radiology reports from 1992 to 1997 into a customized database, which enables our residents and faculty to retrieve cases by querying the report text for a particular diagnosis or search term (4). Once this database was complete, we developed a browser interface and placed the database online with password-restricted access. We are in the process of building a database of computed tomographic and magnetic resonance imaging protocols to be placed online. Once this database is made the "system of record," it will ensure that technologists and residents always have the most up-to-date techniques at hand.


    DEVELOPMENT PROCESS
 Top
 Abstract
 INTRODUCTION
 WHAT IS AN INTRANET?
 DEPARTMENT OF RADIOLOGY PROFILE
 HARDWARE REQUIREMENTS
 SOFTWARE REQUIREMENTS
 KNOWLEDGE REQUIREMENTS AND...
 CONTENT DEVELOPMENT
 DEVELOPMENT PROCESS
 CONCLUSIONS
 References
 
Project Inception
Our intranet was begun by one of our faculty members (S.J.W.) employing Microsoft FrontPage running in Windows 95 on his own desktop computer. As is typical of such projects, the scope expanded over time as ideas and suggestions came in from many quarters and the technology for the server hardware and software advanced. As the potential became apparent for this system to replace labor-intensive functions and improve access to accurate, up-to-date information, higher levels of administration came to support development of the system and assist in integrating it with departmental functions.

Platform Selection
A Windows platform was chosen by the project developer, who had considerable experience programming in the Windows environment, was acquainted with the operating system and Windows-based Web applications, and had a Windows workstation immediately available. Present alternatives would include the MacOS platform, Novell IntranetWare, or some variety of UNIX (at least 16 different versions are in wide use, many of which can run on PC-compatible and Macintosh systems). At about this time, the information services department had decided on Windows 95 as the common hospital-wide platform and we were at the beginning of a major rollout of new workstations throughout the medical center.

We suggest that the primary criteria in choosing a platform for the intranet service are experience and cost. Whoever assumes the role of developer should begin with whatever platform is most familiar and accessible. Material costs can be insignificant if the intranet can be implemented on existing hardware. Conversely, software costs can vary greatly. A complete intranet can be deployed with free software by using Linux (a free version of UNIX) and the Apache Web server, but the learning curve is quite steep, and this would be a poor choice for a newcomer to UNIX. Large discounts and even free software are often available to academic sites, so the cost of software may differ significantly for academic versus commercial sites according to the package selected. For institutions that do not have the programming resources or cannot commit them to a project such as this, commercial Web developer services are widely available. Because we receive substantial academic discounts on software, the total project cost has remained under $5,000, most of which was for hardware upgrades and replacement. All of the programming has been done in-house by the project developer using many of the tools described in this article.

For very busy sites, server performance can become an issue. All configurations are not equal, and real performance disparities exist between competing platform alternatives. Performance comparisons are available from a variety of sources, and someone who is about to install a new system might consider this information before committing to a particular configuration.

Implementation Process
Because this project began as a project of the senior author, the original server was his desktop computer. As the intranet became more widely used, interruptions in service caused by development activity or alternative use became unacceptable and improved performance was desired. At that point, a new system was purchased and set aside as an intranet server.

The clearest economic benefit of an intranet comes from the permanent replacement of inefficient manual processes, but such a move requires greater commitment and effort than the implementation of an intranet as just another means of accessing information. We consider the following strategic factors key to deployment of an intranet as a permanent replacement for manual processes: (a) the support and commitment of management; (b) technical reliability based on fault-tolerant hardware and software; (c) a system for technical support on a 24 hours per day, 7 days a week basis; (d) a self-sustaining system for updating content; (e) provision for training and support of users; (f) promotion of ease of use for the end users; and (g) a gradual transition from manual processes to the intranet rather than an abrupt switch to intranet use.

To accomplish these objectives, we created a team of technical, administrative, and support personnel. The strategic factors were addressed as follows:

1. Administrative support and commitment had already been obtained before the team was formed.

2. Technical reliability was enhanced by acquiring new, higher-performance server hardware and software.

3. Creating a plan for technical support required that individuals besides the developer be assigned to support the system and that the server itself be accessible for maintenance. A schedule was established so that someone would always be available to service a malfunction.

4. In terms of updating content, we tried to preserve existing roles when we moved from a paper-based to a Web-based system. For example, schedules for various subsections have been independently maintained by respective section personnel. Rather than centralize the process, we trained the appropriate people to maintain their schedules on the intranet server and designated an "intranet system manager" to monitor that content was kept up-to-date. Otherwise, obsolescence of data on the intranet would be almost inevitable because of the tendency of some individuals to revert to familiar manual practices.

5. If an organization expects its members to use an intranet as the system of record for important information, then the members deserve training and support as regards access and use of the system. Group training (formal or informal) supplemented by written communications encourage use of the system and help monitor the system's usefulness. User feedback should be sought on desired features and the quality of the interface. Rather than holding formal training sessions, we chose to communicate by means of existing channels such as resident conferences and staff meetings. Team members provide informal support for those who need assistance. Residents with computer skills have been particularly helpful in this area because they have broad and frequent contact with radiology personnel.

6. As for ease of use, people accustomed to computer use usually have no difficulties learning to navigate the intranet. Whereas problems regarding ambiguous terminology or advanced features are easily corrected by redesign, barriers to use of the system are often purely psychologic. Yet, these too may be surprisingly minor and easily overcome. For example, users may be unaware that sites can be "bookmarked" in a browser, eliminating the need to retype the site address each time. Users can be shown how to create a desktop icon to provide a shortcut to the site or how to leave the browser running in the background, minimizing the time required to access the site. When access requires a double click of a desktop icon with a 1–3-second response time, resistance to use of the system often dissolves.

7. Because an intranet can affect so many aspects of routine operation and involves so many people in different roles, implementing features incrementally rather than abruptly has been effective for us. However, with manual processes in place, some individuals feel no urgency to convert, and occasional encouragement and even insistence are helpful. The continued presence of paper records creates a "why bother to learn" attitude among some people. Conversely, overwhelming people with an abrupt change may foster resistance that is even more difficult to overcome.


    CONCLUSIONS
 Top
 Abstract
 INTRODUCTION
 WHAT IS AN INTRANET?
 DEPARTMENT OF RADIOLOGY PROFILE
 HARDWARE REQUIREMENTS
 SOFTWARE REQUIREMENTS
 KNOWLEDGE REQUIREMENTS AND...
 CONTENT DEVELOPMENT
 DEVELOPMENT PROCESS
 CONCLUSIONS
 References
 
The intranet paradigm has become established as a practical, powerful, state-of-the-art method for information management within organizations of all sizes. The technology required is highly scalable, and it is possible to implement an intranet for even a small department or company by using inexpensive and readily available hardware and software. Our intranet continues to evolve and become more integrated into the routine functioning of our department.



View larger version (74K):
[in this window]
[in a new window]
[Download PPT slide]
 
Figure 5. Figures 4–6. (4) ActiveX Popup Menu control. RIS = radiology information system, SSA = Systems Administration. (5) ActiveX Spinner controls. (6) ActiveX Calendar control. This control was programmed so that the user can obtain personnel on call for any date—past, present, or future—by simply clicking on the date.

 


    Footnotes
 
Address reprint requests to S.J.W.

Presented as an infoRAD exhibit at the 1997 RSNA scientific assembly.

Abbreviations: ASP = Active Server Pages HTML = hypertext markup language HTTP = hypertext transfer protocol IP = Internet protocol RIS = radiology information system SQL = Structured Query Language TCP = transmission control protocol WYSIWYG = what you see is what you get

Received for publication May 6, 1998. Revision received June 12, 1998. August 24, 1998. Accepted for publication September 8, 1998.


    References
 Top
 Abstract
 INTRODUCTION
 WHAT IS AN INTRANET?
 DEPARTMENT OF RADIOLOGY PROFILE
 HARDWARE REQUIREMENTS
 SOFTWARE REQUIREMENTS
 KNOWLEDGE REQUIREMENTS AND...
 CONTENT DEVELOPMENT
 DEVELOPMENT PROCESS
 CONCLUSIONS
 References
 

  1. Channin DS, Chang PJ. Primer on computers and information technology. III. Internets and intranets. RadioGraphics 1997; 17:1293-1296.
  2. Orfali R, Harkey D. Client/server programming with Java and CORBA New York, NY: Wiley, 1997; 226-227.
  3. ; http://www.omg.org..
  4. Willing SJ, Bream PR, Jr, Smith JK. Enabling free text searches of RIS report database (abstr). Radiology 1998; 209(P):679.



This article has been cited by other articles:


Home page
RadioGraphicsHome page
A. G. Ryan, L. J. Louis, and W. C. Yee
Informatics in Radiology (infoRAD): HTML and Web Site Design for the Radiologist: A Primer
RadioGraphics, July 1, 2005; 25(4): 1101 - 1118.
[Abstract] [Full Text] [PDF]


This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Willing, S. J.
Right arrow Articles by Berland, L. L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Willing, S. J.
Right arrow Articles by Berland, L. L.
Related Collections
Right arrow Informatics


HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOGRAPHICS RADIOLOGY RSNA JOURNALS ONLINE