For an understanding of our position, it is important to trace the developments that have led to the recent avalanche of Internet use. The Internet is an outgrowth of the U.S. Department of Defense network ARPANET (Advanced Research Project Agency network), which was introduced in 1969 [4]. This project spawned the development of a standard protocol for network communication known as TCP/IP (transmission control protocol/Internet protocol) [5]. This protocol is currently used to connect all Internet computers. The expansion of the Internet was initially slow, growing from four facilities at ARPANET's inception to a mere 235 hosts (defined as computers on the Internet that can provide services) by May 1982 [6]. By 1986, when the National Science Foundation formed NSFNET, 2308 Internet hosts had been registered [6]. Eventually, NSFNET and ARPANET merged and opened connections to universities and private corporations, thereby forming the foundation of today's Internet.

After this, the Internet rapidly expanded. By the end of 1989, more than 150 000 hosts provided services for more than 2 million users [6]. In 1991, members of the European Laboratory for Particle Physics introduced the World Wide Web (WWW), a protocol for the exchange of hypertext and multimedia (digital images, audio, and video) through the Internet [7]. Then, in late 1993, the U.S. National Center for Supercomputing Applications released Mosaic software free of charge for computers running Unix, Microsoft Windows, and Macintosh operating systems. Mosaic provided a graphic-based user interface to what had previously been the text-based environment of the Internet. This software was recognized by The New York Times as a novel computer application capable of revolutionizing the way people use the Internet [8]. Today, many WWW navigational programs—known as browsers—are available [9]. Together, these browsers and the WWW have ignited the most recent explosion of Internet activity.

Although the Internet has provided a way to connect all of our computers, training for the use of such technology among medical professionals is lacking. As a result, our use of the Internet to exchange medical information has been sluggish. In 1994, the National Information Infrastructure Interagency Task Force stated that "the health sector. lagged far behind the other sectors of our economy in applying information and communication technologies" [10].

In contrast, the importance of the Internet as a way to exchange medical information is clearly recognized by our government. The National Institutes of Health [11], the National Cancer Institute [12], the National Library of Medicine [13], the National Institute of Environmental Health Sciences [14], the Centers for Disease Control and Prevention [15], and the Food and Drug Administration [16] all use the Internet to publish vast amounts of health-related information. Physician Data Query, a synthesis of cancer information distributed over the Internet by the National Cancer Institute, is a prime example. This database, created "to help physicians cope with the information explosion by translating the medical literature into usable forms," has been available since 1984 and on the Internet since 1994 [17].

Recently, universities, medical centers, and even patients have become more prominent "publishers" of medical information on the Internet [1]. Some have chosen to distribute information on a specific field of medicine intended for physicians and the public [18], whereas others have developed resources exclusively for physicians [19]. We support these early endeavors and hope that others follow.

Several factors are contributing to the apparent increase in health-related Internet publishing. Direct Internet connections are available in many academic centers, some at no charge to faculty and students. Public and commercial Internet connections are widely available for reasonable fees, thereby opening the door to persons outside of academia. With the use of inexpensive or free software, it is possible to become a host to gopher, telnet, FTP (file transfer protocol), and WWW services on the Internet with an entry-level personal computer. Because the Internet itself is unregulated, there are few legal restrictions on what can be broadcast.

From a traditional publisher's perspective, providing information on the Internet offers unparalleled convenience. Publications can be made instantly available around the world. Revisions, errata, updates, and retractions can be posted with similar ease. Services with which to search for these publications among the mass of Internet data are widely available and have themselves been integrated into the WWW [20-22].

Not surprisingly, the ease of publishing over the Internet will probably present significant challenges. For example, there are few obstacles to distributing potentially harmful medical information over this medium. Human error, ignorance, paucity of data, conflicts of interest, and the absence of appropriate peer review processes may all lead to such problems, which are magnified many times by the size of the audience. This may already have occurred over a medically oriented on-line bulletin board [23]. In another instance, a medical school wrested control of an Internet host from a nonphysician who was using it to provide medical information without adequate review [24].

On the Internet, the concept of the minor or obscure journal does not exist. The services capable of locating and retrieving data from any Internet host have eliminated that paradigm. Thus, any "publisher" offering information about a therapeutic regimen that is false, biased, or unbalanced may appear as prominent as even the most reputable information source. Although savvy readers may be cautious about the source and validity of such information, an audience unfamiliar with the rigors of the scientific method and the peer review process may not be as critical. Although some would endorse caveat emptor, this is a position that we, as physicians, believe is irresponsible.

What do we propose? First, we should prepare for these inevitable changes by becoming more familiar with Internet technologies through formal and informal training. Coursework aimed at providing basic computer and Internet literacy should be incorporated into medical school and continuing medical education curricula.

To satisfy the public's desire for medical information, we must continue to establish Internet-based resources. In addition to providing health information, these resources should emphasize the importance of the scientific method, peer review mechanisms, and the editorial process. Finally, it will become increasingly important that we communicate our existing peer-reviewed medical writings in an electronic, Internet-based venue. We believe that information traditionally conveyed through paper-based medical journals and texts should be made available in this manner.

It is imperative that the medical community embrace the use of Internet-based information. As health care providers, we must also be the prominent providers of Internet-based medical information for both our colleagues and the public.