The Internet has tremendous potential for the delivery of instructional materials, allowing collaboration among a virtual community of users (VCU). The future success of the Internet in medicine, among other areas, will depend on their ability to enable physicians to participate in continuous medical education at a time and location convenient for them, along with useful communication. Functions such as electronic mail, chat rooms, mailing lists and computer conferencing enable physicians to establish and maintain communication and collaboration with colleagues investigating the same subjects. This encourages creation of virtual networks, such as UniNet, that allow physicians to share experiences and learn from one another [1, 2].

The UniNet project (University Network of Thematic Resources for Virtual User Communities: http://www.uninet.edu) is the first real pilot experience that proposes integrated virtual thematic services for a VCU [3]. A VCU is a group of Internet users who share a set of common aspects of the "knowledge society". UniNet began operating by the end of 1996. The UniNet project aims to be universal, free of language or geographical boundaries and open to all the interests represented in the "knowledge society", but it covers mainly scientific, academic and cultural topics [4]. Currently the major emphasis is on medical and health science VCUs.

The UniNet project is based on the voluntary and altruistic cooperative work of scientists and professionals of many countries across five continents. UniNet also intends to supply information and communication channels on the Internet to every member of the "knowledge society". Another important mission of UniNet is to provide the best resources for the inexperienced user, even one who has no previous knowledge of computers [5, 6]. UniNet supports a virtual network on Internet, for a community of doctors oriented to supporting the best practices in patient care.

Our objective was to evaluate the usefulness of a thematic network such as UniNet, specifically in medical family practice. The goal of the study is to analyse the effects of the thematic network UniNet in the following four outcomes of a primary care medical practice in a rural area: patients' referrals, number of hospital stays, use of prescriptions of drugs of low therapeutic utility and generic drug prescriptions. We tried to quantify the benefits of doctors having access to medical information of high quality, more continuing medical education and easy coordination with specialist physicians.

At the end of 1998, we presented a research proposal based on the study – before and after – to connect the doctors from a basic zone of health (BZH), located in a rural area, to UniNet through the public Internet. A grant helped us to provide computer resources and Internet connection to the rural medical center, allowing rural health care professionals to join a VCU at the UniNet network for collaborative work with medical specialists from the General Yagüe Hospital by web, text conference and e-mail.

We designed a controlled interventional study on family doctors in a rural health area. A before-and-after study was performed with a control group in order to evaluate the educational intervention.

We pursued this study in a BZH in the province of Burgos, with 5700 inhabitants affiliated with our public health system, and eight doctors – seven of them primary care physicians and one paediatric specialist – who care for patients in three different doctors' offices, located in the larger towns of this rural area: a central office and two peripheral doctors' offices. This BZH is located more than 100 km from their reference center, the General Yagüe Hospital, and the road has three mountain passes. A control group was used: an adjacent BZH, with a population of 4378 affiliated to the public health system, five doctors and a similar environment, nutritional habits, etc. The centers of these areas are 7.5 km apart, and they have the same kind of population.

All these interventions, including the initial training in Internet tools, occurred during a 12-month period. The global coordination of the project was made by the Research Unit of the General Yagüe Hospital. The educational activities were conducted by specialists from diverse medical departments: nephrology, haematology, rheumatology, neurology, gynaecology, etc. The specialists from the Department of Nephrology led the communication with the BZH and took the lead role in coordinating the educational activities.

For these purposes, each doctor had a personal computer installed in his or her office. In addition, these doctors had access to the databases at the hospital's library, as well as other existing ones freely available via the Internet (Pubmed, Embase, etc).

It took nine months to make the site operational. The first three months of the project were spent on administrative efforts, including equipment orders and installation. During the next three months the software was installed and the network and computers were configured. Afterwards, another six months was used for scheduling and training.

The work plan began with the installation of the local area network (LAN) and the computing equipment. Next, an Introduction to Internet course was delivered, and in the following months, the doctors were trained in computer operation and how to use the diverse services available via the Internet: Web, electronic mail, online databases and text conferencing.

Between January and June 2001, when it was agreed that the doctors had had enough training, we launched the first steps of the continuing medical education program, including clinical sessions, bibliographic reviews, conferences to maintain currency, debate, etc., using live communication by text conferencing, hosted on the UniNet network.

Generally we planned and ran a weekly conference session with the hospital's doctors along with those in the BZH location. A program of subjects of interest presented by the rural doctors was developed. The procedure was as follows: a written document related to the item was sent to the participants before the conference via e-mail, so that it could be read by the rural doctors in advance. Later, at a text conference, all the doctors could discuss the subject more effectively. Every session included bibliographic directions and practical questions. In addition, there was continuous interaction between the specialists and the primary care doctors during the week via e-mail, etc.

All these data were gathered through a blind and independent method with respect to this study by the Directorate of Public Health in Burgos. The doctors of the BZH of the study did not know the evaluation parameters, and the BZH control remained isolated from this investigation.

The results of our study showed positive changes in medical practice, reflected in the improvement of the evaluated parameters in the study BZH compared with the control BZH during the study period. We will now examine the real effect of the intervention on those positive changes; that is to say, the influence that it is possible to attribute to the UniNet network on the improvement of the medical practice indicators. The limitations of our study are related to its design, in that it is quasi-experimental. However, the existence of a control group of similar characteristics permits us to avoid many of these difficulties and lends weight to the results.

The analysis of Dorsch [7] indicates that rural doctors appear to have the same basic information needs as their urban counterparts, and that both groups rely on colleagues and personal libraries as their main sources of information. Rural practitioners, however, tend to make less use of journals and online databases and ask fewer clinical questions, a difference that correlates with geographical and demographic factors. Rural practitioners find many barriers to information access, including lack of time, isolation, inadequate library access, lack of equipment and skills, higher costs and an inadequate Internet infrastructure.

The Internet removes some of the isolation barriers and it has the potential to facilitate communication among rural health professionals and urban specialists. This was the goal of UniNet, and it was specifically the objective of this study. The integration of professionals into a VCU such as UniNet will not totally compensate for the lack of local library services, but electronic communications, such as synchronous text conferencing, provide a way to incorporate learning activities through expert collaboration, with a useful and practical model of continuing medical education. The formation of these virtual learning networks can allow physicians to reflect upon, generalize and discuss the applications of new information with their peers [8]. A common pitfall of many information technology programs in rural health centers is to give priority to building the technical infrastructure rather than focusing on meeting local needs and developing local expertise and ownership [9]. Text conferencing is a long-distance communication tool that is simple, powerful and cheap [10], all reasons why it is a useful and valid tool for the objectives.

Curran et al. [1] consider that the Internet is useful for continuing medical learning because it enables personal one-to-one communications: that is, the doctors, both students and instructors, may communicate with their peers through electronic mail. The Internet enables access to library resources, interactive multimedia tutorials and other related clinical and academic resources. Asynchronous or synchronous group communications allow participation in collaborative discussions with colleagues, instructors and experts via asynchronous computer conferencing or online, real-time chat groups and the like.

The limited Internet usage in our professional environment constituted the main hurdle for the study [11]. When we initiated this study, only one of the doctors used Internet services. It required a relatively long training period to achieve the mandatory prerequisite that most professionals in the BZH develop basic Internet abilities. The motivational level among the clinical staff to use the workstations varied, depending on their age, work style and the presence of a leader who would lead by example, as we found in the rural health center. Regardless of the importance of the information needed, busy clinicians will use the service only if the computer connections and interfaces are convenient and easy to use [12].

The project achieved its goals of ensuring that family doctors had access to good-quality, user-friendly, cost-effective medical information. In addition, the family doctors were assisted by a group of specialists at the hospital who served as mentors. The location of the computer was a key factor in ensuring that it was well used for clinical decision making. The doctors at the BZH under study learned how to search Medline, how to locate and gain access to the best information resources on the Internet and how to request materials through the General Yagüe Hospital library. Recently, the National Library of Medicine has developed a similar project, introducing aids that allow access to bibliographic databases (Medline), through the Internet in rural health centers, with initial success [13].

More than a technical tool to facilitate access, information technology can also serve to build and strengthen communities of like-minded individuals and institutional networks. Therefore it is necessary to form a VCU that allows for information selection. A deluge of information, much of which concerns technologically sophisticated treatments or expensive medications, is not always applicable to a rural health center. Information is relevant and useful if it emphasizes prevention and promotes evidence based, cost-effective treatment strategies [14–16].

The Internet differs from other media in one major way: the communication process is bidirectional, with an information source in the active role and the receiver in a more passive role [17]. Westberg and Miller [18] propose a model in which the academic health center integrates and distributes a wide range of electronic and human resources. This model will require substantial funding for online, full-text journal collections and networked bibliographic databases that may be even more expensive than print collections. The advent of electronic full-text journals may improve the information access of rural health professionals.

According to Marshall [19]: "the timely use of publicly accessible, electronic databases containing bibliographic and full-text information has the potential to assist in the maintenance of health professional competence, decrease the isolation and lack of up-to-date knowledge experienced by health professionals practicing outside of major population centers, and improve the quality of patient care by narrowing the gap between the publication of scientific findings and their application by researchers and clinicians". Searching for medical information via computers is certainly still not the norm [20]. Although computers provide access to an overabundance of up-to-date information, many physicians still find it more convenient to read a textbook or a journal or to ask a colleague. The utility of a VCU such as UniNet lies in its access to biomedical knowledge, colleagues and to faculty [21–24].

Our study found that providing computer resources and Internet connection to a rural medical center enabled rural doctors to join a VCU hosted by the UniNet network for collaborative work with medical specialists and allowed for access to high-quality medical information. Through this network, rural doctors had an effective, useful, user-friendly and cheap source of medical information, which may be related to the improvements observed in the medical quality indices.