Canadian Medical Association Journal 1996; 155: 665-674
[résumé]
Paper reprints may be obtained from: Daniel N. Mendelson, The Lewin Group, Inc., 9302 Lee Highway, Ste. 500, Fairfax, Virginia 22031; dmendelson@lewin.com
© 1996 Canadian Medical Association (text and abstract/résumé)
Fuelled by the promise of quick and user-friendly access to large volumes of information on a wide range of subjects, interest in electronic information sources has grown rapidly.[4,5,17,18] However, barriers to the use of online information -- including lack of training, poor understanding of computer technology, limited access to hardware and software and concerns about the quality and completeness of information from nontraditional sources -- have slowed progress toward this goal.[10,1923] Given these barriers, a growing body of articles in the trade press has emerged to describe electronic information sources.[4,6,20,24,25]
Electronic sources of medical information can be divided into three types: static sources such as informational CD-ROMs and diagnosis-support software; clinical information systems run by managed-care organizations, hospitals and physician practices; and online medical networked information (OMNI) accessible via personal computer to a large audience. Others have reviewed the first two areas15 as well as the field of telemedicine.[26] In this article we will focus primarily on OMNI.
OMNI sources can typically be found through commercial online services and on the Internet and have made many new types of information more readily available to physicians. The scope of this information is broad: it includes electronic versions of traditional media, such as journals and clinical practice guidelines, as well as more innovative resources such as "virtual" hospitals and patients for training and education. Furthermore, OMNI sources provide physicians with access to information and organizations from all parts of the world.
Published work to date has offered pedagogic discussions of the Internet and the World Wide Web (WWW), explanations of basic concepts and terminology[4,6,2730] and samplings of OMNI sources. Although these articles have helped make physicians aware of the possibilities suggested by online information systems, they have offered little in the way of a conceptual basis for understanding what information is available and how it might be useful in practice.
In this article, we move beyond simple descriptions of the Internet, the WWW and commercial online services by setting out an organizational framework to describe the range of available OMNI. We provide examples of how this framework can be used in medical practice and discuss the current limitations of online medical information as well as potential applications in the future.
[Table of contents]
Identifying online medical sources
Because the published literature offers no comprehensive listing, we used the WWW to locate OMNI sources. We began by using nine major commercial WWW subject guides and their associated search engines to identify home pages of important OMNI sources (Table 1). Search engines scan text on the Internet in much the same way that MEDLINE scans abstracts in the medical literature. Twelve key words were used in our search: "biomedical," "clinic," "doctor," "health," "hospital," "information," "medical," "medical education," "medical informatics," "medical school," "medicine" and "physician."
The searches identified a number of sites that focus on medical information on the WWW and offer online indexes of OMNI sources (Table 2). Despite many areas of overlap, each index listed some sites not included in the others. By examining the OMNI indexes we were able to identify the types of OMNI sites on the WWW and to select representative sites from each type to explore further.
We then investigated home pages and examined active links to other home pages and commercial sites. In addition, we identified further OMNI indexes and sites by searching Internet newsgroups with the use of the Find Newsgroups search engine.
[Table of contents]
Anatomy of networked information
The organizational framework summarized in Fig. 1 makes a primary distinction between informational and interactive sources. Informational sources present information on a given topic, whereas interactive sources facilitate the exchange of information, often in the form of real-time dialogue. Users can revise their own informational WWW sites, and an informational OMNI source can become interactive through the use of links to other sites.
The subject matter of both informational and interactive OMNI sources can be either "clinical" or "nonclinical."
[Table of contents]
Informational sources
Informational sources are often formal vehicles of information dissemination that provide services, report study results and lend access to data and publications. Information flows in one direction, from the provider to the consumer. Mechanisms of accessing these sources on the Internet include WWW home pages, Gopher sites, mailing lists, Telnet and anonymous file transfer protocol (FTP) sites.
Informational sources can also be accessed through commercial physician information management services (Table 3). These services organize online resources for physicians who lack other means of Internet access or value the structure provided. These commercial services may also provide proprietary content.
The anatomy of informational OMNI sources is represented in Fig. 2; Table 4 and Table 5 describe examples of each subcategory of clinical and nonclinical informational sources.
Clinical informational sources
Many physicians are interested in online information that relates directly to their clinical practice. Most noninteractive sources of clinical information address the following subject areas.
Sources that address the nonclinical informational needs of physicians may be categorized as follows.
[Table of contents]
Interactive OMNI sources
Interactive OMNI sources feature a flow of information between users and "providers" of information, or among multiple users (Fig. 3). Interactive sources can be accessed through email, discussion groups, newsgroups, and certain types of mailing lists (sometimes called "discussion lists"). Interactive OMNI sources link users throughout the world, are less formal than informational OMNI sources and tend to reflect opinions and experiences rather than objective study results. Some interactive sources are "moderated sites," in which all information provided is screened; others are "unmoderated sites" that allow any information sent by users to be viewed by other users. In either case, the informal nature of interactive sources requires users to make their own judgements about the quality of the information they receive, particularly when no reputable organization warrants accuracy.
The anatomy of interactive OMNI sources is represented in Fig. 3; Table 6 and Table 7 provide examples of each subcategory of clinical and nonclinical sources.
Clinical interactive sources
Interactive OMNI sources provide physicians with an opportunity for informal discussion of difficult research questions and of issues that arise in patient encounters. We found no interactive sites focusing on education; those we did locate address the following subject areas.
Interactive forums on the Internet facilitate the discussion of nonclinical aspects of medical practice in the following subject areas.
[Table of contents]
Discussion
Physicians can access a wide range of clinical and nonclinical information in both informational and interactive modes. We hope that our framework will help health care professionals conceptualize the range of information available and understand the relevance of new information sources as they emerge. Our framework should also be of use to physicians searching for specific informational sources.
There are essentially two strategies for accessing OMNI when an address is not known. The first and more common strategy is to obtain relevant site references by using online search engines (e.g., Lycos and WebCrawler) to carry out keyword searches. This technique has been described elsewhere[2729] and was used as a primary method for obtaining information for this article. Keyword searches are particularly useful when the user knows that a relevant source exists but does not know its location.
The second strategy is to access known sites that contain closely related subject matter. Because the vast majority of informational sites on the Internet contain references to related sites, such "surfing" is often productive. Our hope in describing the various categories of OMNI and presenting examples in each category is to give physicians entry points from which they may move to sites of related interest.
[Table of contents]
Using OMNI in practice
The current and potential impact of OMNI on physician practice patterns can be addressed only briefly here and certainly merits further research. The OMNI sources established to date have made a wealth of data and practical information readily available to physicians. They also allow physicians to interact with colleagues from a distance and to consult with highly specialized practitioners in a variety of clinical areas. Physicians in academic and research settings likely have greater access to and realize greater benefits from the types of online information currently available. However, few data are available to indicate what physicians perceive to be the greatest benefits of online information sources.
Barriers to the use of online information sources include lack of easy access, fear of computers and lack of time to explore the wealth of available resources. In a survey of 580 000 US physicians, 70% of respondents reported that they used a computer and 46% that they had a modem.[30] The lack of consistent indexing makes it difficult to locate many sites. Answers to questions are rarely found on the Internet and rarely through a single route.
Other potential barriers include concerns about the quality and security of information. There are few formal quality control measures for OMNI sources, and mechanisms for preventing individual users or organizations from manipulating data are rarely used. There is no accrediting body to prevent the posting of biased guidelines and outcomes measures. Similarly, while electronic "consultation" with other physicians is valuable, it may be hard to assess the quality of information received in this way.
[Table of contents]
The future of OMNI
Understanding physicians' decision-making processes with respect to information use is crucial to determining how best to influence physician practice. Given the growing emphasis on medical costs and outcomes, providers, payers, administrators, manufacturers and policy-makers must all be concerned about the gap between the production and the consumption of medical information. As electronic information sources become increasingly important, it is worth exploring whether the focused development of newly emerging OMNI sources can help bridge this gap.
To maximize the potential of OMNI sources and address the discontinuity between the production and use of medical information, certain fundamental questions about the current and future users of OMNI must be answered. It is difficult to ascertain how many physicians are accessing the wealth of information available on line and how they are assimilating the information they receive. Statistics about users are generally poorly documented, and many online sources are posted with no mechanism for collecting such information. Moreover, although some research on the physician market for online services has likely been conducted by interested companies, any data that have been compiled are not publicly available. The issue is further complicated by the dynamic nature of the market, which makes data obsolete not long after they are collected and analysed.
We have already seen an explosion of interest in the online delivery of services from vendors of information. Medical journals such as CMAJ, British Medical Journal, New England Journal of Medicine and Journal of the Americal Medical Association have also begun to explore the online provision of information. Such trends point to an increased role for online services, assuming that limitations can be addressed. As these services become better organized and develop more formal quality-control measures, expanded use by physicians is likely to follow.
It is not difficult to fantasize about the future of OMNI. Possible innovations include interactive medical textbooks updated instantaneously with the results of new medical studies;[31] instant, online access to well-organized information on pharmaceutical products as soon as they are approved by the FDA, and localized registries of antibiotic sensitivities in clinical situations in which resistance is an issue. Public health notices could be posted instantaneously on physicians' computers, and immunization records from clinics might be easily accessible through remote access. All of these ideas are possible and would help physicians practise more effective and perhaps more cost-effective medicine. Understanding the strengths and limitations of existing sources is a crucial step in creating useful sources for the future.
This study was funded in part by a contract from the Assistant Secretary for Planning and Evaluation (ASPE) of the US Department of Health and Human Services. The authors thank Cheryl Austein of ASPE and Dr. Robert Rubin of The Lewin Group for guidance in all phases of the study. We also thank Dawn Bartoszewicz, Amy Morrill and Rebecca Barker for critical review of the manuscript.
[Table of contents]
References