| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
Division of Anaesthesiology, Department APSIC, University Hospitals of Geneva, Geneva, Switzerland
Address correspondence and reprint requests to Martin R. Tramèr, MD, DPhil, Division of Anaesthesiology, Department APSIC, University Hospitals of Geneva, 1211 Geneva 14, Switzerland. Address e-mail to martin.tramer{at}hcuge.ch
Coping with the scientific literature has become increasingly difficult. One of the main reasons is the "information explosion." Today, more than 40,000 scientific journals contain more than 1 million articles per year (1). In biology and medicine, more than 20,000 journals and approximately 17,000 new books are published annually (2). In anesthesia alone, 53 different journals were published worldwide in 1992 (3). Nobody can cope with this enormous amount of new publications. It has been suggested that new models of medical knowledge acquisition are needed (2). Another issue is the poor quality of a substantial part of the published data. There is evidence that only half of the articles appearing in medical journals are ever cited by anybody (4). In addition, it has been suggested that approximately 15% of medical interventions are supported by solid scientific evidence, and that only 1% of the articles in medical journals are scientifically sound (5).
Then came the Internet. With the advent of this innovative, speedy, and interactive medium, even more data have become available, virtually without any limitation, almost instantly, and at a low price. The Internet provides absolute literary freedom. For the first time since the invention of letters, every human being can write and publish whenever and whatever he or she wants, without restriction, anonymously or not, and with the guarantee of a potentially worldwide publicity and fan community. For "scientific" reports, no more annoying, stupid, and ignorant peer reviewers. No worries about endless delays with the publishing process. No journal policies that prohibit duplicate publication. The data on the Internet are still increasing (http://www.nw.com/ zone/WWW/report.html).
Perhaps the Internet is simply a new carrier of data and nothing has actually changed (6). However, Internet information is different from printed information in that there is a complete lack of quality control. Health information on the Internet may be flawed, biased, misleading, duplicated, of poor quality, and unreliable. The dilemma is obvious: the anarchistic nature of the Internet is desirable for fostering open debate without censorship; this, however, raises questions about the quality of the available information (7).
Conventional peer-reviewed journals increasingly insist on professional statistical assessment of original data and provide standards on how to report trial results (8). Validity scores, such as the Oxford scale (9), facilitate standardized peer review and ensure rigorous assessment and adequate reporting of original data. For Internet consumers, however, the distinction between high-quality data and the rest is a major problem. "Upstream filtering" (i.e., data selection by an intermediary) is almost impossible because the Internet is too dynamic and rapidly changing to be reviewed by external peers (7). Another control would be to label home pages with a quality score. With such "downstream filtering," third parties would communicate selection criteria to consumers to help them to evaluate information on their own (7). Yet another approach is described by Hernández-Borges et al. (10) in this issue of the Journal. They estimated the relative scientific quality of anesthesia-relevant mailing lists, compared with peer-reviewed journals and conferences, taking the impact factor of the publishing authors as a measurement of validity. Mailing lists, in which subscribers discuss a variety of issues using electronic mail, belong to the medical resources on the Internet. Hernández-Borges et al. regard them as valuable tools in information exchange. They express, however, their concerns that mailing lists could actually do more harm than good because of the lack of external peer review and quality control. The reasoning behind the study is convincing: if we knew that a publishing author was reliable (i.e., scientifically sound), we could more easily trust this author’s data when they are published on the Internet. Hernández-Borges et al. conclude from the results of their study that qualified authors may be found among the subscribers of Internet mailing lists on anesthesiology and critical care medicine (10). At first view, this is good news, although both the usefulness of the impact factor as an indicator of academic merit and the validity of opinion may be questioned. The impact factor as a tool for evaluating medical journals has been criticized as sometimes misleading (11). Although the impact factor is an indicator of the frequency with which articles published in a particular journal will be cited in other journals, it does not tell us anything about an individual author. Science citation index or volume of external funding may be more reliable in identifying an individual’s academic impact. Whatever its limitations, unlike other measures of scientific impact, citation analysis is concrete—not anecdotal—and it is valuable for investigating science on the level of units larger than individuals (12).
Another issue is validity of opinion. The revolution of modern clinical medicine is characterized by a dramatic shift away from eminence and toward evidence. Using criteria of evidence-based medicine, there are no different levels of opinion. Opinion may be reported in a letter to the editor, in a book chapter, in a conventional (narrative) review, in a mailing list, or in an editorial (like this one). All are potentially biased. There has never been an attempt to score the validity of textbooks. Thus, why make a fuss about the validity of opinion expressed in mailing lists? A short look at anesthesia mailing lists reveals that their academic content is rather poor and that they are, at best, entertaining. On the home page of an anesthesiology discussion group (http://gasnet.med.yale.edu/discussion/anesthesiology/), the reader is told that "this list represents the individual opinions of each contributor," and that, therefore, "neither Yale University School of Medicine, the Department of Anesthesiology, GASNet, nor Keith Ruskin assume any responsibility for its contents." If nobody wants to be responsible for its contents, such a list is hardly of value. Should, then, anesthesiologists be encouraged to read them?
What needs to be done? Years before the dramatic growth of the Internet, it was suggested that more analytical training was needed in medicine and that doctors needed to be better at assessing the quality of the evidence on which they base their practice (13). This would resolve all problems: the methodologically trained consumer chooses valid data independently and ignores the rest. Ideally, such training would start in medical school. In the meantime, anesthesiologists who are searching anesthesia-related data on the Internet may rely on web sites of leading, peer-reviewed journals (14). In a recent study, published validity criteria of health-related web sites most frequently included content (quality, reliability); design and esthetics (layout, presentation); disclosure of authors, sponsors, and developers (authorship, origin); and currency of information (frequency of update, maintenance of site) (15). For anesthesiologist, using information from the Internet, for instance from mailing lists, caution concerning the validity of such data is recommended.
Footnotes
MRT is supported by a PROSPER grant from the Swiss National Research Foundation (Grant No. 3233-051939.97).
References
This article has been cited by other articles:
|
|
 |
|
  K. Ruskin, J. Oysten, and A. A. Hernandez-Borges Quality of Medical Information on the Internet Response Anesth. Analg., May 1, 2000; 90(5): 1246 - 1246. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
