Internet teaching files in radiology: a call to action.

Increasingly, many of us are integrating our teaching files into our daily practice of radiology. We use our teaching file cases to instruct residents and fellows. When faced with a confusing case, we refer to the teaching file to look for similar examples. We may refine or broaden our differential diagnosis after reviewing a series of cases with similar clinical or radiographic aspects. Most radiologists are visually oriented, and the teaching file has become an increasingly essential extension of our work experience in imaging. The Internet has allowed the dissemination of teaching files worldwide. Advances in technology now potentially allow any case in a hard-copy teaching file to be shared through the Internet within a matter of minutes. As a result we now have an explosion of cases to learn from and to share. We also have a problem, however. Internet teaching file cases, like all information on the Internet, can be -and frequently are--posted without peer review or any other as-, surance of quality. How often have we heard someone say, "That's a great case! Why don' t you put it on the Internet?" Computer technology today makes it all too easy to disseminate information. The Internet is an almost limitless source of information, but it is an equally limitless source of misinformation. To be useful, the information included in teaching files must above all be accurate. First, is the diagnosis correct? In its essence, a teaching file case is no m o r e a n d no less--than a single case report. Those of us who are reviewers for traditional journals would never recommend publication of a case report without reasonable documentation of the accuracy of the reported information. With regard to verification of our own local teaching file cases, we might review the radiology film jacket, read the patient's medical record, or review the pathologic findings. On the Internet we are at the mercy of the person who wrote up the case. If we know who the person is, we might rely on his or her reputation. Frequently, however, we do not know who the author is. Second, is the material that accompanies the case accurate? Is the description of the images accurate? Is the differential diagnosis appropriate and complete? Are references included, and are they relevant and reasonably complete? Are there other sources that might render the issue more controversial or call the entire discussion into question? Again, locally in our own institutions we would probably know the person who entered the case personally. Brief, off-the-cuff notes are easily separated from carefully prepared descriptions and supporting documentation. All too often, we do not know the source of the information that we are viewing on the Internet. Seeing a case appear on our computer screen makes the included information somehow more authoritafive. An e-mail message or telephone call, however, frequently reveals that the teaching file case we saw was digitized from an older hard-copy film collection. It is easy to convert the original text to hypertext markup language without reviewing the accuracy or appropriateness of the included information. The urge to provide a complete teaching file has unfortunately led some Web authors to place a premium on uploading a large number of cases, rather than ensuring the quality of each case in their collection. The individual in charge of the teaching file may not even be a radiolog i s t a n d , sometimes, not even a physician. How can each of us help to improve this situation? First, caveat emptor (let the buyer beware). Don ' t use or recommend teaching file cases without validating their accuracy in some way. Residents and fellows should be taught that verification of the accuracy of information they view on the Internet is an essential aspect of using this wonderful technology. We routinely evaluate printed research articles with our residents and fellows at regular journal club meetings. This would be an appropriate forum in which to incorporate instruction on the evaluation of Internet publications. Second, those of us who chose to be Web authors should post only our best cases on the Internet. Cases should include detailed clinical and laboratory data that have been carefully validated. Pathologic confirmation should be presented. The case material should include selected, but representative, images from an integrated imaging evaluation, as well