J.F. Myles Clough, M.D., FRCSC
Many people have become so frustrated with the difficulty of finding useful orthopaedic information on the Internet that they have concluded there is not any worth the bother. While it is possible to develop the skill needed to search more effectively and find good stuff, there could be a shortcut: an index of addresses for web pages with orthopaedic content.
This article addresses the heart of my personal ambitions and philosophy about the Internet, so it should be considered in that light.
If you use a search engine to estimate the size of the orthopaedic Internet, you will get some very large numbers. Table 1 shows the results from using various search strings on the Google Search Engine (www.google.com)
Table 1. Estimates of the size of the Orthopaedic Internet
| Search String sent to Google || Number of pages found || Comment |
|Orthopaedic OR Orthopedic || 268,000 ||Basic collection |
|Orthopaedic OR Orthopaedic AND fracture || 40,000 ||Restricts collection to those with "fracture" |
|Orthopaedic OROrthopaedic AND (fracture OR arthroplasty OR hand OR "sports medicine") || 188,000 ||Collection with a number of orthopaedic topics |
|"sports Medicine" OR hand OR fracture OR arthroplasty (orthopaedic OR orthopedic) -clinic - center - hospital || 87,000 ||Number of "Clinic" pages reduced |
The last search string in the table is designed to find pages with orthopaedic content, avoiding orthopaedic mattresses and limiting the number of pages advertising specific orthopaedic clinics. Framing search strings is an art form and this one is not completely successful in eliminating orthopaedic clinic pages.
More than two-thirds of the pages reviewed in the first 800 were of that type. However, many of the "clinic" pages have legitimate orthopaedic information, particularly information for patients with specific complaints. The point of the exercise is to demonstrate that there are tens of thousands of pages on the Internet with interesting orthopaedic content. Eighty-seven thousand pages can't all be rubbish!
A big problem is also graphically illustrated by this example. In contrast to the innocence of the early Internet, an overwhelming proportion of pages are trying to sell you something instead of informing you. The result is that using search engines to find orthopaedic content is frustrating, even to an expert. I will be discussing ways to limit this frustration in a future article. The current point is that the "orthopaedic Internet" is very large but difficult to "find". Thinking that there is nothing of value on the Net is analogous to suggesting that arthroscopy is a difficult skill to learn and that we should continue to do open meniscectomies. I start from a different premise; there is lots of valuable information about orthopaedics on the Internet and we should be devising ways to make the search for it easy and circumvent the 'conventional' search engines. The commercial Internet may have overshadowed the information Internet but the latter still is very much there and potentially very valuable.
How do we currently find new and relevant orthopaedic information in "the literature"? As shown in Figure 1, the publication process starts with an author researching or reviewing and producing some new information. There is then an interchange of ideas between the editor, reviewers and the authors following which the "paper" is published in a journal.
Figure 1. Current Process of Information Publication and Retrieval
New issues of the journal are indexed by, for example, the National Library of Medicine (NLM), which then produces the annual publication, Indexus Medicus, and the online database of citations, Medline. When orthopaedic surgeons wish to find information on a subject they may typically search the Medline database or use Science Citations Index. Although we may not be aware of it or indeed appreciate it, there is a massive effort going on to produce these indices.
Contrast that with what currently happens on the Internet (Figure 2). The same author may produce the same information and may even undergo the same peer review process if he/she wishes to post on an official site. More commonly, authors post their information directly. Either way, the outcome is a web page but there is no attempt at an indexing system. The search engines may index the site but it is done robotically, with no understanding of content.
Figure 2. Current Process of Web page Production and Retrieval
It is the missing element of indexing which makes trying to find good quality information on the Internet so frustrating. You would face the same frustration going into a medical library to look for papers if no one was doing the indexing and cataloguing work.
It is for this reason that indexing sites have been posted on the Internet. They presuppose a new paradigm about presenting your work on the Internet. Figure 3 suggests that authors would notify the Index Site when they post orthopaedic information. Alternatively, when an orthopaedic surgeon has searched successfully, he or she might notify the site about the web pages found.
Figure 3. Role of an Index site on the Orthopaedic Internet
The key feature of the orthopaedic index as visualized here is that everyone can contribute to it. If you post an orthopaedic webpage or if you find an interesting one you can sign on to the Index site and give it the address. Because you publish directly on the Internet, there is no organization to do the indexing for you.
There are several Orthopaedic Links Collections posted by respected library institutions. The Hardin Meta-Directory at the University of Iowa lists 8 "large" lists at http://www.lib.uiowa.edu/hardin/md/ortho.html. Typically these lists are flat form HTML pages compiled by the librarians and run up to several hundred links. There is little provision for outsiders to add to these lists so they cannot form an index site of the type envisaged in Figure 3.
The main contenders for an English index to which orthopaedic surgeons can contribute are Orthopaedic Web Links (OWL) http://www.orthopaedicweblinks.com (of which I am the editor), OrthoGuide http://www.orthoguide.com/ortho/, and Orthopaedics.com http://www.orthopaedics.com/orthopaedics/. OrtoPortal http://www.ortoportal.com/ provides the same service in Spanish. The OWL collection has been pirated in a number of sites across the world. These sites are the front-ends of web-based databases and thus are very rapidly updated. The pages displayed are recreated every time the database is updated. OWL has over 7,000 links; the others have significantly less.
There are several desirable features for an index site performing the function envisaged in Figure 3:
- It should be comprehensive and give access to a high fraction of the pages on the Internet with legitimate useful orthopaedic content. Since we have estimated the number of such pages to be 87,000, there is a long way to go. No individual or team of individuals could keep up with the task of finding new pages as they are posted. Notification to the index site by authors or readers is essential. All orthopaedic surgeons benefit from the site and everyone should contribute to maintaining it. This is an important shift in responsibilities without which the Internet will not realize its full potential.
- There should be only one site, or one central collection posted on several sites. We have seen above that all orthopaedic surgeons should be involved in notifying the site - that would be difficult enough. So long as there are several competing sites and we don't know which one to notify, it will be impossible.
- The site should be organized using a universally agreed classification of orthopaedic subject headings (see previous article published in the August/September 2003 - Bulletin #62 edition).
- The orthopaedic community must run the site, and be seen publicly doing so, all for the benefit of orthopaedics. There should be strong connections between the index site and the leading institutions of orthopaedic surgery. A library is not the ideal institution to undertake such a project. While such issues as classification of subject headings and quality rating of web sites are unsettled, it is vital for orthopaedic surgeons to be laying the foundations of such a key institution in Orthopaedic Informatics.
- The index needs a strong editorial policy and staff to back it up. Editorial tasks include validating links presented by authors and readers of orthopaedic web pages, keeping the index up to date by eliminating "dead" links and devising a quality review structure.
This collection of desiderata makes funding an index site a really complex problem. The cost of running the site is small in terms of software and server space. The cost of staffing is different. The site deserves the attention of a group of orthopaedic surgeons experienced enough to understand the scope of orthopaedics and interested in Orthopaedic Informatics. They would need to be paid either in cash or professional recognition and advancement. Looked at from a different angle, an orthopaedic index site promises to be a high traffic site with the potential to be commercially profitable. Currently the monthly page view statistics for OWL is 30,000. If the scope of the index was fully realized, this number would be increased dramatically, with most orthopaedic surgeons and a large number of orthopaedic patients using the site.
It has always seemed to me that expecting orthopaedic surgeons to contribute to the index site in order to make it comprehensive and responsive to the orthopaedic community is in conflict with running the site for a profit. To put it bluntly, why would other orthopaedic surgeons contribute so that OWL would make money? Yet, passing up the opportunity to make a profit will, in the end, cause the demise of the institution unless other 'cleaner' sources of income are found. It isn't possible to put in the editorial effort indefinitely and run a Canadian orthopaedic practice.
The best option, from my admittedly biased point of view, is for the Orthopaedic Index Site to be supported by the beneficiaries, the international orthopaedic community. If national organizations such as the COA and the AAOS supported the institution, it would be seen to run for the benefit of the community (like a journal) and the organizations would exert some control over the site as well as offering much needed expertise. In many respects, the index sites are performing a service that is of benefit to COA members and others. It is not unreasonable for this relationship to become formal.
My objectivity in this area is influenced by a sense of history. I think it is inevitable that the Internet will become the dominant medium of academic communication. When that happens, a strong orthopaedic index site will be a central institution. Somehow, someday, somebody will create it. I would like this institution to have foundations that ensure it operates for the benefit of the subject of orthopaedics. There is nothing to stop this happening if there is the will, energy and attention to this central issue of Orthopaedic Informatics.