Feature Story

Collected Works

June 2001
Karen Southwick

The genesis for a new Web site devoted to laboratory tests and disease diagnosis came from an old medium indeed: books.

It all started about five years ago when Kenneth Sims, MD, a professor of pathology at Creighton University, Omaha, was asked to do a book review of the 19th edition of a longtime standard reference, Clinical Diagnosis and Management by Laboratory Method.

To do the review properly, Dr. Sims figured he should read the book. In addition, he wanted to scan the predecessor, the 18th edition, to do a comparison. But both books were about 2,000 pages long. "I couldn’t read that much by myself," he says.

So he split the book chapters between two groups—faculty and residents—and had different people read different sections. Nearly everyone came back with similar criticisms: Despite the length of the books, the limited space meant that many topics were presented only superficially and, due to publishing cycles, recent developments were not covered.

"I started to think, ’what could you do that’s better?’" Dr. Sims recalls. At the time, the Internet was bursting upon the public scene, where such technology as hypertext, links, and unique locator names was making online information accessible in a way that previously was impossible.

Dr. Sims and his Creighton team realized, though, that simply to dump the textbook information onto the Web "would give you the worst features of both without improving anything." That is, the material would still be superficial and outdated, and reading dense text on a monitor is difficult. In addition, the process would leave control of the information with the textbook publisher.

Thanks to the Web’s low barriers to entry, "I saw [online publishing] as an opportunity for us to take control of our own academic destiny in deciding what to create," says Dr. Sims. Thus was born the idea of enlisting a consortium of pathology departments to develop a Web site, aimed at clinicians and laboratorians, that would present up-to-date information on disease diagnosis using lab tests as well as other methods.

Over the next couple of years, five other universities joined Creighton: Stanford University, University of California at Davis, University of California at San Francisco, University of Southern California, and University of Iowa. The six together put up $1.4 million to form the University Pathology Consortium, LLC, and to establish a Web site, UPCMD.com. It wasn’t easy, he recounts. "Universities are accustomed to taking in money, not giving it away. It took a number of years to convince both private and state universities to get involved."

The consortium was formed in August 1999 with this mission: "to provide current, coherent, accessible knowledge regarding human disease and its diagnosis to medical professionals as they do their work." UPCMD.com went live in September 2000 with 180 topics on disease diagnosis. Online now are about 240 topics, and that number is growing monthly by 10 to 15. A second module, "Evaluating Abnormal Test Results," was introduced in December 2000 and now offers material on about 40 common tests. UPCMD.com is available by subscription only: $149 annually per individual for both modules, $599 annually for medical libraries. Rates are discounted for group practices, clinics, and teaching programs. The six consortium members receive the Web site free.

How it works

The development group (with extensive help from University of Iowa faculty) worked from the start from a set of guiding principles: speed and ease of use over fancy graphics, exceptional scholarship and the highest academic standards, filtered and selected information, and a bulleted format that gives basic information first. Those who want to delve more deeply can do that by following hyperlinks within the site.

"There are no links to other sites," Dr. Sims says. "The idea is to provide a specific kind of information, available in 30 seconds." Thus, the vast majority of the information is text-based, although occasional images—for instance, depictions of basal-cell carcinomas—are included. Dr. Sims expects that images will become more commonplace as technology improves and bandwidth increases.

The intended audience is primarily clinicians seeking guidance on what tests to use to make a particular diagnosis, and secondarily medical technologists and pathologists. The site information currently resides in the two modules: Disease Diagnosis and Evaluating Abnormal Test Results. Modules to come will be Diagnostic Medical Testing in Psychiatric Disorders (available in July), CytoPath, Transfusion Medicine, SurgPath, and Forensic and Autopsy Pathology. Separate CME sections will be available beginning in July.

Each module contains several subject areas, as well as curriculum vitae of the authors and citations for the material. There is also a section on any controversy related to a topic. Disease Diagnosis includes sections on disease definition, recommended tests, test descriptions, rationale for using the tests, interpretation of results, and alternative tests. Evaluating Abnormal Test Results includes evaluation approach, critical value, pertinent history, common causes, analytical factors, associated testing, and followup testing.

Take the section on diagnosing systemic lupus erythematosus, or SLE, as an example of how the site functions (see "Diagnosing SLE: what UPCMD.com says," page 28). Disease definition includes bullet points such as "highest prevalence in childbearing-age women" and "diagnosis established when at least four of 11 clinical and laboratory criteria are satisfied." Then, criteria for diagnosis are listed, including malar rash, discoid rash, photosensitivity, oral ulcers, arthritis, and other conditions.

Under recommended testing, initial and confirmatory tests are described, including antinuclear antibody (ANA) and anti-double-stranded DNA (dsDNA) antibody. Also included is how to monitor those tests using 24-hour urine protein and complete blood count, hemogram, and platelet count. And the reasons for using those tests are presented under rationale, including a discussion of sensitivity and specificity.

Under interpretation, it says a negative result on the ANA test "virtually rules out active SLE." However, it’s noted that testing for anti-Ro/SS-A antibody has been advocated for rare patients with highly suggestive symptoms and persistently negative ANA results. When the ANA result is positive, it is "consistent with SLE" but not totally indicative, since the test is nonspecific and the incidence of a positive result increases with age among healthy people. With the confirmatory test, the anti-dsDNA antibody, a positive result "is highly specific for SLE," especially if a high titer is shown. A negative result does not rule out SLE, because up to 50 percent of SLE patients may lack the antibody.

Next comes a description of alternative tests, such as anti-Smith antibody, complement levels, urinalysis, renal biopsy, and anti-phospholipid antibody. Finally, a section entitled controversies in SLE testing discusses whether laboratory tests are useful in monitoring the disease or predicting its progress, and cites arguments for and against. For instance: "Monitoring tests may identify patients at risk for certain complications but don’t predict when or if disease exacerbation will occur."

Shorter sections cover reference articles, the author’s credentials, and the peer and editorial review by UPCMD.com. "We have permission from 450 out of over 1,100 peer-review journals to reproduce abstracts," Dr. Sims reports, and negotiations are underway with additional journals.

The site doesn’t discuss specific costs, Dr. Sims says, or whether tests are covered by insurance, because those items are so variable that it would be difficult to give reliable information. However, the rationale section may indicate whether tests are low- or high-cost and what demands they make on the patient, for example, whether they are invasive.

Compared to a medical textbook, which might devote a paragraph or two of information on testing for a diagnosis, "we have provided 70 to 100 Web pages of testing information on each disease," Dr. Sims says.

Recruiting authors

Signing up authors and then "somehow getting them to actually write the material" was the biggest challenge in assembling the site, Dr. Sims says. The 140 authors to date encompass multiple specialties and institutions, representing nearly every clinical area. "We emphasize only that authors be expert in their topic," Dr. Sims says. They even have a few surgeons writing about testing. "They are quite exceptional, despite their handicap of a surgery background," he jokes. Authors typically devote 50 to 150 hours of work to provide the discussion related to one disease. Each topic undergoes two levels of grammar and style review, and two of peer review.

Although authors get paid $250 per topic and are entitled to a share in the proceeds once UPCMD.com becomes profitable, "it was extremely difficult recruiting them," he says. Not only was there the time involved, but "we have a demanding format because it forces you to strip away verbiage and write in this bulleted style."

In addition, authors sign a five-year contract under which they agree to completely revise the topic at least once every year. If new developments warrant, "we can update monthly," Dr. Sims says. "Everything you see on the site has been developed from scratch."

He initially wanted to open the site with 800 disease topics, but had to reduce the number because of the struggle to get authors. Over the next several years, Dr. Sims expects to scale up to 800 to 1,000 topics. With 240 topics, "we’re close to having the core test environment, which is around 300 diseases," notes Clive R. Taylor, MD, PhD, chairman of pathology and senior associate dean for academic affairs at USC. He’s also editor-in-chief of the site.

Dr. Taylor, who wrote the sections on Hodgkin’s disease, praises the site as well constructed to achieve its goal of "helping clinicians order the right tests at the right time." He points out that a lot of time and money is wasted by "ordering the wrong test," which happens because textbooks are out of date and because medical students don’t get much training in diagnostics.

He believes the low number of subscriptions to the site will be remedied as word spreads: "It’s a matter of getting familiar with it and accustomed to using it." But consortium members disagree on whether to obtain commercial sponsorship for the site, Dr. Taylor says. The advantage is additional funding, but the disadvantage is possible undermining of credibility. "There are pluses and minuses regarding commercial sponsorship, and these issues are still being debated," he says.

Another author/editor is John D. Olson, MD, PhD, professor of pathology at the University of Texas and director of the clinical laboratories for University Health System, San Antonio. He was previously at the University of Iowa, one of the consortium members, and Dr. Sims recruited him to handle submissions on hemostasis.

"In two years’ time, the authors have generated something like 50,000 pages of information," Dr. Olson says. "It has been a big effort." He does it because he believes the material "will be very valuable to practicing clinicians," not because he expects to make a lot of money. "Nobody ever wrote anything in the medical field with the hope of getting rich."

Dr. Olson says the Web site is an interesting project well worth his participation. "I thought it had a chance of making it because it has a very uniform structure. Once an individual becomes familiar with the site, it’s easy to find your way to the information you need. That’s a major strength for clinicians."

At University Health System, pathology faculty and residents have access to the material. Dr. Olson himself taps into the site when clinicians call him for advice on what tests to use to make a diagnosis. "The place where I think it will be most helpful is for the medical student or primary care physician struggling to make a diagnosis," he says.

Pathology use

Even though pathologists are not considered the site’s primary audience, they can make good use of it. Richard E. Horowitz, MD, clinical professor of pathology at the USC and UCLA Schools of Medicine and a hospital pathologist for many years, was tapped as a consultant last fall by Dr. Sims to bring to the project the perspective of the community practitioner.

The site was originally devised for medical students and residents to learn how to make better use of the laboratory, Dr. Horowitz recalls, although its mission has since been broadened. Community hospital pathologists, he says, will find the site helpful, especially as they seek to expand their contribution in areas like disease diagnosis and disease management.

He did a survey of 75 pathologists on what skills they deemed most useful in a pathology resident who aspired to community practice. At the top of the list, for clinical pathology, were test interpretation and clinical consultation, both of which can be greatly aided by UPCMD.com, Dr. Horowitz notes.

"If a clinician calls me and says, ’how do you diagnose lupus?’ I can go to the Web site, find out current information about appropriate tests, and discuss how to use them," he says. "The UPC site provides the pathologist with immediate access to credible and current information on diagnostic tests."

Another way the pathologist can use the site is in developing disease management strategies or test algorithms. For example, as a hospital pathologist, Dr. Horowitz says he would review abnormal test results and advise clinicians on how to respond. UPCMD.com enhances that ability. "You can be proactive about calling the clinician and have the Web site at your fingertips to provide a useful and concise consultation."

Finally, once the Web site adds its CME component, pathologists, clinicians, and others, including nurses and medical technologists, can access it to meet their certification requirements. A final use is providing a credible, up-to-date knowledge source for pathologists’ presentations to professional staff conferences or meetings and in preparing articles for hospital or medical staff bulletins.

Signing up subscribers

Getting practitioners to sign up for the site hasn’t been easy. "Most physicians have a very small repertoire of tests they’re familiar with, and they’re not accustomed to accessing knowledge sources as they do their work," Dr. Sims says. "Because this is innovative, it’s difficult to explain, and getting acceptance has been tough." Then, too, physicians are resistant to something of unknown value they have to pay for themselves.

Dr. Sims admits to being disappointed with the number of subscriptions. To date, "we’ve only sold about $10,000 worth," he says. And the site is generating about 1,000 hits a day, minimal by Web standards. Dr. Sims believes he can reach break-even financial status with about 5,000 subscriptions. Once UPCMD begins to make money, the profits will be split between the six members and the authors/editors.

He is hopeful that getting the word out, through organizations such as the CAP, and licensing the Web site for broader use might help. "We are really excited about our association with Labdat.com, an order-entry portal, where UPCMD. com ’content’ is embedded in the order-entry and result-reporting process, so that anyone who’s interested in a particular test is just one click away from our material," Dr. Sims says. The consortium is also discussing with the Northern California region of Kaiser Permanente use of the site as a tool to improve laboratory use in the Kaiser system.

And the aim is to provide the service free to U.S. medical students, with the hope, of course, that they will subscribe when they start their practices. "We’re already providing this free of charge to about one-fourth of the third-year medical students," Dr. Sims says.

The consortium has been careful not to use the word "pathology" in titling its Web site, because "if you have ’pathology’ in the title, nonpathologists will tend to dismiss it as just for pathologists."

UPC also is looking to take its material into formats beyond the PC. "We’re developing an abbreviated version for handhelds," says Dr. Sims, such as the palmtop devices some physicians now carry with them. "We take the core [of the information] and end up with a compact format for handhelds," minus such things as the citations.

However, UPC won’t be trying to market its material to the general public. "We undercut our position if we say this is for everybody," Dr. Sims explains. "It’s designed for health care professionals," including internists, family practitioners, nurse practitioners, physician assistants, and anyone else who orders or interprets tests. Says Dr. Sims, "We encourage everyone to look at it and judge for themselves if we have created a useful tool for laboratorians and clinicians alike."

Karen Southwick is a freelance writer in San Francisco.