Medical information overload can confuse and mislead

The practice of performing lobotomies has long been discredited, and people searching the Internet today would find few data in support of the procedure. But while some medical treatments-such as lobotomy-are easy to vet for viability, much of the medical information available through the Internet can overwhelm patients and referring physicians.

The literature is rife with studies that invalidate or at least dampen the impact of previously accepted studies and treatments. Physicians and others involved in scientific research understand that published information is not necessarily gospel. They know that interpretation requires knowledge of previous findings and that the accepted results of today's study can be refuted tomorrow.

"Very few articles are correct forever," said Dr. Robert J. Stanley, editor-in-chief of the American Journal of Roentgenology. "I worry when the person with no medical background gets 50,000 hits on Google about a disease and really doesn't know what is reliable."

A study by Dr. John P.A. Ioannidis at the Institute for Clinical Research and Health Policy at Tufts-New England Medical Center found that one-third of 45 frequently cited studies were eventually refuted or downgraded in importance by subsequent research (JAMA 2005;294:218-228). Not surprisingly, nonrandomized, smaller studies were more likely to be contradicted than larger randomized trials. But even the most highly cited randomized trials, especially the small ones, may be challenged over time, Ioannidis said.

While Ioannidis did not look at studies specific to radiology, imaging is not immune to its share of false starts. A few years ago, whole-body "wellness" screening was the rage, until issues of radiation exposure, false positives, and downstream costs came to the forefront. Researchers investigating MRI of the prostate initially claimed that high-signal-intensity lesions were cancer. Today, it is accepted that these cancers are usually low-signal-intensity lesions on MRI. In the last few years, several studies have supported virtual colonoscopy, while others have found in favor of conventional colonoscopy. How can patients and referring physicians make sense of it all?

While the Internet might be part of the problem, it may also be part of the solution. Advocacy sites have sprung up to help people sift through the information overload. One such site, patientInform (www.patientinform.org), provides links to journal articles as well as commentary that discusses the significance, strengths, and limitations of a study and its relevance for treatment and self-care. The American Roentgen Ray Society and the Society of Nuclear Medicine have joined two dozen other participating scientific publishers in this advocacy project, which went live in June.

Readers visiting the Web sites of the American Diabetes Association, American Cancer Association, and American Heart Association will have access to patientInform. Specialists at these three organizations comb the participating journals looking for content of interest to their constituency. The patient can link directly to the original article, free of charge, or just read the commentary.

"We have been making our own efforts to provide more access to journal articles," Stanley said. "But to have a place where patients can read a layperson's synopsis of the study and have it put into context in terms of previous research makes good sense to me."

PatientInform germinated from an organization called the Washington DC Principles for Free Access to Science, of which the ARRS and SNM are members. The DC Principles group, which comprises scores of small not-for-profit scientific publishers, was formed a few years ago to combat the negative press heaped on publishers for charging exorbitant journal prices and essentially denying the public easy access to clinical information. Many of these publishers have since made their contents freely available six months to a year after publication. Another issue, raised by the National Institutes of Health, was that government-funded research was not freely available to the public once it was published in scientific journals.

"We wanted to address those concerns," said Becky Haines, ARRS director of communications. "We wanted to differentiate ourselves from the folks who make lots of money publishing journals that cost thousands of dollars."

Much to the dismay of the DC Principles, the NIH in May 2005 created PubMedCentral, a central repository for immediate free access to manuscripts supported by government funding. Opponents of the repository say it is redundant and ignores free access infrastructure already in existence. The ARRS complained in a letter to NIH Director Dr. Elias A. Zerhouni that immediate free access to articles shortly after publication could negatively affect its subscription base and, consequently, its membership. A group of nonprofit organizations contacted Senator Arlen Specter (R-PA) in July requesting a full evaluation of the costs and impact of the NIH policy on private sector publishing.

Both groups want the same result: easy access for the public to valuable clinical information. In the meantime, advocacy groups are spearheading efforts to help the public and referring physicians better understand the wealth of research available through the Internet.