How to Find the Latest
On Results of Clinical Trials

The current push to require drug companies to disclose their unpublished clinical trials could unleash a flood of new information for doctors and patients. But in the meantime, there is already a growing effort by medical publishers, scientific groups and government agencies aimed at helping people find and interpret clinical-trials results online.

Currently existing registries are run by the National Institutes of Health and a host of private organizations. Web sites like MedlinePlus.gov offer direct links to most published medical studies, which in some cases are free or else can be purchased directly from the journal. And more help is under development, such as an upcoming guide from the National Library of Medicine on understanding reports.

GO FIGURE Find out where to learn more on clinical trials and see a list of medicines for children being developed. TRACKING DOWN TRIAL RESULTS See a list of sites that can help you track down clinical-trial results.

Most of these efforts center around published studies, but unpublished research is sometimes available on the Web sites of medical specialty conferences, where researchers may present work that isn't later published. One study in the 1990s of research submitted to a meeting of the Society for Academic Emergency Medicine said that of 223 unpublished investigators studied, only 20% had submitted manuscripts to a journal.

If anything, the challenge is to sort through it all: There are close to 50,000 clinical trials going on at present in the U.S. alone, and every day there seems to be a new study, sometimes reinforcing and sometimes contradicting previous research. Much of the research is in dense medical language, which can be tough slogging for even the savviest patient. But there are also resources that can help consumers find information about trial results on plain language as well as tips on how to weigh the their legitimacy by assessing such factors as how many patients were involved and who paid for the research.

The feud over what research should be published flared up again earlier this month following a recent study published by the medical journal Lancet. The study found that unpublished research on some antidepressants included more negative findings than the published results. The authors argued that the suppression of such research "can lead to erroneous recommendations for treatment."

The American Medical Association voted Tuesday to recommend that the Department of Health and Human Services develop a registry of all clinical trials and their outcomes. Also, editors of some medical journals say they have discussed requiring companies to register in a trial database as a condition of publication in the journals.

The AMA proposal faces stiff hurdles. Alan Goldhammer, associate vice president for regulatory affairs at the Pharmaceutical Research and Manufacturers of America says the drug-company trade group plans to review the registry proposal, but is concerned about the protection of drug makers' proprietary data.

Some drug companies may voluntarily post unpublished trial results in hopes of pre-empting any federal requirement. But Mr. Goldhammer says such results can be confusing for consumers. Research documents that GlaxoSmithKline just released on its gsk.com Web site on the effects of the antidepressant Paxil in children run to thousands of pages, and it isn't clear whether any significant conclusions that can be drawn from the data.

Nevertheless, trial results can provide critical information for consumers. Trial results submitted for Food and Drug Administration approval of a drug must describe side effects, drug interactions and the various risks different types of patients might face. For example, data from the Women's Health Initiative, a study of hormone-replacement therapy (www.nia.nih.gov/menopause/faq-detailed.htm), offers details on risk of heart attack, colon cancer and hip fractures that woman wouldn't be able to find just from the pill package.

So it can be worthwhile for patients to wade through the medical data. Among the most comprehensive sources for clinical-trial data is the Cochrane Collaboration. This U.K.-based nonprofit group, which advocates broad disclosure of evidence from clinical trials, has volunteers who search the world's medical literature to find randomized clinical trials. They also say they look to report on unpublished studies. Consumers can sift through the group's free reports at www.trialscentral.org.

Several agencies of the federal government also provide information on trial results. The National Cancer Institute's cancer.gov site reports regularly on findings from published cancer clinical trials. The site's search engine lets users look for trials by phase of the trial, the type of treatment or intervention -- such as chemotherapy or vaccine therapy, the drugs being tested, and other criteria.

At the National Library of Medicine's ClinicalTrials.gov site, which lists more than 7,000 trials, users can click on the "What's New" icon, then ask that the search include all completed trials submitted in the past seven days or past 30 days, and get detailed data about findings. By law, all trials of drugs for life-threatening conditions must be entered in this government registry, though the FDA is looking into a report that only 49% of cancer-related trials are submitted to the site.

The library is also developing a guide to be available online to help consumers understand reports from clinical trial research, and already offers glossaries on the various government sites, according to Alexa McCray, director of biomedical communications for the library.

MedlinePlus.gov, the consumer-oriented version of the National Library of Medicine's site, helps distill much of the data from drug tests in drug dictionaries, encyclopedia and glossary that makes it easier to understand risks and benefits.

Ultimately, should consult with their doctor before relying on any one medical study or clinical trial result, says Ken Getz, founder of the nonprofit Center for Information and Study on Clinical Research Participation. Results of one study have to be duplicated by other researchers at different locations before they are accepted as general medical practice.

Some guidelines for evaluating studies:

What's the source? Studies published in peer-reviewed journals such as the New England Journal, the Journal of the American Medical Association or Lancet have editorial boards who review studies for publication. Specialty journals often contain the best studies on specific diseases and conditions, such as circulation for heart conditions.

Studies conducted by university teaching hospitals or funded by the National Institutes of Health are often the most reliable.

Who was studied? Results of studies often seem contradictory, but that may be because they used different eligibility criteria or different age, gender or ethnic groups. "If the study says it looked at white males between 34 and 50, it isn't going to have a lot of relevance for a 64-year-old white woman," says John Schneider, an internist who is chairman of the AMA's council of scientific affairs.

How was the study conducted? Double-blinded randomized controlled trial studies are usually considered the most valid. That means neither the investigator nor the trial participant know who is receiving a drug and who is receiving a placebo.

What do the statistics mean? In general, clinical-trial results are considered statistically significant if there is a less than 5% probability that the difference observed would occur by chance alone. A primer from Englewood, Colo.-based Craig Hospital for Spinal Cord Injuries (www.craighospital.org), titled "Those Scary Statistics," helps explain statistics used in research.