A Tough Test for PhRMA’s New Clinical Study Results Database
Posted On October 4, 2004
The new, free Clinical Study Results database for medical practitioners and the general public (http://www.clinicalstudyresults.org) from the Pharmaceutical Research and Manufacturers of America (PhRMA) went live as expected on Oct. 1, 2004. It would be unreasonable to expect comprehensive coverage on day 1 for this industrywide voluntary listing of all controlled "hypothesis-testing" clinical trials (primarily Phase III and IV studies) completed since October 2002 for U.S.-approved drugs. Bringing a database online less than 4 weeks after its Sept. 7 announcement (http://www.phrma.org/mediaroom/press/releases/07.09.2004.1063.cfm) is, perhaps, enough of a feat. However, the announcement the day before of a worldwide withdrawal of Merck & Co.'s popular arthritis and acute pain medication Vioxx, available for sale in the U.S. since 1999, created a high-bar test that shows the holes in the new database concept.
Need for the DatabaseA number of recent events have highlighted the lack of full disclosure from drug companies about results in the clinical trials that occur prior to and following FDA approval of a substance. Earlier this year, New York Attorney General Eliot Spitzer filed suit against several companies, accusing them of being less than forthcoming about negative results in clinical trials. (One suit alleged that GlaxoSmithKline hid data indicating its antidepressant Paxil might increase suicide risk in children.) Many of those suits have since been settled, with defendant companies announcing they would in the future post all results on their own Web sites.
In June, the American Medical Association (AMA) suggested a national clinical trials registry be created by the U.S. Department of Health and Human Services, making registration in such a database a condition of approval by review boards. The International Committee of Medical Journal Editors, which includes representatives from the Journal of the American Medical Association, The New England Journal of Medicine, and The Lancet, reportedly also is considering a proposal requiring that trials be listed in a registry before journal publication of results.
Presumably in an attempt to forestall a government mandate, PhRMA entered the playing field with its announcement of its own centralized database, providing a standard format that would include for all records a sponsoring company name, proprietary and generic drug names, links or reference to the FDA-approved drug label, the studied indication or disease, a bibliography of published studies with links to printed articles, and a summary of unpublished studies that describe the study center, design of trial, number of patients studied, dose and mode of administration, and a summary of conclusions and outcomes, whether positive or inconclusive. Also to be available were search options by drug name, company name, disease state, indication, and study name.What's in ClinicalStudyResults.org?
On Oct. 1, most of the data elements and search options promised were evident. You can search on the drug Avandia to find six study results for GlaxoSmithKline's diabetes type 2 drug and see the best example of individual records. It's easy to find "the best" example—so far there are only 14 records in the entire database. The population consists of nine drugs from seven of the 49 members of PhRMA listed on its Web page: Eli Lilly, Abbott, Bristol-Myers Squibb, GlaxoSmithKline, Ortho-McNeil, Pfizer, and Wyeth.
Notably, Merck is not included. In a Sept. 7 press release, Merck indicated that it intended to participate in the PhRMA database, but clearly preferred to take "advantage of the full capabilities of the current government-administered clinical trials registry" at http://www.clinicaltrials.gov. ClinicalTrials.gov, a service of the U.S. National Institutes of Health, currently details approximately 11,600 federally and privately supported human clinical research studies. Originally planned to recruit subjects for trials, the registry now also includes trials that are no longer recruiting. Merck pledged, by the end of September, to register all company-sponsored hypothesis-testing Phase III and post-marketing trials—whether open for enrollment or not.
Merck's Sept. 30 press release announcing its voluntary and immediate worldwide withdrawal of Vioxx (generic name rofecoxib) says the decision was based on "new, three-year data from a prospective, randomized, placebo-controlled clinical trial, the APPROVe (Adenomatous Polyp Prevention on VIOXX) trial," which began enrolling patients in 2000 and was designed to evaluate the efficacy of Vioxx 25 mg. in preventing recurrence of colorectal polyps. By Merck's own statements, this study should have been listed on the ClinicalTrials.gov site, and perhaps it is—there is a study called "Rofecoxib After Surgery in Treating Patients with Stage II or Stage III Colorectal Cancer." But even with its high standards in the amount of data collected, ease of searching, precision of results, and quality of information, ClinicalTrials.gov failed to yield any results when searching the study name APPROVe. The text of the complete record of the most likely-sounding "Vioxx" search match, while informative, does not show conclusively that it is the same as the APPROVe study. Requests to Merck representatives failed to elicit a response in time for this report.What's Ahead?
One failure does not break the system, particularly on its inaugural day, but it can show the weak spots. Very importantly, PhRMA's database is voluntary—no one is forced to contribute. There are also concerns about the speed of entry and updating of the database, which aims only to include studies completed since 2002. (The APPROVe study was not yet "completed," though it is now discontinued.) Most surprising is the discovery that the search capabilities enumerated (drug name, company name, disease state, indication, and study name) were the only ways to search. Without full-text searching, can it even be called a database?
To PhRMA's credit, some of the current entries are from Phase I and II trials—earlier than the promised scope. There is a clean and facile layout, but I wonder whether the underlying architecture is robust enough to handle the numbers and complexities of the additions that must come to make this effort worthwhile.
The future depends on how strong an arm the industry—or the government—wields. Neither ClinicalStudyResults.org nor ClinicalTrials.gov is currently equipped to provide a central, high-quality, comprehensive database of all the results of all clinical drug trials accessible to the general public. The government site would need to take on an additional mission, mandate compliance, add a few data elements, and fine-tune some fields. The PhRMA site needs to mandate compliance, increase its size, add retrospective coverage, and improve searchability. Common to both: mandate compliance.
[Editor's Note: Clinical Study Results is featured as the CyberSelection for November-December in The CyberSkeptic's Guide to Internet Research (http://www.cyberskeptic.com).]
Susanne Bjorner provides editorial services to publishers, librarians, authors, and researchers from her base in Spain.
Email Susanne Bjørner
Add A Comment