The evolution of trial registries and their use to assess the clinical trial enterprise.

THE ORIGINAL PURPOSE OF REGISTRIES OF CLINICAL trials was to reveal the existence of all trials, published or not, to investigators and systematic reviewers. Trials left unpublished because results were unfavorable to their sponsors, or simply because investigators never submitted them to journals for publication, could then be discovered, the trial investigators contacted, and the available trial evidence involving medical interventions could then be assessed. This would help eliminate publication bias, shown originally in the 1980s, demonstrated by Simes to affect the treatment of patients, and later revealed to be widespread by the expanding efforts of the Cochrane Collaboration. A seemingly arcane statistical point became a pressing clinical problem. Although the 1997 US Food and Drug Administration Modernization Act (FDAMA) established a US-based trial registry, ClinicalTrials.gov, the mandated content was narrowly defined by law, and trial investigators, whether funded by commercial sponsors, government agencies, or academic institutions, largely ignored any legal requirement to register. In the United States, the lines were drawn at a small invitational meeting in April 2000 that we both attended when pharmaceutical company executives stated emphatically that they would resist any effort to have a public registry of trials registered at inception. By that year, 2000, the US ClinicalTrials.gov registry was made publicly available (FIGURE), and in addition, more than 200 individual trial registries existed. Some investigators registered trials before the first patient was randomized, others when the trials were complete. The proliferation of registries by hospitals and academic centers was at least in part because their leaders could see the importance of trial enrollment as a way of attracting both grant funding and patients to their centers. Pharmaceutical companies established their own registries and contributed to existing registries as well. Many of these early trial registries were selective in the information they displayed and in their public accessibility, and they were remarkably limited on facts (eg, a drug might simply have been listed as “investigational drug”) at a time when developments in the techniques of systematic review and meta-analysis were demanding easier and fuller access to more uniform registries. Acknowledging that many others, chiefly in the United Kingdom, Europe, and elsewhere, had been pushing for trial registration, we responded with an explanatory editorial in 2003 designed to build the scientific, clinical, and ethical case for such registries and to secure support from academic leaders and journal editors, particularly in the United States. Two key events occurred in 2004. First, the New York State Attorney General Eliot Spitzer sued GlaxoSmithKline (GSK) for failing to reveal the results of trials that showed an antidepressant might be harmful. Within a matter of days, GSK and several other pharmaceutical companies announced that they would henceforth register their trials and had intended to do so all along. Shortly thereafter, the International Committee of Medical Journal Editors (ICMJE) announced that their journals, which included those regarded as among the most prestigious, would not publish reports of trials unless they had been registered, adding to their definition of trials and medical interventions the necessity of a concurrent control group and concentrating on phase 3 trials. For many, including the pharmaceutical industry and those in academia, the ICMJE decision was the most important motivator to register their trials, given the substantial gain to investigators and sponsors from positive results published in high-profile journals. The World Health Organization declared support for clinical trials registration and in 2006 launched the WHO International Clinical Trials Registry Platform (ICTRP), a portal through which more than a dozen major trial registries from around the world can now be accessed. They include ClinicalTrials.gov; ISRCTN; the Australian New Zealand