|
||||||||||||||
 |
||||||||||||||
|
Academy News: November 2004 Reporting data from clinical research trials Dr. D works for a pharmaceutical company overseeing their clinical trial studies. He recently worked with Dr. E on a study of Schnozshtop. The study compared Schnozshtop to placebo and a currently available medication, Acho-less. Now that the study is completed, Dr. E analyzed the data and found that Schnozshtop was less effective than Acho-less, but more effective than a placebo. He asks Dr. D what should be done with these data. Dr. D suggests that a manuscript be prepared with just the data comparing Schnozshtop to placebo. When Dr. E questioned this, Dr. D explained that they are simply reporting the positive findings of the study, since the negative findings probably wouldn’t get published anyway. Could this really happen? Data suppressed from clinical research trials or negative trials not even reported? It can happen, according to an editorial in the American Medical Association (AMA) Medical News, where it was estimated that nearly half of all results are never published.1 As part of an investigation into clinical trials, the AMA Council on Scientific Affairs found that positive findings were more likely to be published than negative ones.1, 2 This bias seemed to be greater for industry-sponsored research, where most results were found to support “pro-industry conclusions.” In most “comparison trials, the sponsor’s drug is almost always deemed equivalent or superior to the comparator.”2 This potential bias is even more concerning when it is multiplied by the fact that greater than 70% of all human clinical trials are financially sponsored by industry.2 So you may ask, this is a worrisome bias, but is it ethically inappropriate? To answer this question, we should examine the AMA Principles of Medical Ethics. Principle V states “a physician shall continue to…advance scientific knowledge,…make relevant information available to patients, colleagues, and the public….”3 But “what is relevant information?” Is negative data relevant? In some situations it is relevant, such as preventing unnecessary exposure to a potentially toxic therapy, or preventing use of ineffective therapies that increase the time until appropriate care (such as in the case of chemotherapy). As with many ethical issues, the decision on what is relevant cannot be stated by a hard and fast rule. The AMA Principles, while not defining relevance, do mention that physicians and institutions undertaking medical research must conduct their studies in an ethically appropriate manner.4 Related to this, is an expectation that “publication of results will not be unduly delayed or otherwise obstructed by the sponsoring company.”5 So, we might conclude that if a physician associated with the research feels the data are relevant, then the ethically appropriate response is to publish all data–regardless of its potential positive or negative impact on the sponsoring institution. In response to this rather nebulous issue, the International Committee of Medical Journal Editors (ICMJE) released a statement in their 11 member journals that they would require registration of all clinical trials as a prerequisite to publication.6 The major argument for registration is that failing to report negative studies means that these data are not available to “influence the thinking of patients, clinicians, other researchers, and experts who write practice guidelines or decide on insurance-coverage policy.” Therefore, registering a trial at its inception in a public database, would allow for the evaluation of the body of negative evidence against a given treatment. In addition, if advance registration is required, omission of data, like specific arms of a study, would be much less likely to occur, since the astute reviewer could quickly surmise from the database that there were missing data in the published report. The ultimate goal of this requirement is to make data and clinical evidence more transparent to all who might seek it. Of course, this mechanism demands listing in a registry, it does not require publication of all data obtained in a study. This is where the ethical considerations remain murky. If we return to our vignette, what would be considered appropriate? Clearly, if the study was to be reported in an ICMJE journal, it would need to be registered. This would show that a comparison between the two drugs was a part of the study. Reviewers would then likely assume that if the data were not included in the published account, it was not a positive outcome. However, the larger question is whether it is ever appropriate to omit data from a published study. Obviously the easy answer is to strive for openness with full disclosure. In the vignette, however, this does not happen. What the ethically appropriate action would be depends on your personal view, but as mentioned above, the AMA Principles do put a burden on physicians practicing research to strive for full publication.5 Therefore, if Dr. E felt the omitted data were relevant, he would have an obligation to push for its publication; similarly, the sponsoring company should not obstruct the publication. In general, it seems it is best to publish too much information rather than too little. Study registries may help in this matter, but perhaps all studies, positive or negative, should be published and available for all to view. References
The Ethics Committee provides these discussions as a way to open a dialogue on the various ethical issues that confront our specialty on a daily basis. These issues are often quite complex and do not have simple “right” or “wrong” solutions. The articles are meant as a way to highlight the various issues that are involved in these ethical dilemmas, they should not be viewed as the Ethics Committee or the AAAAI’s particular stance on an issue. |
|||||||||||||
|
© 1996-2008 · All Rights Reserved · American Academy of Allergy Asthma & Immunology |
||||||||||||||
