文章来源：Scientific American

【翻译】

隐藏的副作用：医学研究经常略去不良结果

一项新的分析估计，对于近一半的临床研究，数据在发布时“缺失”

根据PLOS医学最近的一项分析，大约一半发布的新医学治疗研究略去了他们发现的至少一些不良影响。一组英国研究人员在遇到医学文献中缺失副作用的个案后，对此进行了审查，其中包括来自制药公司、医院和学术界的研究。为了确定问题的广泛性，他们分析了28篇期刊论文，将来自500多个临床研究的已发表数据与其原始数据集全部进行再确认。英国东安格利亚大学的医生和讲师Yoon Loke说，该审查的结果定量地证实了一些药可能有连医生都不知道的副作用，这意味着治疗可能不如他们表现得那样安全。科学美国人和Loke谈到了关于临床数据透明度的重要性。

接下来是已编辑的摘录：

科学美国人：为什么这些结果令人不安？

Yoon Loke：我们的发现证实了我们所怀疑的：缺失数据是非常普遍的。期刊出版物经常报道的测量不良事件比我们在临床研究中观察到的比例更小。我们发现这个比例是惊人的。你想要为病人做到最好，但如果你只能获得一半的信息，那么选择一种特定的药物或设备的决定可能不如你想要的那么可靠。

为什么不良事件没有被报道？

我认为其中一个问题是期刊受到空间和他们可以发布的范围的限制。 我自己是一个科学杂志的编辑，通常你想发布人们想阅读的有趣的、积极的事情。这是一种乐观的偏见。还有很多其他问题也被暗示了。例如，对于公司销售产品，发布更有利的结果可能更有益，而不是不良事件。

患者可以做什么？

如果你是患者并参加了临床试验，当你签署同意书时，你应该能够规定你希望将研究结果提供给公众。在Alltrials网站，有大量人正在竞选所有试验注册和所有结果被报道，这样就没有这么多的缺失数据。 我希望通过公众的声音，以及美国食品和药物管理局最近收紧监管要求，有关不良事件的数据将提供给更广泛的受众。

自本月起，进行临床试验的美国调查人员必须公开所有的调查结果，无论研究结果如何，这要归功于美国卫生和人类服务部和美国国家卫生研究院的新规则。同时，牛津大学的循证医学数据实验室发布了一个名为TrialsTracker的在线工具，它能准确揭示谁在扣压数据。——R.F.M.

【原文】

Hidden Side Effects: Medical Studies Often Leave Out Adverse Outcomes

Anew analysis estimates that for nearly half of clinical studies, data goes“missing” when published

Approximately half of studies published onnew medical treatments leave out at least some of the adverse effects theyuncovered, according to a recent analysis in PLOS Medicine. A teamof British researchers conducted the review after coming across individualcases of missing side effects in medical literature, which includes studiesfrom pharmaceutical companies, hospitals and academics. To determine howwidespread the problem was, they analyzed 28 journal articles that togethercross-checked the published data from more than 500 clinical studies with theiroriginal data sets. The review's results quantitatively confirm that some drugsmay have side effects not even doctors know about—which means treatments maynot be as safe as they appear, says Yoon Loke, a physician and lecturer at theUniversity of East Anglia in England. Scientific American talkedwith Loke about the importance of clinical data transparency. Edited excerptsfollow.

Scientific American: Why are these results troubling?

Yoon Loke: Whatwe found confirmed our suspicions: missing data are very common. Journalpublications often report a smaller proportion of the measured adverse eventsthan were observed in the clinical research. We found it alarming. You want todo the best for the patient, but if you can access only half the information,then a decision on choosing a particular drug or device might not be asreliable as you'd like.

Why do adverse events go unreported?

I think one of the problems is thatjournals are limited by space and the scope of what they can publish. I myselfwas an editor of a scientific journal, and often you want to publishinteresting, positive things that people want to read. It's an optimism bias.There are a lot of other issues that have been hinted at, too. For example, fora company to market a product, it may be more beneficial to publish morefavorable results, as opposed to adverse events.

What can patients do?

If you are a patient and take part in aclinical trial, when you sign the consent form you should be able to stipulatethat you want the results of the study to be available to the public. OnAlltrials.net, a large body of people are campaigning for all trials to beregistered and all results to be reported so there aren't so many missing data.I'm hoping that through the public voice—as well as the U.S. Food and DrugAdministration recently tightening up their regulatory requirements—data aboutadverse events will become available to a much wider audience.

Starting this month, U.S. investigatorsconducting clinical trials will have to make all their findings publiclyavailable—no matter what outcome a study has—thanks to a new rule from the U.S.Department of Health and Human Services and the U.S. National Institutes ofHealth. Meanwhile the Evidence-Based Medicine Data Lab at the University ofOxford released a new online tool called TrialsTracker that reveals exactly whois withholding data. — R.F.M.