Medical Communications Consultants, LLC, 103 Van Doren Place, Chapel Hill, NC 27517, USA.
J Med Ethics. 2011 Aug;37(8):498-503. doi: 10.1136/jme.2010.041830. Epub 2011 Feb 22.
Media attention to retracted research suggests that a substantial number of papers are corrupted by misinformation. In reality, every paper contains misinformation; at issue is whether the balance of correct versus incorrect information is acceptable. This paper postulates that analysis of retracted research papers can provide insight into medical misinformation, although retracted papers are not a random sample of incorrect papers. Error is the most common reason for retraction and error may be the principal cause of misinformation as well. Still, one-quarter of retracted papers are fraudulent, and misinformation may also arise through fraud. This paper hypothesises that error and fraud are the main sources of misinformation and that error is more common than fraud. Retraction removes misinformation from the literature; bias is non-retracted misinformation. Bias arises when scientific impropriety results in false research findings. Impropriety can involve experimental design, data collection, data analysis, or data presentation. Yet impropriety also arises through earnest error or statistical naiveté; not all bias is fraud. Several measures are proposed to minimise misinformation in the medical literature, including: greater detail in the clinical trial registry, with rigorous definition of inclusion and exclusion criteria and primary endpoints; clear statistical criteria for every aspect of clinical trials, especially sample size; responsibility for data integrity that accrues to all named authors; increased transparency as to how the costs of research were paid; and greater clarity as to the reasons for retraction. Misinformation can arise without malicious intent; authors of incorrect papers are owed a presumption of incompetence, not malice.
媒体对撤回研究的关注表明，大量论文被错误信息所污染。实际上，每篇论文都包含错误信息；问题在于正确信息与错误信息的平衡是否可以接受。本文假设，对撤回研究论文的分析可以深入了解医学错误信息，尽管撤回的论文并不是错误论文的随机样本。错误是撤回的最常见原因，错误也可能是错误信息的主要原因。尽管如此，四分之一的撤回论文是欺诈性的，错误信息也可能是通过欺诈产生的。本文假设错误和欺诈是错误信息的主要来源，并且错误比欺诈更为常见。撤回将错误信息从文献中删除；偏倚是非撤回的错误信息。当科学不当导致虚假研究结果时，就会出现偏倚。不当行为可能涉及实验设计、数据收集、数据分析或数据呈现。然而，不当行为也可能源于认真的错误或统计上的幼稚；并非所有偏见都是欺诈。提出了一些措施来减少医学文献中的错误信息，包括：临床试验注册中增加更详细的信息，包括严格定义纳入和排除标准以及主要终点；明确临床试验各个方面的统计标准，尤其是样本量；对所有署名作者的数据完整性负责；增加对研究费用支付方式的透明度；以及更明确地说明撤回的原因。错误信息可能是无意产生的；错误论文的作者应该被假定为能力不足，而不是恶意。
