Evidence contamination: what it is and why it matters

Since its seminal movement initiated through the efforts of the visionary pioneers in the mid-twentieth century, evidence-based medicine (EBM) has gradually gained widespread recognition and support from the medical community.1–4 Today, it stands as a cornerstone of contemporary healthcare practice.5 At its core, EBM seeks to integrate individual clinical expertise with best available research evidence and patient preferences.6 7 For over a quarter of a century, methodologists have devoted significant efforts to the development of structured frameworks to support informed decision-making.8–17 Together, these efforts constitute the current paradigm of identifying the best available evidence. They have driven the evolution of the evidence ecosystem and continue to guide routine clinical practice—an approach that, while imperfect, has contributed to significant advances in healthcare. Nevertheless, the growing concerns about research misconduct (see related terminologies in box 1) bring a new and serious challenge for this well-established system.18 19 During the past decades, retraction rates increased from 10.7 to 44.8 per 100 000 publications, with research misconduct accounting for the most (66.8%).20 Research misconduct can undermine the foundations of evidence-based decision-making: what seems to be the best available evidence may actually be fabricated or manipulated.21 If not properly addressed, the best available evidence may be distorted by fraudulent studies.22 Addressing this problem is challenging, as Grey et al pointed out, ‘ retracting untrustworthy, problematic academic publications can take years, and most often never occurs ’.23 However, this should not prevent us from taking the first step forward. Box 1 ### What is evidence contamination and what is not. #### Basic terminology