Vesal Azad Roya, Riahinia Nosrat, Azimi Ali, Baradaran Hamid
Department of Knowledge and Information Science, Faculty of Psychology and Education, Kharazmi University, Tehran, Iran.
Center for Educational Research in Medical Sciences (CERMS), Department of Medical Education, School of Medicine, Iran University of Medical Sciences, Tehran, Iran.
Med J Islam Repub Iran. 2025 May 6;39:64. doi: 10.47176/mjiri.39.64. eCollection 2025.
According to the pyramid of evidence, systematic reviews hold the highest position among studies used in healthcare systems and policy-making. Avoiding systematic and methodological errors are demanding responsibility for authors. Clearly, erroneous studies can have irreparable consequences on health and treatment decisions. Therefore, this study aims to identify potential errors in systematic reviews within the field of health.
To systematically identify potential errors in systematic reviews, we conducted a comprehensive literature search using keywords such as "Bias," "Error," and "Systematic Reviews" across databases like PubMed, Web of Science, Scopus, Embase, Cochrane Library, and ProQuest without any time restrictions. This yielded 2333 articles and 11 books initially.After removing duplicates and unrelated sources based on predefined inclusion/exclusion criteria tailored for this study context (e.g., relevance to error identification in systematic reviews), we closely examined 88 relevant sources.
Upon analyzing the full texts of these sources with strict adherence to our criteria, we identified 77 distinct types of errors that could occur either within or between studies. These findings highlight the complexity of maintaining accuracy in systematic review methodologies.
Given the critical role systemic reviews play in informing clinical decisions and health policies, ensuring their quality is paramount. Accurate methodology ensures validity; biased studies risk leading to suboptimal patient care outcomes. By pinpointing error sources-such as selection bias or information bias-and implementing strategies to mitigate them through rigorous methodologies like robust search protocols or transparent reporting standards (e.g., PRISMA guidelines), researchers can enhance review quality significantly.
根据证据金字塔，系统评价在医疗保健系统和政策制定所使用的研究中占据最高地位。避免系统和方法学错误对作者来说责任重大。显然，错误的研究可能会对健康和治疗决策产生无法弥补的后果。因此，本研究旨在识别健康领域系统评价中的潜在错误。
为了系统地识别系统评价中的潜在错误，我们使用“偏倚”“错误”和“系统评价”等关键词，在PubMed、科学网、Scopus、Embase、考克兰图书馆和ProQuest等数据库中进行了全面的文献检索，没有任何时间限制。最初得到了2333篇文章和11本书。根据为本研究背景量身定制的预定义纳入/排除标准（例如，与系统评价中错误识别的相关性）去除重复项和无关来源后，我们仔细审查了88个相关来源。
在严格按照我们的标准分析这些来源的全文后，我们识别出了77种可能在研究内部或研究之间出现的不同类型的错误。这些发现凸显了在系统评价方法中保持准确性的复杂性。
鉴于系统评价在为临床决策和卫生政策提供信息方面所起的关键作用，确保其质量至关重要。准确的方法确保有效性；有偏倚的研究可能会导致患者护理结果不理想。通过查明错误来源，如选择偏倚或信息偏倚，并通过严格的方法（如稳健的检索方案或透明的报告标准，如PRISMA指南）实施减轻这些错误的策略，研究人员可以显著提高评价质量。
