Table_1_Spin in Published Reports of Tinnitus Randomized Controlled Trials: Evidence of Overinterpretation of Results.docx

Background: Spin refers to reporting practices that could distort the interpretation and mislead readers by being more optimistic than the results justify, thereby possibly changing the perception of clinicians and influence their decisions. Because of the clinical importance of accurate interpretation of results and the evidence of spin in other research fields, we aim to identify the nature and frequency of spin in published reports of tinnitus randomized controlled trials (RCTs) and to assess possible determinants and effects of spin.
Methods: We searched PubMed systematically for RCTs with tinnitus-related outcomes published from 2015 to 2019. All eligible articles were assessed on actual and potential spin using prespecified criteria.
Results: Our search identified 628 studies, of which 87 were eligible for evaluation. A total of 95% of the studies contained actual or potential spin. Actual spin was found mostly in the conclusion of articles, which reflected something else than the reported point estimate (or CI) of the outcome (n = 34, 39%) or which was selectively focused (n = 49, 56%). Linguistic spin (“trend,” “marginally significant,” or “tendency toward an effect”) was found in 17% of the studies. We were not able to assess the association between study characteristics and the occurrence of spin due to the low number of trials for some categories of the study characteristics. We found no effect of spin on type of journal [odds ratio (OR) −0.13, 95% CI −0.56–0.31], journal impact factor (OR 0.17, 95% CI −0.18–0.51), or number of citations (OR 1.95, CI −2.74–6.65).
Conclusion: There is a large amount of spin in tinnitus RCTs. Our findings show that there is room for improvement in reporting and interpretation of results. Awareness of different forms of spin must be raised to improve research quality and reduce research waste.

背景：报道偏向（Spin）指为使研究结果呈现出较实际情况更为乐观的态势，从而扭曲解读、误导读者，进而可能改变临床医生的认知并影响其诊疗决策的报道行为。鉴于临床研究结果准确解读的重要性，且其他研究领域已证实存在此类偏向性报道，本研究旨在明确耳鸣随机对照试验（randomized controlled trial, RCT）已发表文献中偏向性报道的类型与发生频率，并评估其潜在影响因素与效应。方法：本研究于2015年至2019年间在PubMed数据库中系统检索收录了涉及耳鸣相关结局指标的随机对照试验。按照预先设定的评估标准，对所有符合纳入标准的文献开展实际偏向性报道与潜在偏向性报道的识别与评估。结果：本次检索共获取628项研究，其中87项符合评估纳入标准。95%的研究存在实际或潜在的偏向性报道。实际偏向性报道多出现于文章结论章节，具体可分为两类：一是结论与报告的结局点估计值（或置信区间（confidence interval, CI））不符（n=34，占比39%）；二为存在选择性聚焦行为（n=49，占比56%）。17%的研究存在语言类偏向性报道，如使用"趋势"、"边缘显著"或"存在效应倾向"等表述。由于部分研究特征类别对应的试验样本量较少，本研究未能评估研究特征与偏向性报道发生之间的关联。未发现偏向性报道与期刊类型[比值比（odds ratio, OR）=-0.13，95%CI=-0.56~0.31]、期刊影响因子（OR=0.17，95%CI=-0.18~0.51）或论文被引次数（OR=1.95，CI=-2.74~6.65）存在显著关联。结论：耳鸣随机对照试验中存在大量偏向性报道。本研究结果显示，研究结果的报告与解读环节仍存在较大改进空间。需提升科研人员对各类偏向性报道的认知水平，以提升研究整体质量、减少研究资源浪费。