Reporting of randomised factorial trials: development of extensions to the CONSORT 2010 and SPIRIT 2013 guidance statements

CONTEXTRandomised试验提供了关于卫生保健干预措施有效性的最有力证据。存在报告随机试验的标准，以便证据使用者可以清楚地看到研究设计或实施中可能影响结果有效性的优势或缺陷。SPIRIT（标准方案项目：干预性试验的建议）和CONSORT（报告试验的统一标准）声明被广泛采用并以证据为基础，为报告随机试验的方案和结果提供了最低限度的建议。析因随机试验代表了在同一研究中有效评估多种干预措施的潜力，而无需更大的样本量。在测试两种干预措施（A和B）的最简单的2x2形式中，参与者被随机分配接受无干预措施、单独接受A、单独接受B或A和B。除SPIRIT和CONSORT指南中规定的标准项目外，析因试验还有特定的设计和分析考虑。虽然扩展SPIRIT和CONSORT存在的方法的变化，包括不同的试验设计，干预措施和类型的数据，扩展还不存在factorial trials.AIMThe本研究的目的是开发扩展报告因子随机试验的SPIRIT和CONSORT statements.METHODSWe将采用类似的方法，以前使用的开发SPIRIT和CONSORT的主要声明和扩展。我们将首先对文献进行范围审查，以制定一个我们认为可能对报告析因试验有用的项目的长列表。范围审查将在本研究开始前完成，不构成资金申请的一部分，但为了完整起见，在此纳入。然后，我们将进行一项德尔菲研究，包括（1）在线调查和（2）共识会议。对于在线调查，我们将联系利益相关者，包括具有析因试验设计/实施/分析/报告经验的临床试验研究人员，知识用户，资助机构代表，期刊编辑和PPI（患者和公众参与）代表。我们将从相关出版物和研究网络（如MRC-NIHR Trials Methodology Research Partnership）中确定个人。由于没有正式的方法来估计德尔菲调查所需的样本量，我们将致力于在可用时间内招募最大的样本。以往的经验表明，最多将邀请130人。将要求参与者完成两轮调查，并指出将范围界定审查中确定的项目纳入最终清单的重要性。他们也可以在第一轮中提出新的项目。在第二轮中，他们将获得第一轮的分数以及其他参与者的总结分数。在第二轮调查后，我们将使用预先定义的标准将每个项目定义为“共识”，“共识”或“没有共识”。在调查结束后，定义为“共识”的项目将被丢弃，所有其他项目将在共识会议上进行讨论。与会者（约30人）将包括参与调查的利益攸关方。每个项目都将被讨论和投票，以纳入SPIRIT和CONSORT扩展项目的最终列表。研究的最后阶段包括编写和分发SPIRIT和CONSORT扩展指南和检查表，以及共享的解释和验证文件。析因随机试验的解释和传播。CONSORT和SPIRIT扩展将有助于资助评审员、资助机构、试验方法学家、期刊编辑、评审员和读者的关键评估和解释。

CONTEXTRandomised trials provide the strongest evidence available about the effectiveness of health care interventions. Standards exist for reporting randomised trials so that users of evidence may clearly see strengths or deficiencies in study design or conduct that could affect the validity of the results. Widely adopted and evidence-based, SPIRIT (Standard Protocol Items: Recommendations for Interventional Trials) and CONSORT (Consolidated Standards of Reporting Trials) statements provide minimum recommendations for reporting protocols and results of randomised trials.Factorial randomised trials represent the potential for efficient evaluation of more than one intervention in the same study without the need for a larger sample size. In its simplest 2x2 form to test two interventions (A and B), participants are randomised to receive either no intervention, A alone, B alone or both A and B. Factorial trials have specific design and analysis considerations in addition to standard items specified in SPIRIT and CONSORT guidance. While extensions to SPIRIT and CONSORT exist for methodological variations, including different trial designs, interventions and types of data, extensions do not yet exist for factorial trials.AIMThe aim of this study is to develop extensions for reporting factorial randomised trials to the SPIRIT and CONSORT statements.METHODSWe will adopt similar methods as used previously to develop SPIRIT and CONSORT main statements and extensions. We will first conduct a scoping review of the literature to draw up a long list of items that we consider may be useful in reporting factorial trials. The scoping review will be completed prior to starting the present study and does not form part of the request for funding, but is included here for completeness.We will then conduct a Delphi study comprising (1) an online survey, and (2) a consensus meeting. For the online survey, we will contact stakeholders including clinical trials researchers with experience in the design/conduct/analysis/reporting of factorial trials, knowledge users, representatives from funding bodies, journal editors, and PPI (patient and public involvement) representatives. We will identify individuals from relevant publications and through research networks such as the MRC-NIHR Trials Methodology Research Partnership.As there is no formal method for estimating required sample size for a Delphi survey, we will aim to recruit the largest sample in the time available. Previous experience suggests that up to 130 individuals will be invited. Participants will be asked to complete two survey rounds in which they indicate importance of including items identified in the scoping review in a final list. They can also suggest new items in round 1. In round 2, they are presented with their own score from round 1 as well as summary scores from other participants. After survey round 2, we will use pre-defined criteria to define each item as 'consensus in', 'consensus out', or 'no consensus'.Following the survey, items defined as 'consensus out' will be discarded and all others will be taken forward for discussion at a consensus meeting. Participants at the meeting (approximately 30) will comprise stakeholders represented in the survey. Each item will be discussed and voted on for inclusion in the final list of items for SPIRIT and CONSORT extension.The final phase of the study involves writing and disseminating SPIRIT and CONSORT extension guidance and checklists, and a shared Explanation and Elaboration document.BENEFITSThe outputs of this research will provide investigators with a comprehensive framework for the design, conduct, analysis, interpretation and dissemination of factorial randomised trials. CONSORT and SPIRIT extensions will aid in critical appraisal and interpretation for grant reviewers, funding bodies, trial methodologists, journal editors, reviewers, and readers.