The Institute of Cancer Research, 15 Cotswold Road, Sutton SM2 5NG, United Kingdom.
Contemp Clin Trials. 2023 Aug;131:107277. doi: 10.1016/j.cct.2023.107277. Epub 2023 Jun 29.
This project aims to: (1) assess the completeness of information in flow diagrams of published early phase dose-finding (EPDF) trials based on CONSORT recommendations, and if additional features on dose (de-)escalation were presented; (2) propose new flow diagrams presenting how doses were (de-)escalated throughout the trial.
Flow diagrams were extracted from a random sample of 259 EPDF trials, published from 2011 to 2020 indexed in PubMed. Diagrams were scored out of 15 following CONSORT recommendations with an additional score for presence of (de-)escalation. New templates were proposed for features that were deficient and presented to 39 methodologists and 11 clinical trialists in October and December 2022.
98 (38%) papers included a flow diagram. Flow diagrams were most deficient in the reporting of reasons for lost to follow up (2%) and reasons for not receiving allocated intervention (14%). Few (39%) presented sequential dose-decision stages. Of voting methodologists, 33/38 (87%) agreed or strongly agreed that for participants recruited in cohorts, presenting the (de-)escalation steps in the flow diagram is a useful feature, also expressed by the trial investigators. Most workshop attendees (35/39, 90%) preferred a larger dose to be displayed higher up within the flow diagram than a smaller dose.
Most published trials do not provide a flow diagram, and for those that do, essential information is often omitted. EPDF flow diagrams capturing information on participant flow in the trial's journey, encapsulated within one figure, are highly recommended to promote transparency and interpretability of trial results.
本项目旨在：(1) 根据 CONSORT 建议评估已发表的早期剂量发现 (EPDF) 试验流程图中信息的完整性，如果呈现了额外的剂量 (降) 级特征；(2) 提出新的流程图，展示整个试验过程中剂量是如何 (降) 级的。
从 2011 年至 2020 年在 PubMed 中索引的随机抽取的 259 个 EPDF 试验中提取流程图。根据 CONSORT 建议对流程图进行评分，满分为 15 分，另外还有(降)级特征的评分。对于缺失的特征，提出了新的模板，并在 2022 年 10 月和 12 月向 39 名方法学家和 11 名临床试验人员展示。
98 篇(38%)论文包含流程图。流程图在报告失访原因(2%)和未接受分配干预的原因(14%)方面最不完善。很少(39%)呈现了连续的剂量决策阶段。38 名投票方法学家中有 33 名(87%)表示同意或强烈同意对于按队列招募的参与者，在流程图中呈现(降)级步骤是一个有用的特征，试验研究者也表达了这一观点。大多数参加研讨会的人员(39 人中有 35 人，90%)更喜欢在流程图中更高的位置显示较大的剂量，而不是较小的剂量。
大多数已发表的试验都没有提供流程图，对于那些提供了流程图的试验，通常会遗漏重要信息。强烈建议使用 EPDF 流程图，以一个图形的形式捕获试验过程中参与者的流程信息，以提高试验结果的透明度和可解释性。
