Reply: Clinical trial registry alone is not adequate: on the perception of possible endpoint switching and P-hacking.
Reply: Clinical trial registry alone is not adequate: on the perception of possible endpoint switching and P-hacking.
复制标题
答复：仅凭临床试验注册是不够的：对可能的终点切换和 P-hacking 的看法。
DOI：
10.1093/humrep/dex360
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
Abbara A
中科院分区：
文献类型：
--
作者：
Abbara A
先生，我们很感激有机会回复希尔、康奈尔和帕图纳基斯的信。我们可以向作者保证，他们对“端点切换和p -黑客”的印象是不正确的。在试验开始前，我们的试验统计学家（LH）先验地确认了主要结果，并在试验结束后由他独立进行了分析。本试验的样本量由基于该主要结果和我们先前试验的相关数据的功率计算确定（Abbara et al., 2015）。因此，主要结果被预先定义并记录在统计分析计划（SAP）中，在提交过程中由Human Reproduction编辑团队和审稿人审查。Hill等人要求进一步澄清本试验中主要结局的选择。正如论文的引言和讨论中所述（Abbara等人，2017年），我们故意选择了以患者为中心的主要结果，以解决我们之前使用单剂量kisspeptin的试验中观察到的反应可变性（Abbara等人，2015年）。从患者的角度来看，获得临床有效结果的机会比一组患者的平均卵母细胞产量更有意义，因为反应的可变性可能导致特定患者遇到的不良或良好反应。高影响力期刊上的许多试验选择了类似的主要结果，评估达到疗效阈值的患者比例，而不是仅呈现一组患者的平均差异（Garg等人，2017）。在英国，鼓励选择对患者而不仅仅是对研究人员有意义的主要结果（称为“患者和公众参与研究”；PPI），患者被视为临床研究试验结果设计和选择的合作伙伴，而不仅仅是参与者（Bagley et al., 2016）。因此，本试验的主要结局选择是在与我们之前试验的患者讨论后做出的（Jayasena et al., 2014; Abbara et al., 2015）。我们同意Hill及其同事的观点，即重要的是额外提供传统的卵母细胞成熟标记，如成熟卵母细胞数量和卵母细胞成熟率（如手稿表2所示）（Abbara et al., 2017），然而，重要的是要认识到这些标记作为触发有效性的衡量标准也有局限性。成熟卵母细胞的数量在很大程度上取决于触发当天大小合适的卵泡的数量。2007年，Shapiro et al.（2007）对hCG和GnRH激动剂触发卵母细胞成熟的效果进行了回顾性分析。他们发现，与hCG（21.6）相比，GnRH激动剂的卵母细胞数量显著增加（28.8）（Shapiro et al., 2007）。然而，在本回顾性研究中，接受GnRH激动剂治疗的患者在触发当天有更多的卵泡（GnRH激动剂34.2个卵泡；hCG 21.7个卵泡），因此难以准确比较两组之间的触发效果（Shapiro et al., 2007）。因此，在他们后来的工作中，Shapiro引入了“卵母细胞产量”的概念，即收集的卵母细胞数量被校正为触发当天的卵泡数量（Shapiro et al., 2011）。他们报告了GnRH激动剂触发后的成熟卵母细胞产量（触发当天成熟卵母细胞占卵泡≥10 mm的比例）为63% （Shapiro et al., 2011）。Chen等人（2012）采用类似的方法报告了卵母细胞产量（取卵当天卵母细胞占卵泡≥10mm的比例），并确定了…
Sir, We are grateful for the opportunity to respond to the letter written by Hill, Connell and Patounakis. We can reassure the authors that their impression of ‘endpoint switching and P-hacking’is incorrect. The primary outcome was confirmed a priori with our trial statistician (LH) before the start of the trial and analysis independently carried out by him following conclusion of the trial. The sample size for this trial was determined by a power calculation based on this primary outcome and relevant data from our previous trial (Abbara et al., 2015). The primary outcome was thus predefined and recorded in the statistical analysis plan (SAP) which was reviewed by Human Reproduction editorial team and reviewers during the submission process. Hill et al. request further clarification for the choice of primary outcome used in this trial. As stated in the Introduction and Discussion of the manuscript (Abbara et al., 2017), we deliberately chose a patientcentric primary outcome to address the variability in response observed in our previous trial using a single bolus of kisspeptin (Abbara et al., 2015). From a patient’s perspective, the chance of achieving a clinically effective outcome is a more meaningful outcome than the average oocyte yield across a group, for which the variability in response could result in either a poor or good response being encountered for a particular patient. A number of trials in high-impact journals have chosen similar primary outcomes assessing the proportion of patients who meet a threshold for efficacy, rather than presenting only the mean difference across a group (Garg et al., 2017). In the UK, there is a drive to encourage the choice of primary outcomes that are meaningful to patients rather than only to researchers (termed ‘patient and public involvement in research’; PPI), with patients viewed as partners in the design and choice of outcomes for clinical research trials rather than merely as participants (Bagley et al., 2016). Hence, the choice of primary outcome for this trial was made following discussions with patients from our previous trials (Jayasena et al., 2014; Abbara et al., 2015). We agree with Hill and colleagues that it is important to additionally present traditional markers of oocyte maturation such as the number of mature oocytes and oocyte maturation rate (as presented in Table 2 of the manuscript)(Abbara et al., 2017), however, it is important to recognize that these markers also have limitations as measures of trigger efficacy. The number of mature oocytes retrieved heavily depends on the number of appropriately sized follicles on the day of trigger. In 2007, Shapiro et al.(2007) conducted a retrospective analysis comparing the efficacy of hCG and GnRH agonist to trigger oocyte maturation. They observed that GnRH agonist resulted in a significantly higher number of oocytes retrieved (28.8) when compared with hCG (21.6)(Shapiro et al., 2007). However, in this retrospective study, patients receiving GnRH agonist had a greater number of follicles on the day of trigger (GnRH agonist 34.2 follicles; hCG 21.7 follicles) making it difficult to accurately compare trigger efficacy between the two groups (Shapiro et al., 2007). Thus in their later work, Shapiro introduced the concept of an ‘oocyte yield’, whereby the number of oocytes collected is corrected for the number of follicles on the day of trigger (Shapiro et al., 2011). They reported a mature oocyte yield (mature oocytes as a proportion of follicles of≥ 10 mm on the day of trigger) of 63% after GnRH agonist trigger (Shapiro et al., 2011). Chen et al.(2012) used a similar approach reporting oocyte yield (oocytes as a proportion of follicles≥ 10mm on the day of oocyte retrieval) and determined an …
登录
查看更多内容
用促性腺激素释放激素激动剂“触发”后诱导黄体生成激素激增的功效。
影响因子：
6.7
作者：
B. Shapiro;S. Daneshmand;H. Restrepo;F. C. Garner;M. Aguirre;C. Hudson
通讯作者：
C. Hudson
影响因子：
4
作者：
Haas J;Zilberberg E;Dar S;Kedem A;Machtinger R;Orvieto R
通讯作者：
Orvieto R
影响因子：
15.9
作者：
Jayasena, Channa N.;Abbara, Ali;Dhillo, Waljit S.
通讯作者：
Dhillo, Waljit S.
影响因子：
0.2
作者：
Schulz KF;Altman DG;Moher D
通讯作者：
Moher D