关于预测英国限时多中心临床试验所需站点数量的读者来信
本文分享在英国限时预算内成功交付罕见病多中心临床试验的经验:初期预测需20个站点,实际将站点数翻倍至48个伦理批准站点,才在目标时间内完成招募。站点从伦理批准到开放招募的中位时间为182天(95%CI 143–245天),范围18–613天;4个站点(9%)延迟部分源于药房签核滞后。作者给出站点数计算公式以支持限时资助的规划与成本核算。
要点
- 限时多中心试验应一开始就开放足够站点,避免后期追赶。
- 站点伦理批准到开放招募中位182天,个体差异大(18–613天)。
- 药房签核与R&D能力确认不同步可造成额外延迟。
核心结论
- 在限时预算内交付多中心试验,将初始预测站点数约翻倍可成功完成目标样本量招募。
研究设计与样本
- 基于一项罕见病多中心临床试验的经验分享,最终有48个站点获伦理批准。
- 试验注册号EudraCT 2017-001171-23,伦理批件来自NRES Committee South West(IRAS 225959)。
主要发现
- 初始预测需20个站点,试验早期将站点数增至原来的两倍以上以按时完成招募。
- 48个伦理批准站点从伦理批准到开放招募的中位时间为182天(95%CI 143–245天),范围18–613天。
- 4个站点(9%)的延迟部分原因是R&D已发出能力与容量(C&C)时药房尚未签核。
- 药房签核导致的延迟从10天到超过3个月不等,两个站点分别延迟94天和102天。
关键数值与推荐
- 站点数=所需样本量÷(每站点每月合格患者数×招募率×( accrual月数−6个月))。
临床怎么用
- 规划限时多中心试验时,应尽早按公式估算并一次性开放足够站点,而非后期补救。
- 应同步推进药房签核与R&D能力容量确认,以减少站点开放延迟。
局限性
- 本文为单中心经验分享,未报告对照或验证数据。
- 站点数公式的实际预测性能未在文中评估。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
When planning a multicentre clinical trial, it can be difficult to predict the time needed to open individual sites, and this in turn impacts on the total number of sites needed, the budget and the time frame for a clinical trial to be delivered successfully. This is of particular importance for funding applications with a limited time frame and budget such as NIHR RfPB. It is more efficient and cost-effective to open the total number of sites needed at the outset of a trial, rather than to respond later to slow site opening and recruitment. Here, we share our experience of successfully delivering a multicentre clinical trial for a rare disease within a limited time frame and budget by approximately doubling the number of sites initially predicted to be needed. We initially predicted 20 sites would be needed to deliver the clinical trial, but early on in the trial, the number of sites was more than doubled to allow successful recruitment of the target sample size within the desired time frame. Of the 48 ethically approved sites, the median time from ethical approval of a site to opening for recruitment was 182 days (95% confidence interval [143 to 245 days]) and ranged from 18 to 613 days. In four (9%) of these sites, part of the delay was due to pharmacy sign off not being given when R&D had issued capacity and capability (C&C). Delays due to pharmacy sign off varied from 10 days to over 3 months delay in two sites (94 days and 102 days). A mathematical solution to the problem of planning a study with a short recruitment window has been given to support the planning and costing of grants with fixed time constraints: number of sites = required sample size divided by (number of eligible patients per site per month times recruitment rate times (the number of months accrual minus 6 months)). We expect these results to help others who are planning multicentre clinical trials in the UK. Ethical approval from NRES Committee South West (IRAS number 225959). TRIAL REGISTRATION: EudraCT Number 2017-001171-23 . Registered on 26 June 2017.
引用
- Greenwood R, Pell J, Foscarini-Craggs P, et al. Letter on Predicting the number of sites needed to deliver a multicentre clinical trial within a limited time frame in the UK. Trials. 2020. PMID: 33087161. doi:10.1186/s13063-020-04798-x
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。