School of Social and Community Medicine, University of Bristol, Bristol BS8 2PS, UK.
Clin Trials. 2013 Feb;10(1):104-24. doi: 10.1177/1740774512467405.
Monitoring the conduct of clinical trials is recommended by International Conference of Harmonisation Good Clinical Practice (ICH GCP) guidelines and is integral to trial quality assurance. On-site monitoring, that is, visiting trial sites, is one part of this process but little is known about the procedures that are performed in practice.
To examine and summarise published on-site monitoring methods for health-care clinical trials, including evaluations of their benefits and costs to trials.
A systematic literature review identified all articles reporting the methods and practices of on-site monitoring of randomised controlled trials (RCTs). Articles were categorised into (1) reports from research groups and organisations, (2) reports from individual RCTs, (3) randomised trials of on-site monitoring interventions, (4) cost simulations, or (5) surveys of trial staff and monitors. Data were extracted on the characteristics of the trials and groups reporting on-site monitoring (e.g., geographical origin, sponsor, and trial focus). Information from articles in categories (1)-(3) was summarised on the frequency and scope of site monitoring visits, monitoring team size and composition, activities during site visits, and reporting structures. Evaluations of the benefits and disadvantages of on-site monitoring were examined for all included articles.
In total, 57 articles were identified, comprising 21 articles about the on-site monitoring practices of 16 research groups, 30 articles from 26 RCTs, 1 on-site monitoring intervention RCT, 2 cost simulations, and 3 surveys. Publications in categories (1)-(3), mostly originated from the United States (33/52, 63%) or Europe (15/52, 29%), were predominantly describing non-commercial organisations or trials (45/52, 87%), with heart disease (9/26, 35%) or cancer (5/26, 19%) the commonest focus of individual RCTs. The frequency of visits ranged from every 6-8 weeks up to once every 3 years, with mostly all trial sites visited. The number of monitors visiting a site varied between 1 and 8. The most common on-site monitoring activity was verifying source data and consent forms, with a focus on data accuracy. Only six articles evaluated their on-site monitoring process, with improvements observed in recruitment rates and protocol adherence but with direct costs and staff time viewed as the major disadvantages. The on-site monitoring RCT ended prematurely so preventing full assessment.
Trialists and organisations may utilise additional unpublished on-site monitoring systems. The varied terminology used to describe monitoring may have limited identification of some relevant articles.
This review demonstrated that on-site monitoring is utilised in trials worldwide but systems vary considerably with little evidence to support practice. These on-site monitoring practices need to be evaluated empirically, including costs, to provide robust evidence for the contribution of site visits to trial performance and quality.
国际协调会议良好临床规范（ICH GCP）指南建议监测临床试验的开展情况，这是保证试验质量的重要组成部分。现场监测，即访问试验现场，是这一过程的一部分，但对于实际中进行的程序却知之甚少。
检查和总结已发表的关于医疗保健临床试验现场监测的方法，包括对其给试验带来的益处和成本的评估。
系统文献综述确定了所有报告随机对照试验（RCT）现场监测方法和实践的文章。文章分为（1）研究小组和组织的报告，（2）个体 RCT 的报告，（3）现场监测干预措施的 RCT，（4）成本模拟，或（5）试验人员和监测员的调查。提取了报告现场监测的试验和小组的特征（例如，地理位置、赞助商和试验重点）的数据。从类别（1）-（3）的文章中总结了现场监测访问的频率和范围、监测团队的规模和组成、现场访问期间的活动以及报告结构。对所有纳入文章中现场监测的益处和弊端进行了评估。
共确定了 57 篇文章，其中包括 21 篇关于 16 个研究小组现场监测实践的文章、26 个 RCT 中的 30 篇文章、1 篇现场监测干预 RCT、2 篇成本模拟和 3 篇调查。类别（1）-（3）的出版物主要来自美国（33/52，63%）或欧洲（15/52，29%），主要描述非商业组织或试验（45/52，87%），个体 RCT 中最常见的心脏病（9/26，35%）或癌症（5/26，19%）。访问的频率从每 6-8 周一次到每 3 年一次不等，大多数情况下所有试验地点都被访问。访问一个地点的监测员人数在 1 到 8 人之间。最常见的现场监测活动是验证源数据和同意书，重点是数据准确性。只有 6 篇文章评估了他们的现场监测过程，发现招募率和方案依从性有所提高，但直接成本和工作人员时间被认为是主要的缺点。现场监测 RCT 提前结束，因此无法进行全面评估。
试验人员和组织可能会使用额外的未发表的现场监测系统。用于描述监测的不同术语可能会限制一些相关文章的识别。
本综述表明，现场监测在全球范围内的试验中得到了应用，但系统差异很大，几乎没有证据支持实践。这些现场监测实践需要进行实证评估，包括成本，以提供现场访问对试验绩效和质量的贡献的有力证据。
