Centralized Monitoring in Clinical Trials: Everything You Should Know

What Is Centralized Monitoring?

Centralized monitoring is a component of risk-based quality management (RBQM) that aims to detect emerging quality-related risks (either pre-identified or unanticipated risks) proactively during the conduct of a clinical trial. There are three main elements of centralized monitoring:

• Statistical Data Monitoring (SDM) serves as an unsupervised method for quality oversight in clinical trials, aiming to identify unanticipated risks. This approach systematically reviews all clinical data gathered during a trial, enabling the detection of potential issues such as fraud, inaccuracies in data recording, training deficiencies, or malfunctions in study equipment.
• Key Risk Indicators (KRI) concentrate on predefined risk indicators, conducting a focused examination of critical clinical and operational elements crucial to the success of a trial.
• Quality Tolerance Limits (QTLs) share a conceptual similarity with KRIs and can be regarded as a specialized category of study-level KRIs. They contribute to a comprehensive framework for monitoring and maintaining the quality and integrity of clinical trial data.

SDM, KRIs, and QTLs trigger the creation of risk signals that the study team should mitigate, and any confirmed issues are addressed to drive high-quality outcomes.

Centralized monitoring can, for example, detect a site with an atypically low or high rate of adverse events (AE) and flag it for follow-up. This gives teams the opportunity to investigate and, if necessary, take corrective action before the issue has a chance to negatively impact data quality or patient safety, thus giving studies every possible chance of success.

Regulatory Guidelines & Best Practices for Centralized Monitoring

RBQM and centralized monitoring are backed by the world’s most significant clinical development regulators, including the International Council for Harmonisation (ICH), which tends to lead the way for regional and national authorities.

ICH E6 (R2), published in 2016, emphasizes the need to develop risk-based approaches to monitoring clinical trials, including a clear expectation for centralized monitoring. It also supports the need to document centralized monitoring activities and encourages sponsors to utilize new technologies that can make clinical research safer and more efficient.¹

How Much Do KRIs Help in Improving Data Quality?

KRIs are vital to centralized monitoring and adept at quickly pinpointing issues and discerning trends.

In each study, around ten to twenty different KRIs are created, each with a predefined value range. They mainly focus on the trial’s reliability and compliance and include metrics such as adverse event reporting rates, protocol deviation rates, and data entry timelines. Sites breaching these thresholds during ongoing data review will be earmarked as high risk and flagged for follow-up.

To quantify the potential benefits of KRIs during clinical trials, CluePoints analyzed 212 studies, comprising a total of 1,676 sites having risk signals on at least one of the nine commonly used KRIs covering a broad range of risks. 83% of the site KRIs improved at the closure of the risk signal, as measured by an improved P-value between signal creation and closure.²

How Much Does SDM Help in Improving Data Quality?

SDM can reveal unexpected risks to data quality by assessing hundreds of variables collected during the clinical trial and converting the mass of P-values computed for each single site into a global score, named the site Data Inconsistency Score (DIS). This score is calculated for each site of a clinical trial on a regular basis as the study data accumulates. A site is considered ‘at-risk’ or atypical when its DIS reaches the significant threshold (DIS ≥ 1.3 corresponding to a P-value > 0.05).

In a bid to quantify the potential benefits of SDM during clinical trials, CluePoints analyzed data from 159 studies involving a total of 1,111 atypical sites with a significant DIS. In 83% of the sites, the DIS improved when all risk signals created at the site level were closed by the study team.³

How Does Centralized Monitoring with SDM Compare to Traditional Monitoring?

As part of the same project, the team retrospectively analyzed two studies that did not use any centralized monitoring tools, such as SDM or KRI, and compared their data to those from the 159 studies that did use SDM. They found that forty-three of the non-centralized monitoring sites would have had a significant DIS during the trial if SDM analysis had been conducted. Among those forty-three sites, only 56% had an improved DIS at the end of the clinical trial, which is much lower than the 83% of sites using SDM.³

Additionally, a Transcelerate analysis confirmed that traditional monitoring provides a relatively low return on investment. This analysis of more than 1,100 trials by 50-plus sponsors found that just 3.7% of EDC data is corrected after initial site entry. Of these, just 1.1% of data is amended as a result of SDV/SDR, with 1.4% being the result of EDC auto-queries and 1.2% the result of all other reviews, such as data management, biostatistician review, or medical/safety monitoring. In addition, it was unclear whether the detected errors would’ve had any bearing on data quality or patient safety.⁴

Centralized Monitoring in Clinical Trials Moving Forward

The application of artificial intelligence (AI) and machine learning (ML) to centralized monitoring is already driving further efficiency. The “risk signals” used to monitor and track investigations and corrective action contain vast amounts of free text entered by various users as they document their findings. Therefore, using natural language processing (NLP) and deep learning techniques, CluePoints screens all this information and flags risk signals that lack the required documentation. By running this algorithm retrospectively on all risk signals processed prior to the release of this ML algorithm, we observed that 40% of the signals weren’t clearly documented. This has been reduced to 30% since the release, or a 25% overall improvement in signal documentation.⁵

Challenges & Opportunities for Clinical Trials

By enabling sponsors to identify potential risks and take timely corrective action—before issues can negatively impact data reliability—centralized monitoring can help sponsors give their studies the very best chance of success. Regular data reviews can safeguard patient well-being and maintain the integrity of results throughout the trial, contributing to inspection readiness and accelerating overall timelines.

Yet hurdles to adoption remain, and chief among them is change management. Traditional methods, such as on-site monitoring, are heavily embedded into our processes and our mindsets. It means that successfully implementing change requires both a bottom-up and top-down approach that demonstrates the value of centralized monitoring techniques, provides easy-to-use systems, and, crucially, secures buy-in from all levels at all functions. Want to learn more? Contact us to continue exploring centralized monitoring in clinical trials.

References:

1. Integrated Addendum to ICH E6(R1): Guideline for good clinical practice: E6(R2) [Internet]. 2017. https://www.ich.org/page/efficacy-guidelines. Accessed 27 July 2022
2. de Viron, S.  et al. Does Central Monitoring Lead to Higher Quality? An Analysis of Key Risk Indicator Outcomes. Ther Innov Regul Sci 57, 295–303 (2023). https://doi.org/10.1007/s43441-022-00470-5
3. de Viron, S. et al. Does Central Statistical Monitoring Improve Data Quality? An Analysis of 1,111 Sites in 159 Clinical Trials. Ther Innov Regul Sci (2024). https://doi.org/10.1007/s43441-024-00613-w
4. Sheetz, N. et al. Evaluating Source Data Verification as a Quality Control Measure in Clinical Trials. Ther Innov Regul Sci 48, 671–680 (2014). https://doi.org/10.1177/2168479014554400
5. Young, S. Using Machine Learning and NLP to Improve Central Monitoring Documentation. Applied Clinical Trials. December 2023.