Good Clinical Practice (GCP) focuses on data quality and integrity. Clinical trial sponsors must demonstrate strict oversight of studies to ensure proper conduct, safety of study subjects and accuracy and completeness of clinical data. 
 
Centralized Risk-Based Monitoring (RBM) of clinical trials greatly enhances this process. Traditionally, oversight of a clinical study includes on-site data monitoring, performed by or on behalf of the sponsors, with monitors visiting each study site at defined intervals. Data monitoring confirms procedures are being carried out according to protocol. Monitors perform Source Data Verification (SDV), validating that the data in the Case Report Form (CRF) accurately reflects the source. This oversight is a labor-intensive, costly and inefficient process.
 
Electronic CRFs and Technology Enhance Trial Oversight
Increasing use of electronic CRFs opened the door to alternatives that provide more efficiencies and cost advantages than the SDV approach. The centralized Risk-Based Monitoring alternative improves monitoring cost-effectiveness without compromising quality and integrity. It identifies trial areas at greatest risk and implements targeted measures and controls to manage trial quality. Additionally, Risk-Based Monitoring helps improve clinical trial design, conduct, oversight, recording, and reporting, while ensuring human subject protection and reliability of trial results.
 
With ongoing RBM, cumulative data can be examined at subject and site levels, flagging potential errors that must be queried or systematic errors/errors in process that may occur at a site (e.g. measurements that look too low or too high compared to other sites). The data monitoring team can then take remedial action This could trigger an on-site monitoring visit, or further site training. Review of query rates by site, subject or form can reveal possible data quality issues. Quality Tolerance Limits (QTL) can be set and monitored to focus resources on vulnerable areas to guide the level of action required. Centralized monitoring as guided by the FDA¹ and as per the EMA Reflection Paper² is encouraged.
 
Key Risk Indicators Detect Potential Issues
Key Risk Indicators (KRIs) are critical data and other study variables or operational data that can detect potential issues at site, country or trial levels. Operational data can highlight site level concerns but potentially limited direct impact on subject safety and data integrity. They can be visualized in a dashboard format for ease of monitoring. An example of a monitored query is duration of Open Queries (see Figure 1).   

Quality Tolerance Limit Can Trigger Investigations
A Quality Tolerance Limit is a level, point, or value associated with a trial variable that should trigger an investigation if a deviation is detected in order to determine if there is a possible systematic issue (i.e. a trend has occurred). QTLs are monitored at the trial level and pre-defined before the trial commences from  a review of historical data from similar trials and where possible, using statistical methods and modelling.
 
A QTL for protocol deviations can be created and tested using simulated data. Unusually high levels of deviations may indicate an issue at that site, but unusually low levels of deviations may indicate underreporting. QTLs should identify both. Investigations made in real time increase the chance of determining root causes.
 
Easy data visualization against the QTL is key to successful RBM implementation. The above example may be plotted against calculated limits to identify breaches and seen easily in a dashboard format (see Figures 2 and 3).



Statistical Methods can Help Identify Patterns
Centralized monitoring provides access to cumulative data across sites. The use of statistical methods helps find unusual or implausible patterns in the data to indicate potential manipulation or Rates of Adverse Events (AEs): if one site has a low rate of AEs comparatively, this might indicate under-reporting, or difficulty in how to classify AEs based on symptoms, which should flag further investigation. 
 
• Lack of variability: if a site or subject shows much less variability in a measurement than other subjects/sites, this may indicate the data is not real and trigger further investigation.
 
• Digit preference: data that are invented by people tend to show preferences for certain numbers, like rounding up. Data can be examined to see if the rates of any of the digits are higher than expected.
 
• Inliers: clusters of values very close to the mean may indicate fabricated data
 
Centralized Risk-Based Monitoring – A Successful Clinical Trial
In summary, centralized Risk-Based Monitoring helps streamline trials while alleviating labor-intensive and costly SDVs. At the same time, it improves data quality by guiding and prioritizing-site visits and setting and monitoring Quality Tolerance Limits using statistical methods. All this adds up to the Good Clinical Practice mission of strict oversight and improved and more efficient approaches to clinical trial design, conduct, oversight, recording, and reporting whilst ensuring participant safety and clinical study data accuracy and completeness. 
 

References
1. Guidance for Industry: Oversight of Clinical Investigations – A Risk-Based approach to monitoring, Draft Guidance.  http://www.fda.gov/downloads/Drugs/Guidances/UCM269919.pdf
2. EMA Reflection paper on Risk-based Quality Management in Clinical Trials. www.ema.europa.eu/docs/en_GB/document_library/…/11/WC500155491.pdf

 

---

Sheelagh Aird is head of clinical data operations at PHASTAR. With more than 30 years of experience in clinical data management, Sheelah has directed and delivered projects in all phases of clinical trials across numerous therapeutic areas. She has led PHASTAR’s data management group since 2016. Sheelagh holds a BSC in pharmacology and doctorate in pharmacokinetics from the University of Bath.