18.1 Risk Assessment and Centralised Surveillance

Risk Assessment and Centralised Surveillance

An effectiveness criterion needs an observation period, denominator and reason for the chosen sample. In a hypothetical site review, 5 of 25 expected timed assessments were missed before remediation: 5/25 × 100 = 20%. After a scheduling change, the next 20 expected assessments are all documented within the protocol window: 0/20 = 0% missed in that observed period.

These figures describe the reviewed records; they do not prove that recurrence is permanently impossible or that no unreviewed record contains a problem.

Check that the new denominator includes every assessment actually expected under the defined review, rather than selecting only successful visits. Consider whether staff turnover, weekends or other circumstances that caused the original failure were represented. If the observed period contains no relevant high-risk activity, a zero-failure count may provide little evidence about the control.

Document implementation separately from effectiveness. A revised scheduler can be installed and staff trained while the first suitable effectiveness observations are still pending. The action tracker should show those different states honestly. If failures recur, reassess causes and the adequacy of the measure; do not change the success definition after seeing poor results simply to close the issue.

Risk-Based Monitoring (RBM), Centralized Monitoring & KRIs

Monitoring should address the risks of the particular trial. Comparing transcribed values with source records is useful, but a matching value does not establish that consent was valid, an assessment occurred correctly, or a safety event was reported. Combine data verification with process review, central surveillance and other proportionate controls. No fixed budget saving or discrepancy-detection percentage establishes the right approach for every study.

A Practical Risk-Based Monitoring Framework

Operationalizing RBM requires a structured framework that links protocol risk assessment directly to monitoring execution. One possible operational approach, centers on three foundational components:

Structured Risk Assessment

Prior to trial finalization, a cross-functional sponsor team (clinical operations, biostatistics, medical affairs, data management, and QA) executes a formal structured risk assessment. The tool scores risks across multiple dimensions:

• Protocol Complexity: Number of procedures per visit, volume of protocol-mandated blood draws, invasive biopsy requirements, and tightness of visit windows.
• Subject Population Vulnerability: Pediatric, geriatric, critically ill, or incapacitated participants with diminished capacity to report adverse events.
• Investigational Product (IP) Safety Profile: First-in-human biologics, narrow therapeutic windows, complex cold-chain storage, or hazardous reconstitution procedures.
• Endpoint Objectivity: Objective automated lab markers (e.g., serum cholesterol) versus subjective clinical scales (e.g., pain scores or psychiatric evaluation ratings).
• Site Operational Novelty: Experience level of investigators, unfamiliar electronic platforms, or decentralized home health nursing elements.

Identifying Critical-to-Quality (CtQ) Factors

From the RACT assessment, the study team isolates those critical data fields and processes that directly underpin subject safety and primary study conclusions. In a cardiovascular trial, CtQ data includes signed informed consent, primary major adverse cardiac event (MACE) endpoints, troponin levels, serious adverse events, and study drug dosing logs. Non-critical data—such as routine vital signs at non-dosing follow-up visits or minor demographic parameters—are deprioritized.

Constructing the Integrated Quality and Risk Management Plan (IQRMP)

The IQRMP establishes the operational monitoring plan, defining the specific blend of centralized monitoring, remote data surveillance, and targeted on-site CRA monitoring visits.

Centralized Monitoring vs. On-Site Monitoring vs. Remote Monitoring

Modern clinical trials deploy a multi-tiered monitoring architecture. The CCRA candidate must clearly distinguish the operational scope and technical capabilities of each monitoring modality:

Centralized Monitoring

Centralised monitoring can provide the statistical and analytical engine of RBM. Performed remotely by centralized monitors, clinical data managers, and biostatisticians, it evaluates aggregate clinical data accumulating across all participating investigative sites concurrently.

Core Capabilities of Centralized Monitoring:

• Cross-Site Comparative Analytics: Comparing site performance metrics against national or global study means to detect statistical outliers.
• Data Fabrication & Anomaly Detection:
  • Digit Preference Analysis: Human beings fabricating data rarely generate true random distributions. A central monitor evaluating blood pressure readings or laboratory numbers will identify sites where 80% of recorded systolic values end in "0" or "5," warranting review of rounding, instrument output and recording practices.
  • Zero-Variance Detection: Identifying investigative sites where repeated laboratory measurements or visual analog scales show unnaturally identical values across multiple subjects.
  • Visit Date Clustering: Detecting patterns where multiple subjects are recorded as attending complex clinical visits at identical weekend or holiday hours.
• Underreporting Surveillance: Flagging sites that report zero adverse events, zero protocol deviations, or zero concomitant medications across large participant cohorts over extended treatment durations.

Official sources

Checked October 18, 2026: ICH E6(R3), Principles and Annex 1, ICH E8(R1).