Understanding and Preventing Reporting Bias in Adverse Event Severity Ratings

Introduction: Why Severity Bias Matters

Accurate severity grading of adverse events (AEs) is critical for patient safety, regulatory compliance, and trial credibility. Yet, despite the availability of standardized tools like the Common Terminology Criteria for Adverse Events (CTCAE), severity ratings often suffer from reporting bias. This bias arises when investigators or sponsors consciously or unconsciously misclassify the intensity of an AE, either overstating or understating its seriousness. Regulatory agencies such as the FDA, EMA, and MHRA have repeatedly cited severity misclassification as a major inspection finding, emphasizing the need for consistent and unbiased severity reporting.

Reporting bias has downstream effects: it can distort benefit–risk profiles, delay safety signal detection, and compromise regulatory submissions. For example, underreporting severe AEs may conceal emerging risks, while overreporting minor events as severe can exaggerate drug toxicity. This article provides a comprehensive guide to understanding severity rating bias, its causes, regulatory expectations, case study examples, challenges, and best practices for mitigation.

Causes of Reporting Bias in Severity Ratings

Bias in severity ratings typically arises from multiple sources:

• Investigator subjectivity: Severity assessments often rely on clinical judgment, which can vary by investigator experience and interpretation.
• Therapeutic area influence: Oncology trials tend to report higher severity due to CTCAE familiarity, while non-oncology studies may misapply grading.
• Sponsor pressure: Investigators may consciously or unconsciously downgrade severity to avoid triggering expedited reporting.
• Lack of training: Inadequate familiarity with CTCAE criteria leads to inconsistent application across sites.
• Data entry shortcuts: Busy sites may default to moderate grading rather than carefully considering criteria.

For instance, in a cardiovascular trial, episodes of chest pain were repeatedly graded as “Mild” despite meeting criteria for “Severe” due to hospitalization, reflecting investigator underestimation and poor CTCAE application.

Regulatory Expectations for Severity Ratings

Global regulators mandate consistency and accuracy in severity reporting:

• FDA: Requires severity grading in IND safety reports and NDA/BLA submissions; discrepancies between eCRFs and narratives are frequently inspected.
• EMA: Demands severity alignment across EudraVigilance submissions, eCRFs, and SAE narratives.
• MHRA: Audits commonly flag underreporting of severity, especially when hospitalization criteria are ignored.
• ICH E2A/E2B: Identifies severity as a mandatory data element in clinical safety reporting standards.

Inspection reports show that severity bias is often linked to poor documentation. Regulators expect clear rationale for severity judgments, supported by CTCAE or equivalent criteria, rather than vague investigator impressions.

Case Studies Highlighting Severity Bias

Case Study 1 – Underreporting in Oncology: In a Phase III cancer trial, Grade 3 febrile neutropenia events were downgraded to Grade 2 in eCRFs to avoid expedited reporting. Regulators discovered discrepancies during SAE narrative review, leading to a major finding.

Case Study 2 – Overreporting in Vaccine Trial: In a vaccine trial, mild fever (below 38°C) was frequently classified as Grade 3 by inexperienced investigators. This inflated the perceived safety risks, delaying trial progress until retraining corrected grading errors.

Case Study 3 – Inconsistent Documentation: In a cardiology study, chest pain leading to hospitalization was graded inconsistently—Grade 2 at one site, Grade 3 at another—prompting regulatory queries on lack of harmonization.

Challenges in Eliminating Severity Bias

Even with standardized grading systems, sponsors face challenges in eliminating bias:

• Variability across sites: Multi-country trials face inconsistent interpretation of CTCAE criteria.
• Complexity of criteria: Some CTCAE categories involve lab values, clinical judgment, and intervention requirements, making grading difficult.
• Data reconciliation gaps: Severity inconsistencies often remain unreconciled between eCRFs, narratives, and pharmacovigilance databases.
• Operational constraints: High AE volume can pressure investigators to apply severity grades hastily.

These challenges emphasize the need for robust systems, continuous oversight, and training interventions to reduce reporting bias.

Best Practices for Mitigating Reporting Bias

To strengthen the accuracy of severity ratings, sponsors and CROs can adopt best practices:

• Provide comprehensive CTCAE training at study initiation and refresher sessions during the trial.
• Develop study-specific grading guides with practical examples relevant to the therapeutic area.
• Use eCRF edit checks that prompt investigators to confirm grading against predefined criteria.
• Conduct centralized medical review of severe or unexpected events to ensure consistent grading.
• Reconcile severity across SAE forms, narratives, and pharmacovigilance systems regularly.

For example, in a global immunotherapy trial, centralized safety review identified discrepancies in severity ratings across sites and implemented retraining. This reduced misclassification rates by 25% and improved regulatory compliance.

Regulatory Implications of Severity Misclassification

Misclassification of severity carries serious regulatory consequences:

• Missed expedited reporting: Downgrading severity can delay submission of life-threatening events.
• Regulatory findings: Agencies may issue critical observations for inconsistent severity reporting.
• Risk–benefit distortion: Misreporting can exaggerate or conceal safety signals, affecting drug approval decisions.
• Delayed database lock: Inconsistent severity ratings require reconciliation, prolonging trial timelines.

Regulators increasingly scrutinize severity consistency, making accurate grading both a scientific and compliance requirement.

Key Takeaways

Reporting bias in AE severity ratings undermines clinical trial integrity and poses regulatory risks. To mitigate these issues, sponsors and investigators must:

• Train investigators on CTCAE and regulatory expectations for severity reporting.
• Implement edit checks, centralized reviews, and reconciliation processes.
• Document rationale for all severity judgments in narratives and eCRFs.
• Regularly review severity trends to identify and correct systematic bias.

By applying these practices, trial teams can minimize severity misclassification, strengthen pharmacovigilance, and ensure compliance with global regulatory standards, ultimately improving patient safety and trial reliability.