(A-240) Enhancing The Meaningfulness of Qualitative Adverse Event Report Reviews Through Incorporation of Quantitative and Qualitative Aspects of Causality Assessment

Background: Safety signals are identified in Individual Case Safety Reports (ICSRs) through combined quantitative and qualitative approaches. Despite the benefits of this approach, the manual nature of qualitative review at a case level can be inefficient as most Adverse Event (AE) reports reviewed in isolation lack the necessary information to support assessment of causality.

Objectives: To retrospectively characterise the efficiency gain in qualitative review when incorporating the Clinical Utility Score for Prioritisation (CUSP; Kara et al 2023) approach alongside key qualitative elements of causality assessment.

Methods: AE reports subjected to qualitative review from November 2022 to November 2024 were utilised. The CUSP methodology was employed to determine scores closest to the bottom 25th and 50th percentiles. Two critical data elements commonly used in causality assessment, time-to-onset or positive de-challenge, provided further criteria for prioritisation.

Subsequently, 201 cases across both grouping were randomly selected across 3 distinct therapeutic areas and reviewed independently by two healthcare professionals to identify "clinically interesting" cases, with independent adjudication for discordance. Cases deemed clinically interesting were further reviewed for relevance to benefit-risk by a pharmacovigilance expert.

Results: Of the 184,525 AE reports reviewed, 50,842 cases (27.6%) had CUSP scores between 0-11, corresponding to the lowest 28th percentile, of which 64% (n=18,033) included de-challenge or time-to-onset information. Additionally, 56,596 cases (30.7%) had CUSP scores between 12-17, corresponding to the lowest 58th percentile, with 52% (n=29,388) including these attributes.

Following review of the 201 cases, 5 clinically interesting cases were identified, 2 in the 28th percentile and 3 in the 58th percentile, none of which could be immediately considered relevant to benefit-risk. These included 2 labelled events, 2 non-serious events lacking supporting clinical evidence, and 1 case where the event is listed for a co-reported medication.

Reprioritisation of such cases could occur as qualitative review resulted in 17.8% (32,809 of 184,525 cases) in the lowest 28th percentile and 14.7% (27,208 of 184,525 cases) in the 58th percentile being de-prioritised for review.

Conclusions: Qualitative case reviewed would have been streamlined by integrating aspects of causality assessment into case reviews by deprioritising cases that are not likely to impact benefit-risk assessments, thereby allowing experts to prioritize higher-value cases.

A limitation of our study is the small sample size, requiring further research to understand generalizability of our approach.