A Scoping Review of Case-Level Causality Assessment Tools Developed Between 2008 and 2023; Strengths, Weaknesses and Potential Future Improvements
A Scoping Review of Case-Level Causality Assessment Tools Developed Between 2008 and 2023; Strengths, Weaknesses and Potential Future Improvements

A Scoping Review of Case-Level Causality Assessment Tools Developed Between 2008 and 2023; Strengths, Weaknesses and Potential Future Improvements

Drug Saf. 2026 Mar 30. doi: 10.1007/s40264-026-01665-7. Online ahead of print.

ABSTRACT

BACKGROUND: Various approaches to causality assessment have been developed that generally fall into the following three categories; expert judgement/global introspection, algorithms or probabilistic/Bayesian methods (or a combination of these). Causality assessment tools (CATs) are designed to assess the likelihood that a medicine has caused an adverse event in a specific individual.

OBJECTIVE: To provide an up-to-date assessment of existing CATs, the International Working Group on New Developments in Pharmacovigilance conducted a scoping review to identify CATs developed or updated between 2008 and 2023. The objective was to describe key components and strengths and weaknesses and to categorise CATs based on therapeutic area, clinical setting or patient population to support selection of the most appropriate tool within a specific context. We also discuss additional considerations for the assessment of adverse drug reactions seen with biologics.

METHODS: Searches were conducted in Embase and MEDLINE and to identify relevant grey literature. Review articles that described CATs developed between 1 January, 2008 until 31 December, 2023 (including updates to pre-existing CATs) were included. The key components of each CAT identified were extracted in addition to its strengths, weaknesses and other performance characteristics.

RESULTS: In total, 48 articles and 7 grey literature sources were eligible for inclusion; 18 CATs were identified, categorised as: global introspection (n = 1, 6%), algorithmic (n = 12, 67 %), hybrid (n = 4, 22%) and probabilistic (n = 1, 6 %). Algorithmic CATs included those for use in specific outcomes (severe cutaneous adverse reaction n = 1, drug-induced liver injury n = 4), or in specific populations/settings (paediatric n = 1, neonatal intensive care units n = 1). One CAT (World Health Organization tool for assessing adverse events following immunization, WHO-AEFI) was designed for use with vaccines.

CONCLUSIONS: Causality assessment tools designed to assess certain outcomes, such as drug-induced liver injury and severe cutaneous adverse reaction, may benefit from the future inclusion of biomarkers as predictors of risk. For specific biologics, such as immune checkpoint inhibitors where there is a heightened risk of immune-mediated AEs, there may be a role for biomarkers to help identify immune checkpoint inhibitor-induced toxicity. Additional considerations for causality may include drug quality, potential for medication error and assessment of adherence to risk minimisation measures.

PMID:41910704 | DOI:10.1007/s40264-026-01665-7