Abstract:Introduction Causality assessment of individual case safety reports (ICSRs) is an important step in pharmacovigilance case-level review and aims to establish a position on whether a patient's exposure to a drug is causally related to the patient experiencing an untoward adverse event. There are many different approaches for case causality adjudication, including the use of expert opinions and algorithmic frameworks; however, a great deal of variability exists between assessment methods, products, therapeutic classes, individual physicians, change of process and conventions over time, and other factors. Objective The objective of this study was to develop a machine learning-based model that can predict the likelihood of a causal association of an observed drug–reaction combination in an ICSR. Methods In this study, we used a set of annotated solicited ICSRs (50K cases) from a company post-marketing database. These data were enriched with novel supplementary features from external and internal data sources that aim to capture facets such as temporal plausibility, scientific validity, and confoundedness that have been shown to contribute to causality adjudication. Using these features, we constructed a Bayesian network (BN) model to predict drug–event pair causality assessment. BN topology was driven by an internally developed ICSR causality decision support tool. Performance of the model was evaluated through examination of sensitivity, positive predictive value (PPV), and the area under the receiver operating characteristic curve (AUC) on an independent set of data from a temporally adjacent interval (20K cases). No external validation was performed because of a lack of publicly available ICSRs with causality assessments for drug–event pairs. Results The model demonstrated high performance in predicting the causality assessment of drug–event pairs compared with clinical judgment using global introspection (AUC 0.924; 95% confidence interval [CI] 0.922–0.927). The sensitivity of the model was 0.900 (95% CI 0.896–0.904), and the PPV of the model was 0.778 (95% CI 0.773–0.783). Conclusion These results show that robust probabilistic modeling of ICSR causality is feasible, and the approach used in the development of the model can serve as a framework for such causality assessments, leading to improvements in safety decision making.

Applying causal inference and Bayesian statistics to understanding vaccine safety signals — a simulation study

Applying causal inference and Bayesian statistics to understanding vaccine safety signals using a simulation study

Bayesian Safety Surveillance with Adaptive Bias Correction

Safety signal detection with control of latent factors

A Bayesian method for safety signal detection in ongoing blinded randomised controlled trials

Bayesian learning of COVID-19 Vaccine safety while incorporating Adverse Events ontology

A Bayesian Group Sequential Approach to Safety Signal Detection.

Signalling Paediatric Side Effects using an Ensemble of Simple Study Designs

Non-parametric Bayesian mixture model to study adverse events of COVID-19 vaccines

Using Vaccine Safety Data to Demonstrate the Potential of Pooled Data Analysis

A Supervised Adverse Drug Reaction Signalling Framework Imitating Bradford Hill's Causality Considerations

Improved Apriori Method for Safety Signal Detection Using Post-Marketing Clinical Data

Combining a Pharmacological Network Model with a Bayesian Signal Detection Algorithm to Improve the Detection of Adverse Drug Events

Supervised Adverse Drug Reaction Signalling Framework Imitating Bradford Hill's Causality Considerations

Lessons Learned on Observed-to-Expected Analysis Using Spontaneous Reports During Mass Vaccination

Online causal inference with application to near real‐time post‐market vaccine safety surveillance

A Seasonality‐Adjusted Sequential Test for Vaccine Safety Surveillance

Leveraging Machine Learning to Facilitate Individual Case Causality Assessment of Adverse Drug Reactions

Discovering clinical drug-drug interactions with known pharmacokinetics mechanisms using spontaneous reporting systems and electronic health records

What would have happened anyway? Population data source considerations when estimating background incident rates of adverse events following immunisation to inform vaccine safety

Reps, Jenna M. and Garibaldi, Jonathan M. and Aickelin, Uwe and Soria, Daniele and Gibson, Jack E. and Hubbard, Richard B. (2014) Signalling paediatric side effects using an ensemble of simple study