MOdified NARanjo Causality Scale for ICSRs (MONARCSi): A Decision Support Tool for Safety Scientists

Area of Science:

• Pharmacovigilance and drug safety assessment.
• Development and validation of clinical decision support tools.
• Statistical modeling for causality evaluation.

Background:

• Assessing drug-event causality in Individual Case Safety Reports (ICSRs) lacks a universal method, leading to assessment variability.
• Current methods rely on clinical judgment, probabilistic approaches, and algorithms, with no single accepted standard.
• Consistent and reliable causality assessment is crucial for effective pharmacovigilance.

Purpose of the Study:

• To develop and validate an Individual Case Safety Report (ICSR) Causality Decision Support Tool.
• To assist Safety Professionals (SPs) in performing consistent and documentable causality assessments.
• To introduce the MONARCSi model for standardized drug-event causality evaluation.

Main Methods:

• Developed a model incorporating Naranjo criteria, Bradford-Hill criteria, and internal practices, with nine drug-event pair features.
• Weighted features based on 65 safety professionals' assessments of importance using an ordinal scale.
• Used logistic regression to calculate the probability of a causal relationship and validated against 978 clinical trial drug-event pairs.

Main Results:

• The MONARCSi model demonstrated substantial agreement (Gwet Kappa = 0.77) with expert safety professionals' causality assessments.
• Achieved moderate sensitivity (65%), high specificity (93%), high positive predictive value (79%), and high negative predictive value (88%).
• The model's F1 score was 71%, indicating a strong balance between precision and recall.

Conclusions:

• The MONARCSi model shows potential as a valuable decision support tool for pharmacovigilance.
• It can assist safety professionals in evaluating drug-event causality consistently and documentably.
• This tool contributes to standardizing causality assessments in ICSRs.