Improving the analysis of adverse event data in randomized controlled trials

Area of Science:

• Clinical Trials
• Pharmacovigilance
• Biostatistics

Background:

• Analyzing treatment-induced harm in randomized controlled trials (RCTs) is crucial but hampered by small sample sizes.
• Established practices for efficacy outcome analysis contrast with limited progress in adverse event (AE) analysis.

Purpose of the Study:

• To examine key issues in adverse event (AE) analysis within RCTs.
• To propose improvements for valuing and analyzing harm data in clinical research.
• To offer strategies for better selection and interpretation of AE results in publications.

Main Methods:

• The commentary critically reviews current practices in AE analysis.
• It discusses reframing research questions to focus on detecting adverse reaction signals.
• It advocates for Bayesian analyses and standardized reporting of AE data.

Main Results:

• Harm data in RCTs is often undervalued and AE analysis presents significant difficulties.
• Current AE analysis practices are frequently unsatisfactory, particularly regarding result selection and interpretation.
• Bayesian methods can facilitate cumulative harm assessment across trial phases.

Conclusions:

• Improved strategies are needed to enhance the analysis and reporting of adverse events (AEs) in clinical trials.
• Recommendations include using core outcome sets, focusing on serious and pre-specified events, and analyzing discontinuation events.
• Adopting new practices can lead to better realization of harm data value from RCTs.