Statistical methods to analyze adverse events data of randomized clinical trials

Area of Science:

• Clinical Trials
• Pharmacovigilance
• Biostatistics

Background:

• Current analysis of adverse events in clinical trials often uses crude or exposure-adjusted incidence rates.
• These traditional methods fail to capture individual patient profiles with multiple or recurrent adverse events.
• Statistical assumptions for incidence rates, like constant hazard rates, are often unmet by clinical trial adverse event data.

Purpose of the Study:

• To introduce and demonstrate the utility of the nonparametric mean cumulative function (MCF) approach for analyzing recurrent adverse event data in randomized clinical trials.
• To provide a statistically valid inference method for patient safety profiles that overcomes limitations of traditional incidence rates.
• To enhance the understanding of drug safety profiles through a more comprehensive analysis of adverse events.

Main Methods:

• Employed a nonparametric statistical approach: the mean cumulative function (MCF).
• Applied the MCF method to an actual adverse event dataset from a clinical trial.
• Compared MCF estimates with traditional crude and exposure-adjusted incidence rates.

Main Results:

• The MCF approach provides valid statistical inference for recurrent adverse event profiles without assuming the form of the function.
• MCF estimates offer a more nuanced understanding of drug safety compared to conventional incidence rates.
• Demonstrated the practical applicability and utility of MCF in analyzing clinical trial safety data.

Conclusions:

• The mean cumulative function (MCF) is a superior method for analyzing recurrent adverse events in clinical trials.
• MCF enhances the interpretation of drug safety profiles by accounting for individual patient event histories.
• This nonparametric approach improves statistical inference for adverse event data, particularly when traditional assumptions are violated.