A modelling approach to the analysis of menstrual diary data

Area of Science:

• Clinical Trials Methodology
• Statistical Modeling in Reproductive Health
• Pharmacovigilance of Contraceptives

Background:

• Clinical trials for contraceptives require detailed daily recording of bleeding events by participants.
• Analysis of adverse events like prolonged bleeding or amenorrhoea is crucial for contraceptive safety and efficacy assessment.
• Traditional analysis methods may not fully capture inter-individual variability in bleeding patterns.

Purpose of the Study:

• To describe parametric statistical models for analyzing adverse events from contraceptive trial diary data.
• To specifically analyze the probability of prolonged bleeding or amenorrhoea and the lengths of bleeding episodes.
• To incorporate inter-individual variation into the analysis of contraceptive-related bleeding patterns.

Main Methods:

• Utilized linear logistic models to assess the probability of prolonged bleeding or amenorrhoea.
• Employed log-linear models to analyze the duration of bleeding episodes.
• Incorporated random effects in both models to account for variation between women.

Main Results:

• Demonstrated the application of parametric models for analyzing complex bleeding patterns in contraceptive trials.
• Successfully modeled the occurrence and duration of bleeding episodes, accounting for individual differences.
• Validated the methods using a World Health Organization (WHO) database.

Conclusions:

• Parametric models with random effects offer a robust approach to analyzing contraceptive trial data on bleeding events.
• These statistical methods enhance the understanding of adverse events such as prolonged bleeding and amenorrhoea.
• The described methodology provides valuable tools for contraceptive research and safety monitoring.