A survival analysis approach to modeling human fecundity

Area of Science:

• Reproductive biology
• Biostatistics
• Epidemiology

Background:

• Accurate conception probability estimation is crucial for fertility treatments and identifying reproductive toxicants.
• Current statistical methods, survival analysis and hierarchical Bayesian models, have limitations for human reproduction data.
• Existing models do not fully account for the complex, timed processes of human reproduction.

Purpose of the Study:

• To propose a novel, biologically valid discrete survival model.
• To unify existing statistical approaches for estimating conception probabilities.
• To provide a flexible model for analyzing time-to-pregnancy (TTP) and day-specific conception probabilities.

Main Methods:

• Development of a discrete survival model integrating survival and hierarchical Bayesian approaches.
• Incorporation of covariate effects at both cycle and daily levels.
• Accounting for daily variations in conception probabilities within menstrual cycles.

Main Results:

• The proposed model successfully unifies TTP and day-specific conception probability estimation.
• It relaxes restrictive assumptions of previous models, offering greater biological validity.
• Simulations and cohort study data demonstrate the model's utility and flexibility.

Conclusions:

• The unified discrete survival model offers a more comprehensive approach to understanding human conception probabilities.
• This method enhances the ability to identify reproductive toxicants and inform fertility management.
• The model provides a valuable tool for researchers, with accompanying R code available.