The current duration design for estimating the time to pregnancy distribution: a nonparametric Bayesian perspective

Area of Science:

• Biostatistics
• Reproductive Epidemiology
• Survival Analysis

Background:

• Estimating time-to-pregnancy (TTP) distributions is crucial for reproductive health research.
• The current duration design, while useful, faces challenges with accurately estimating short TTPs due to sampling biases.

Purpose of the Study:

• To investigate the identifiability and estimation of TTP distributions using current duration data.
• To address the instability in estimating short waiting times inherent in the current duration design.

Main Methods:

• Introduction of a novel Bayesian estimation method for TTP distributions.
• Mathematical proof of the asymptotic consistency of the proposed Bayesian method.
• Empirical evaluation using simulated data and real-world TTP data (Slama et al., 2012).

Main Results:

• The Bayesian method demonstrates asymptotic consistency for TTP distribution estimation.
• Comparison of the Bayesian approach with non-parametric maximum likelihood estimators (NPMLEs).
• Empirical studies confirm the robust performance of the Bayesian method.

Conclusions:

• The developed Bayesian method provides a more stable and reliable approach for estimating TTP distributions from current duration data.
• This method effectively addresses the limitations of previous techniques in handling short waiting times.
• The findings have implications for improving the accuracy of reproductive health studies.