|
Abstract:
|
Using the proportional odds model, we develop a method to evaluate covariate effects on time-to-pregnancy (TTP), the length of time that it takes a couple to conceive. Specifically, to assess the effect of within-woman menstrual cycle variance on TTP, we first investigate the cycle variance via a log linear model, then implement the estimated variance as a covariate in modeling TTP. An unbiased estimating equation is used for modeling the variance in cycle length and a pseudo-likelihood procedure is used to estimate the parameters in the proportional odds model. It is shown that the resulting estimator is asymptotically normal, with a closed form variance-covariance matrix and can be consistently estimated by the usual empirical method. We illustrate the proposed approach using the data from the Mount Sinai Study of Women Office Workers (MSSWOW).
|
