Joint Modeling of Multiple Skewed Longitudinal Processes with Excess of Zeroes and Time-to-Event Using a Shared Parameter Model

Joint modeling of longitudinal and time to event data has gained increasing attention in the recent past. Focused study of female fecundity is a relatively new field that provides challenging problems. Studying the patterns in menstrual cycle, for instance, can provide information on female fecundity. Naturally, it is of interest to study the association between the patterns in menstrual cycle and the time it takes for an individual to achieve a successful pregnancy. A menstrual cycle has two distinct phases, the pre-ovular phase and the post-ovular phase, governed by the levels of the luteinizing hormone. Subject specific covariates such as age, age at menarche, weight and smoking habits influence the lengths of the pre-ovular and post-ovular phases, thereby affecting the time to pregnancy. The distribution of pre-ovular lengths tend to be right skewed, due to abnormal cycles where ovulation occurs long after the standard period of days. In addition to skewness, post-ovular phase can also be of length zero. This is observed for cycles with no ovulation which can occur either due to successful pregnancies or due to anomalous causes. Distinguishing the causes for such a zero-length cycle is of interest, since an excessive number of such cycles for an individual can point to serious physiological issues, one such being infertility. This dissertation proposes a joint modeling approach to study the association between the two longitudinal processes, namely, the pre-ovular and the post-ovular phases, the covariates and the time to pregnancy. A mixed effects model with two correlated random effects was proposed to model the longitudinal processes and a discrete hazard model with a log-log link was used to model the time to pregnancy. The joint model was used to analyze the Oxford Conception Study and to observe the patterns in menstrual cycle lengths and their association with time to pregnancy.

Author

• Chatterjee, Somak

Language

• en

Keyword

• Female Fecundity
• Shared Parameter Model
• Excess of Zeros
• Longitudinal Data
• Joint Model
• Time to Event

Date created

• 2018

Type of Work

• Dissertation

License

• All rights reserved

Rights statement

• In Copyright

GW Unit

• Statistics

Degree

• Ph.D.

Advisor

• Sundaram, Rajeshwari
• Kundu, Subrata

Committee Member(s)

• Pan, Qing
• Modarres, Reza
• Ma, Ling
• Landon, Joshua

Persistent URL

•  https://scholarspace.library.gwu.edu/etd/9w032318x