9/15/2017

Post date: Sep 17, 2017 2:19:51 AM

Title: Assessing the distribution of time-to-event based on imperfect recall data

Speaker: Sedigheh Mirzaei Salehabadi, PhD, Eunice Kennedy Shriver National Institute of Child Health and Human Development

Abstract:

Recalled time-to-event is often encountered in cross-sectional observational studies. Thus, resulting in data that may be incomplete, for instance, current status or interval censored data etc. One can then use the (non)-parametric maximum likelihood estimators to assess the underlying distribution of interest based on considerable available literatures (for example, Kaplan-Meier, Turnbull or so). However, the chance of recall may depend on the time span between the occurrence of the event and the time of ascertainment (example, interview).

This talk will focus on recalled time-to-event data from two different studies: the age at menarche data from a recent Anthropometric Study, Kolkata, India and time-to-pregnancy from Upstate KIDS Study, New York, USA with different types of uncertainty contained in recalled time-to-event (continuous and discrete). We will discuss the estimation of the distribution of interest and their large sample properties based on the proposed models adapted to the special nature of the data at hand. Monte Carlo simulations indicate that the proposed estimators performing better than the existing models. We will provide detailed analyses of both the study data. We will conclude with a discussion on usefulness of the proposed method for Upstate KIDS Study. This talk is based on joint works with Professor Debasis Sengupta of ISI and Dr. Rajeshwari Sundaram of NICHD and the two study teams.