%PIC_NPMLE: A SAS Macro For Nonparametric Estimation In Partly Interval-Censored Survival Data

Content Provider	Semantic Scholar
Author	Zhu, Liang Rong Huang, Qinlei
Copyright Year	2015
Abstract	Partly interval-censored (PIC) data arise frequently in follow-up studies. An example of such data is provided by the St. Jude Lifetime Cohort study (SJLIFE), which follows childhood cancer survivors for late effects such as growth hormone deficiency (GHD). For some patients, the exact time of the first occurrence of GHD is reported, but for others, the exact time is unknown and is recorded as occurring between two clinic visits. PIC data is an important topic in medical research; however, no statistical software is available for analyzing it. In this paper, we provide two SAS™ macros to calculate the nonparametric maximum-likelihood estimator (NPMLE) of the survival function in PIC data. We estimate the NPMLE by using an iterative convex minorant (ICM) algorithm and an EM iterative convex minorant (EM-ICM) algorithm based on two different likelihood functions. Our simulation studies showed that both of the proposed algorithms provide consistent estimates for survival functions and the macro %PIC_NPMLE using EMICM algorithm computes much faster than the other macro using EM algorithm. Finally, we illustrate how to use the %PIC_NPMLE macro by applying it to GHD data from the SJLIFE study mentioned above.
File Format	PDF HTM / HTML
Alternate Webpage(s)	http://www.pharmasug.org/proceedings/2015/SP/PharmaSUG-2015-SP10.pdf
Language	English
Access Restriction	Open
Content Type	Text
Resource Type	Article