Biomedical Engineering Reference
In-Depth Information
Section 1.6, and the analysis of competing risks interval-censored data in Sec-
tion 1.7. Section 1.8 contains some general remarks and discussion about some
topics that are not touched upon in the previous sections. However, two im-
portant topics, the analysis of current status data and Bayesian approaches for
interval-censored data, will not be discussed as they are addressed in Chapters
3 through 7.
1.2
More Introduction
To further introduce interval-censored failure time data and set the framework
for the sections below, we first present a real and well-known set of interval-
censored failure time data in this section. The data structure or the censoring
mechanism behind interval-censored data is then discussed, as well as some
key differences between right-censored data and interval-censored data.
A set of interval-censored failure time data that has been investigated by
many authors is that arising from a breast cancer study (Finkelstein and Wolfe
(1985); Sun (2006)). The data consist of 94 early breast cancer patients treated
at the Joint Center for Radiation Therapy in Boston between 1976 and 1980.
For their treatments, the patients were given either radiation therapy alone
(46 patients) or radiation therapy plus adjuvant chemotherapy (48 patients).
Each patient was supposed to have clinic visits every 4 to 6 months to be
examined for cosmetic appearance such as breast retraction. However, actual
visit times differ from patient to patient, and the times between the visits
also vary. As a consequence, with respect to the breast retraction time, only
interval-censored data were observed. Specifically, among the 94 patients, 38
of them did not experience breast retraction during the study, giving right-
censored observations for the time. For all other patients, intervals of the
form (a;b] were observed for their breast retraction times. Here the interval
Search WWH ::
Custom Search