Regression Analysis for Current Status Data - Interval-Censored-Time-to-Event Data: Methods and Applications - page 107

on the kernel function and the copula model, we tried different kernel functions

and copula models. For example, with the use of the triangular kernel function
and the Clayton model, we obtained

^ = 2:2916 and ^ = 0:6023 and the

the estimated standard errors are 0.2164 and 0.1785, respectively. In the case

of the FGM model defined as C (u;v) = uv + uv(1 u)(1 v) and the
normal kernel function, the results become

^ = 2:3314 and ^ = 0:7921

with the estimated standard errors of 0.4393 and 0.3409, respectively. These

results gave similar conclusions.

4.7 Discussion and Remarks

There is copious literature about current status data. Banerjee and Wellner

(2005) considered the current status data from a rubella study conducted in

Austria; Grummer-Strawn (1993) analyzed the current status data from a

breast-feeding study; Ding and Wang (2004) introduced a bivariate current

status data from a community-based study of cardiovascular diseases in Tai-

wan; and Jewell et al. (2005) discussed the analysis for bivariate current status

data on heterosexual transmission of HIV from the California Partners' Study.

Although there is extensive literature about current status data, the regres-

sion analysis, especially for the high-dimensional current status data, is still

very limited. In addition to the proportional hazards model, the proportional

odds model, and the linear transformation model introduced in this chapter,
several other models have also been considered to analyze current status data.

For example, Lin et al. (1998) and Martinussen and Scheike (2002) discussed

regression analysis with the additive hazards model, which is given by

(tjZ) = 0 (t) + Z 0 :