Biomedical Engineering Reference
In-Depth Information
6. Concluding
In this paper, we have reviewed several recently developed statistical mod-
eling procedures for data from an ODS scheme. These procedures include
semiparametric empirical likelihood and semiparametric estimated likeli-
hood. Semiparametric empirical likelihood can be used to deal with the
ODS scheme with an overall SRS sample and several supplemental samples
and the semiparametric estimated likelihood can be used to deal with the
ODS scheme with an overall SRS sample, several supplemental samples and
some information for the underlying population. These are robust methods
as they do not require parametric modeling of the underlying distribution of
covariates. Generally, an ODS design, coupled with an appropriate analysis,
can be a powerful alternative to commonly used sampling scheme.
A complexity in practical studies often involves the cluster- or center-
eects of the study subjects. In this situation a random eects model is
often used since it allows the investigators to interpret their results beyond
the limited participating centers. Zhou, You and Longnecker
16
has extended
the semiparametric empirical likelihood method developed by Zhou, et al.
15
to the setting with center-eects. However, there are no results for other
two semiparametric methods.
Another important case is that the response may be multivariate. An
example is the Collaborative Perinatal Project (CPP) about the study of
left and right hearing level. When the response is multivariate, extending
the above work to multiple responses is still an open problem.
Acknowledgments
This research is supported by a grant from National Institute of Health
(CA 79949)
References
1. J. Corneld, A method of estimating comparative rates from clinical data,
Applications to cancer of lung, breast, and cervix, Journal of National Cancer
Institute, 11 (1951), 1269-1275.
2. M.P. Longnecker, M.A. Klebano, H. Zhou, and J.W. Brock, Maternal serum
level of the DDT metabolite DDE is associated with premature and small-
for-gestational-age birth, Lancet, 358 (2001), 110-114.
3. A.B. Owen, Empirical likelihood ratio condence intervals for a single func-
tional, Biometrika, 74 (1988), 237-249.
4. A.B. Owen, Empirical likelihood for condence regions, Ann. Statist., 18
(1990), 90-120.
Search WWH ::
Custom Search