Abstract
We propose and compare two approaches for regression analysis of multilevel binary data when clusters are not necessarily nested: a GEE method that relies on a working independence assumption coupled with a three-step method for obtaining empirical standard errors, and a likelihood-based method implemented using Bayesian computational techniques. Implications of time-varying endogenous covariates are addressed. The methods are illustrated using data from the Breast Cancer Surveillance Consortium to estimate mammography accuracy from a repeatedly screened population.
| Original language | English (US) |
|---|---|
| Pages (from-to) | 381-398 |
| Number of pages | 18 |
| Journal | Biostatistics |
| Volume | 5 |
| Issue number | 3 |
| DOIs | |
| State | Published - Jul 2004 |
| Externally published | Yes |
Fingerprint
Keywords
- Conditional
- Endogeneity
- Hierarchical models
- Longitudinal data
- Marginal
- Transition models
ASJC Scopus subject areas
- Medicine(all)
- Statistics and Probability
- Statistics, Probability and Uncertainty
Cite this
Marginal modeling of multilevel binary data with time-varying covariates. / Miglioretti, Diana L; Heagerty, Patrick J.
In: Biostatistics, Vol. 5, No. 3, 07.2004, p. 381-398.Research output: Contribution to journal › Article
}
TY - JOUR
T1 - Marginal modeling of multilevel binary data with time-varying covariates
AU - Miglioretti, Diana L
AU - Heagerty, Patrick J.
PY - 2004/7
Y1 - 2004/7
N2 - We propose and compare two approaches for regression analysis of multilevel binary data when clusters are not necessarily nested: a GEE method that relies on a working independence assumption coupled with a three-step method for obtaining empirical standard errors, and a likelihood-based method implemented using Bayesian computational techniques. Implications of time-varying endogenous covariates are addressed. The methods are illustrated using data from the Breast Cancer Surveillance Consortium to estimate mammography accuracy from a repeatedly screened population.
AB - We propose and compare two approaches for regression analysis of multilevel binary data when clusters are not necessarily nested: a GEE method that relies on a working independence assumption coupled with a three-step method for obtaining empirical standard errors, and a likelihood-based method implemented using Bayesian computational techniques. Implications of time-varying endogenous covariates are addressed. The methods are illustrated using data from the Breast Cancer Surveillance Consortium to estimate mammography accuracy from a repeatedly screened population.
KW - Conditional
KW - Endogeneity
KW - Hierarchical models
KW - Longitudinal data
KW - Marginal
KW - Transition models
UR - http://www.scopus.com/inward/record.url?scp=15844390639&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=15844390639&partnerID=8YFLogxK
U2 - 10.1093/biostatistics/kxg042
DO - 10.1093/biostatistics/kxg042
M3 - Article
C2 - 15208201
AN - SCOPUS:15844390639
VL - 5
SP - 381
EP - 398
JO - Biostatistics
JF - Biostatistics
SN - 1465-4644
IS - 3
ER -