Mixtures of marginal models

Rosen, Ori

doi:10.1093/biomet/87.2.391

Cited by 61 publications

(72 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We propose use of an EM algorithm for maximization of the composite-likelihood with respect to the hazard parameters. Although the EM algorithm is traditionally used for maximumlikelihood estimation in missing data problems, a similar algorithm can be more generally applied to any method that is based on unbiased estimating equations [e.g., Rosen et al, 2000]. In our application, the E-step of the algorithm involves computing the conditional probability of each relative being a carrier and a noncarrier, given their individual event history and the genotype of the index proband.…”

Section: Estimationmentioning

confidence: 99%

Adjustment for competing risk in kin‐cohort estimation

Chatterjee

Hartge

Wacholder

2003

Genetic Epidemiology

View full text Add to dashboard Cite

Kin-cohort design can be used to study the effect of a genetic mutation on the risk of multiple events, using the same study. In this design, the outcome data consist of the event history of the relatives of a sample of genotyped subjects. Existing methods for kin-cohort estimation allow estimation of the risk of one event at a time with the assumption that the censoring events are unrelated to the genetic mutation under study. These methods, however, may produce biased estimates of risk when multiple events are related to the genetic mutation, and follow-up of some of the events may be censored by the onset of other events. Using a competing risk framework to address this problem, we show that cause-specific hazard functions for carriers and noncarriers are identifiable from kin-cohort data. For estimation, we propose an extension of a composite-likelihood approach we described previously. We illustrate the use of the proposed method for estimation of the risk of ovarian cancer from BRCA1/2 mutations in the absence of breast cancer, based on data from the Washington Ashkenazi Kin-Cohort Study. We also evaluate the performance of the proposed estimation method, based on simulated data that were generated following the setup of the Washington Ashkenazi Study.

show abstract

Section: Estimationmentioning

confidence: 99%

Adjustment for competing risk in kin‐cohort estimation

Chatterjee

Hartge

Wacholder

2003

Genetic Epidemiology

View full text Add to dashboard Cite

show abstract

“…As each C gi (t), conditional on group specific frailties ν gi , is a nonhomogeneous Poisson process, ability to self-adapt to varying levels of data complexity and provides a rich and flexible framework for counting process models. Quasi-likelihood inference (Heyde, 1997) for this model is discussed in Nielsen and Dean (2007) and utilizes an algorithm similar to the expectation-solution algorithm of Rosen et al (2000). Figure 4 shows plots of the estimated baseline intensities, λ g0 (t), and the estimated treatment effects, γ g1 , of the three-component model fitted to the cherry bark tortrix data.…”

Section: Modeling the Cherry Bark Tortrix Datamentioning

confidence: 99%

Spatial and mixture models for recurrent event processes

2007

View full text Add to dashboard Cite

SUMMARYStudies of recurring infection or chronic disease often collect longitudinal data on the disease status of subjects. Two types of models may be envisioned for the analysis of such data: counting process models or multi-state transitional models. We consider both scenarios in the specific case where the population consists of mixtures. A flexible semi-parametric model for analyzing longitudinal panel count data is presented. Discrete mixtures of smooth counting process intensity forms are considered, including mixtures of splines, which permit time-varying covariate effects, with the so-called proportional intensity model as a limiting case. For recurrent events handled in a multi-state transitional model framework, individuals may be said to occupy one of a discrete set of states and interest centers on the transition process between states. We examine the use of mixed Markov models for the analysis of such longitudinal data where the processes corresponding to different subjects may be correlated spatially over a region. Both discrete and continuous-time models incorporating spatially correlated random effects are discussed. Examples illustrate the methods discussed including a study of recurrent weevil infestation, and one to assess the effectiveness of a pheromone treatment in disturbing the mating habits of the cherry bark tortrix moth.

show abstract

“…The MES algorithm is motivated from the ES (Expectation-Solution) algorithm (Hall and Zhang, 2004;Rosen, Jiang and Tanner, 2000) when the data are clustered. The ES algorithm combines elements of both GEE (Liang and Zeger, 1986) and the EM algorithms, so that one can account for dependency (clustering) in the data.…”

Section: Mes Algorithm Based On a Modified Newton-raphson Methodsmentioning

confidence: 99%

“…The ES algorithm combines elements of both GEE (Liang and Zeger, 1986) and the EM algorithms, so that one can account for dependency (clustering) in the data. However, the ES algorithm as prescribed by Rosen, Jiang and Tanner (2000) has a major limitation in that it is only applicable to an exponential dispersion family which has a form of 0 ÐC à 34 ) cannot be re-expressed in the exponential dispersion family form. As a consequence of this, the expectation of is not only related to but also to making a regression formulation complicated. ]…”

Section: Mes Algorithm Based On a Modified Newton-raphson Methodsmentioning

confidence: 99%

“…In this project, we illustrate two statistical methodologies: MES(Modified Expectation-Solution) algorithm and MPL(Maximum Pseudo Likelihood) method. The MES algorithm, one of our proposed methods is a modified version of an ES algorithm which is introduced by Rosen, Jiang and Tanner, 2000. The MES algorithm shows how to account for the clustered count data with excessive zeros based on a CMP regression model.…”

Section: A Marginal Model Approach To Zero-inflated Clustered Count Datamentioning

confidence: 99%

See 1 more Smart Citation

Inference for a zero-inflated Conway-Maxwell-Poisson regression for clustered count data.

Choo‐Wosoba

View full text Add to dashboard Cite

Mixtures of marginal models

Cited by 61 publications

References 27 publications

Adjustment for competing risk in kin‐cohort estimation

Adjustment for competing risk in kin‐cohort estimation

Spatial and mixture models for recurrent event processes

Inference for a zero-inflated Conway-Maxwell-Poisson regression for clustered count data.

Contact Info

Product

Resources

About