Approximate Inference in Generalized Linear Mixed Models

Breslow, Norman E.; Clayton, David

doi:10.1080/01621459.1993.10594284

Cited by 3,116 publications

(2,915 citation statements)

References 52 publications

Supporting

Mentioning

2,862

Contrasting

Unclassified

Order By: Relevance

“…Formal testing of whether the random effects are required in a model will depend on whether the best-fitting model is a generalized linear model, or a generalized additive model. For the former, we compared the fit using a generalized linear mixed-effects model (Breslow and Clayton, 1993;Wolfinger and O'Connell, 1993) while the latter, when there are significant nonlinear predictors, should be compared using a generalized additive mixed mode (Wang, 1998;Wood, 2004).…”

Section: Generalized Additive Mixed Model For Binary Outcomesmentioning

confidence: 99%

Methods for testing theory and evaluating impact in randomized field trials: Intent-to-treat analyses for integrating the perspectives of person, place, and time

Brown¹,

Wang²,

Kellam³

et al. 2008

Drug and Alcohol Dependence

152

139

View full text Add to dashboard Cite

Randomized field trials provide unique opportunities to examine the effectiveness of an intervention in real world settings and to test and extend both theory of etiology and theory of intervention. These trials are designed not only to test for overall intervention impact but also to examine how impact varies as a function of individual level characteristics, context, and across time. Examination of such variation in impact requires analytical methods that take into account the trial's multiple nested structure and the evolving changes in outcomes over time. The models that we describe here merge multilevel modeling with growth modeling, allowing for variation in impact to be represented through discrete mixtures-growth mixture models-and nonparametric smooth functions-generalized additive mixed models. These methods are part of an emerging class of multilevel growth mixture models, and we illustrate these with models that examine overall impact and variation in impact. In this paper, we define intent-to-treat analyses in group-randomized multilevel field trials and discuss appropriate ways to identify, examine, and test for variation in impact without inflating the Type I error rate. We describe how to make causal inferences more robust to misspecification of covariates * Corresponding author. Tel.: +1 813 974 6672. E-mail address: hbrown@health.usf.edu (C.H. Brown). Conflict of InterestAuthor Muthén is a co-developer of Mplus, which is discussed in this paper. There are no conflicts of interest. in such analyses and how to summarize and present these interactive intervention effects clearly. Practical strategies for reducing model complexity, checking model fit, and handling missing data are discussed using six randomized field trials to show how these methods may be used across trials randomized at different levels. NIH Public Access

show abstract

Section: Generalized Additive Mixed Model For Binary Outcomesmentioning

confidence: 99%

Methods for testing theory and evaluating impact in randomized field trials: Intent-to-treat analyses for integrating the perspectives of person, place, and time

Brown¹,

Wang²,

Kellam³

et al. 2008

Drug and Alcohol Dependence

152

139

View full text Add to dashboard Cite

show abstract

“…For example, Oliver et al ( 1992 ) used geostatistical kriging to analyze the patterns of rare diseases. A notable generalized linear mixed model technique was discussed in Breslow and Clayton ( 1993 ) who extended the disease-mapping approach of Clayton and Kaldor (1987 ) to include exposure using the penalized quasilikelihood method. In this work we use the Bayesian maximum entropy (BME ) mapping approach discussed in Christakos ( 1992Christakos ( , 1998Christakos ( , 2000.…”

Section: The Random Field Modelmentioning

confidence: 99%

“…This methodological viewpoint is based on a powerful combination of spatiotemporal random field theory with Bayesian maximum entropy analysis ( Christakos, 1992( Christakos, , 2000 , which accounts for various sources of knowledge ( scientific theories, soft data, uncertain observations, physical and biological laws, higher-order spatiotemporal moments, etc. ) that cannot be incorporated by existing approaches to the exposure± health effect problem (e.g., Breslow and Clayton, 1993;Briggs and Elliott, 1995 ). Other attractive features of the Bayesian maximum entropy approach ( nonlinear estimators, nonGaussian laws, etc. )…”

Section: Introductionmentioning

confidence: 99%

Spatiotemporal analysis of environmental exposure–health effect associations

Christakos

Serre

2000

J Expo Sci Environ Epidemiol

View full text Add to dashboard Cite

The goal of this work is to discuss a general methodology for studying associations between environmental exposures and health effect by means of the spatiotemporal random field theory. This theory is the tool of choice for rigorously accounting for important spatiotemporal variations and uncertainties related to exposures and effect. Within the framework of the random field theory, the Bayesian maximum entropy model neatly synthesizes various sources of physical and epidemiological knowledge into spatiotemporal analysis. Therefore, unlike technical statistics, this approach relies on the blending of substantive physical knowledge with powerful mathematical techniques and a coherent rationale. Given the well -founded fact that certain health effects may be caused by environmental exposures, the significance of these exposures is assessed in terms of a criterion that is based on the joint stochastic representation of exposure and health -effect distributions in space / time. In view of this criterion, the strength and consistency of the exposure ± effect association are evaluated on the basis of the health -effect predictions that the combined physico -epidemiologic analysis generates in space / time. The main features of the approach are demonstrated by a simulation example and a real case study involving mortality and cold temperatures in North Carolina. The studies demonstrated the practical usefulness of the stochastic human exposure analysis in assessing the exposure ± effect association. The results reported here emphasize the links between spatiotemporal models of physical systems and population health -effect distributions, thus suggesting directions for improving the current understanding of quantitative``exposure ± health effect'' functions.

show abstract

“…The model is a generalized linear mixed model (GLMM) [Breslow and Clayton, 1993;Burton et al, 1999]. If ij ¼ 0, (1) is just the standard logistic regression model of the effect of the measured genetic and environmental factors contained in z on risk of disease.…”

Section: Introductionmentioning

confidence: 99%

Ascertainment adjustment in complex diseases

Glidden

Liang

2002

Genetic Epidemiology

View full text Add to dashboard Cite

Genetic studies of complex diseases must confront two statistically difficult issues simultaneously. First, in many settings, to minimize the number of individuals to be genotyped, families enriched for disease must be oversampled. Also, statistical models in family studies should allow for residual association. This association will represent unmeasured genetic and environmental factors influencing disease risk. Dealing with these features simultaneously is both compelling and challenging. Burton et al. [2000] (Am. J. Hum. Gen. 69: 1505–14) recently discussed this issue and suggested that ascertainment corrections may lead to problematic parameter estimation. We revisit the issues and examples of Burton et al. [2000] (Am. J. Hum. Genet. 69: 1505–14) and present a more optimistic assessment. Estimation in this context is conceptually straightforward, but may be more problematic in practice. Specifically, we find that even slight misspecification of the random effects distribution in ascertainment‐adjusted likelihood can yield severely biased parameter estimates. This result should make scientists wary when interpreting results from ascertainment‐adjusted variance‐component models. Genet. Epidemiol. 23:201–208, 2002. © 2002 Wiley‐Liss, Inc.

show abstract

Approximate Inference in Generalized Linear Mixed Models

Cited by 3,116 publications

References 52 publications

Methods for testing theory and evaluating impact in randomized field trials: Intent-to-treat analyses for integrating the perspectives of person, place, and time

Methods for testing theory and evaluating impact in randomized field trials: Intent-to-treat analyses for integrating the perspectives of person, place, and time

Spatiotemporal analysis of environmental exposure–health effect associations

Ascertainment adjustment in complex diseases

Contact Info

Product

Resources

About