Practical marginalized multilevel models

Clustered data analysis is characterized by the need to describe both systematic variation in a mean model and cluster-dependent random variation in an association model. Marginalized multilevel models embrace the robustness and interpretations of a marginal mean model, while retaining the likelihood inference capabilities and flexible dependence structures of a conditional association model. Although there has been increasing recognition of the attractiveness of marginalized multilevel models, there has been a gap in their practical application arising from a lack of readily available estimation procedures. We extend the marginalized multilevel model to allow for nonlinear functions in both the mean and association aspects. We then formulate marginal models through conditional specifications to facilitate estimation with mixed model computational solutions already in place. We illustrate the MMM and approximate MMM approaches on a cerebrovascular deficiency crossover trial using SAS and an epidemiological study on race and visual impairment using R. Datasets, SAS and R code are included as supplemental materials.

[1]  P. Albert,et al.  Models for longitudinal data: a generalized estimating equation approach. , 1988, Biometrics.

[2]  Brian Caffo,et al.  A User-Friendly Introduction to Link-Probit-Normal Models , 2006 .

[3]  Jeffrey D Blume,et al.  Likelihood methods for measuring statistical evidence , 2002, Statistics in medicine.

[4]  N. Jewell,et al.  To GEE or Not to GEE: Comparing Population Average and Mixed Models for Estimating the Associations Between Neighborhood Risk Factors and Health , 2010, Epidemiology.

[5]  M. Kenward,et al.  Design and Analysis of Cross-Over Trials , 1989 .

[6]  E. Demidenko,et al.  Mixed Models: Theory and Applications (Wiley Series in Probability and Statistics) , 2004 .

[7]  D. Bates,et al.  Nonlinear mixed effects models for repeated measures data. , 1990, Biometrics.

[8]  Michael G. Kenward,et al.  Design and Analysis of Cross-Over Trials, Second Edition , 2003 .

[9]  Brian S. Caffo,et al.  Unequal sampling for Monte Carlo EM algorithms , 2002 .

[10]  Scott L. Zeger,et al.  Marginalized Multilevel Models and Likelihood Inference , 2000 .

[11]  N. Breslow,et al.  Bias Correction in Generalized Linear Mixed Models with Multiple Components of Dispersion , 1996 .

[12]  C. McCulloch.  Maximum Likelihood Algorithms for Generalized Linear Mixed Models , 1997 .

[13]  T. Louis,et al.  Marginalized Binary Mixed‐Effects Models with Covariate‐Dependent Random Effects and Likelihood Inference , 2004, Biometrics.

[14]  Lewis B. Sheiner,et al.  Heteroscedastic nonlinear regression , 1988 .

[15]  L. Corrado.  Generalized Latent Variable Modeling: Multilevel, Longitudinal, and Structural Equation Models , 2005 .

[16]  Wolfgang Jank,et al.  A survey of Monte Carlo algorithms for maximizing the likelihood of a two-stage hierarchical model , 2001 .

[17]  M Davidian,et al.  Some general estimation methods for nonlinear mixed-effects models. , 1993, Journal of biopharmaceutical statistics.

[18]  Brian S. Caffo,et al.  A user-friendly tutorial on link-probit-normal models , 2006 .

[19]  Sanjoy K. Sinha,et al.  Robust Analysis of Generalized Linear Mixed Models , 2004 .

[20]  H. Goldstein.  Multilevel Statistical Models , 2006 .

[21]  E. Page,et al.  Approximations to the Cumulative Normal Function and its Inverse for Use on a Pocket Calculator , 1977 .

[22]  Joanne Katz,et al.  Socioeconomic Status and Visual Impairment Among Urban Americans , 1991 .

[23]  D. Bates,et al.  Approximations to the Log-Likelihood Function in the Nonlinear Mixed-Effects Model , 1995 .

[24]  Russell D. Wolfinger,et al.  Fitting Nonlinear Mixed Models with the New NLMIXED Procedure , 1999 .

[25]  T. Amemiya.  QUALITATIVE RESPONSE MODELS: A SURVEY , 1981 .

[26]  S. Zeger,et al.  Longitudinal data analysis using generalized linear models , 1986 .

[27]  James K. Lindsey,et al.  Nonlinear models in medical statistics , 2001 .

[28]  J. Ware,et al.  Random-effects models for serial observations with binary response. , 1984, Biometrics.

[29]  John M. Olin.  On MCMC sampling in hierarchical longitudinal models , 1999 .

[30]  Sl Zeger,et al.  Marginalized multilevel models and likelihood inference - Comments & Rejoinder , 2000 .

[31]  J. Kalbfleisch,et al.  A Comparison of Cluster-Specific and Population-Averaged Approaches for Analyzing Correlated Binary Data , 1991 .

[32]  Thomas A. Louis,et al.  Matching conditional and marginal shapes in binary random intercept models using a bridge distribution function , 2003 .

[33]  R. Royall.  Statistical Evidence: A Likelihood Paradigm , 1997 .

[34]  R. Wolfinger,et al.  Generalized linear mixed models a pseudo-likelihood approach , 1993 .

[35]  J. Robins,et al.  Adjusting for Nonignorable Drop-Out Using Semiparametric Nonresponse Models , 1999 .

[36]  John D. Kalbfleisch,et al.  Application of Likelihood Methods to Models Involving Large Numbers of Parameters , 1970 .

[37]  T. Liao.  Interpreting Probability Models: Logit, Probit, and Other Generalized Linear Models , 1994 .

[38]  P. Diggle,et al.  Analysis of Longitudinal Data. , 1997 .

[39]  M. Karim.  Generalized Linear Models With Random Effects , 1991 .

[40]  Brian Caffo,et al.  Flexible random intercept models for binary outcomes using mixtures of normals , 2007, Comput. Stat. Data Anal..

[41]  S. R. Searle,et al.  Generalized, Linear, and Mixed Models , 2005 .

[42]  N. Breslow,et al.  Approximate inference in generalized linear mixed models , 1993 .

[43]  P. Heagerty.  Marginally Specified Logistic‐Normal Models for Longitudinal Binary Data , 1999, Biometrics.