New Methods - Speakers and Presentations
Saturday, October 14, 2:30-3:30 p.m.
"Multivariate regression models for analyzing data from multiple informants"
Presentation outline is available by clicking here.
Garrett Fitzmaurice , Associate Professor of Medicine (Biostatistics) at the Harvard Medical School
Dr. Fitzmaurice has research interests in methods for analyzing discrete longitudinal data, models for mixed discrete and continuous outcomes, general missing data problems, methods for detecting and adjusting for overdispersion, and statistical problems in psychiatric epidemiology. A major focus of his research is on the development of statistical methodology for analyzing discrete multivariate data. In particular, he has developed likelihood-based methods for analyzing multivariate binary data arising from longitudinal studies. Recently, he has begun developing methods for handling non-ignorable dropouts in longitudinal clinical trials. Dr. Fitzmaurice's research in psychiatric epidemiology has concentrated on methodological problems surrounding the use of multiple informant data in mental health surveys. Dr. Fitzmaurice is also an affiliate of Brigham & Women's Hospital, the Harvard School of Public Health and the Karolinska Institute in Sweden.
Abstract: "Multivariate regression models for analyzing data from multiple informants"
Presentation outline is available by clicking here.
In community or service-based mental health research, assessments of psychopathology, as well as service use, are commonly collected using multiple informants. For example, in assessments of a child's mental health status, parents, teachers, children and others have served as informants on the underlying psychopathology of the child. In this lecture we review recent developments in statistical methods for the analysis of "multiple source" data. We use the term "multiple source" data to encompass data that are simultaneously obtained from multiple informants or raters (e.g., self-reports, family members, health care providers, administrators) or via different/parallel instruments or methods (e.g., symptom rating scales, standardized diagnostic interviews, or clinical diagnoses). We outline some of the methodological challenges regarding the analysis of multiple source data and discuss recently developed methodology for analyzing ratings from multiple sources (used as outcomes or used as predictors) based on multivariate extensions of generalized linear regression models.
Selected Publications:
O'Brien, L.M., Fitzmaurice, G.M. and Horton, N.J. (2006) "Maximum likelihood estimation of marginal pairwise associations with multiple source predictors." Biometrical Journal, In Press.
O'Brien, L.M. and Fitzmaurice, G.M. (2005). "Regression models for the analysis of longitudinal Gaussian data from multiple sources." Statistics in Medicine, 24, 1725-1744.
Horton, N.J. and Fitzmaurice, G.M. (2004). "Regression analysis of multiple source and multiple informant data from complex survey samples." Statistics in Medicine, 23, 2911-2933.
O'Brien, L.M. and Fitzmaurice, G.M. (2004). " Analysis of longitudinal multiple source binary data using generalized estimating equations." Applied Statistics, 53, 177-93.
Fitzmaurice G.M., Laird N.M. and Ware J.H. (2004). "Applied Longitudinal Analysis." New York: John Wiley and Sons.
Rubio-Stipec, M., Fitzmaurice, G.M., Murphy, J. and Walker, A. (2003). "The use of multiple informants in identifying the risk factors of depressive and disruptive disorders: Are they interchangeable?" Social Psychiatry and Psychiatric Epidemiology, 38, 51-58.
Goldwasser, M. and Fitzmaurice, G.M. (2001) "Multivariate linear regression analysis of childhood psychopathology using multiple informant data." International Journal of Methods in Psychiatric Research, 10, 1-10.
