University of Wisconsin-Madison
School of Medicine and Public Health
Department of Biostatistics and Medical Informatics

Director: David L. DeMets, Ph.D.
Address, phone, e-mail


Advances throughout the biomedical sciences rely to an increasing degree on the principles and methods of statistical analysis. The ability of a lab or clinic to generate data often outpaces its ability to fully analyze that data; and this can limit research progress. Being a science concerned with the collection, analysis, and interpretation of data itself, statistics will play a critical role in resolving the information bottleneck facing biomedical scientists. The research mission of our program is to pursue the most important problems at the interface between statistics and biomedical science - the problems of biostatistics. The aims of our interdisciplinary training program are to recruit and provide pre-doctoral training in cardiovascular and pulmonary biostatistics to students who are interested in careers in biomedical science. The demand is high for scientists with this expertise. In contrast with traditional biostatistics training, the Interdisciplinary Training Program in Cardiovascular and Pulmonary Biostatistics emphasizes and further supports the interdisciplinary elements of biostatistical research. Through course work, pre-doctoral trainees learn theoretical, methodological, and practical underpinnings of statistics, and also relevant topics in biology, bioinformatics, clinical investigation, population-based investigation, and the responsible conduct of research. Through a novel lab rotation system, trainees become familiar with the biomedical context surrounding active investigations. Those who succeed in the program will be well positioned for further success in academia, government or industry. Contributing to the program are leading research/training departments, a high level of collaborative research across many disciplines related to cardiovascular and pulmonary medicine, and a proactive approach to increasing the domestic supply of undergraduates through summer internship programs. We know that biostatistics has a critical role to play in modern cardiovascular and pulmonary medical science, and so we have designed an interdisciplinary training program to most effectively train the next generation of biostatisticians.

Areas of Special Emphasis

The Department of Biostatistics and Medical Informatics in the School of Public Health is the home for this training grant. Faculty and staff are involved in the design and analysis of clinical trials, genetics, and other clinical, epidemiological and laboratory experiments. Areas of research expertise are analysis of longitudinal/functional data, Bayesian methods, clinical trials methodology, clustered data analysis, genomics, design of hybrid phase I/II/III trials, health services research, image analysis, observational studies, semiparametric inference and smoothing, sequential design, spatial statistics, statistical genetics and survival analysis.

The main track of the training program is a major enhancement of our longstanding PhD degree in Statistics with emphasis in Biostatistics. That program is based on PhD requirements in the Department of Statistics and specialized coursework and activities in the Department of Biostatistics and Medical Informatics. Trainees do course work in the theory, methodology, and application of statistics, taking specialized biostatistics courses covering clinical trials and epidemiology. As part of their specialization trainees take a course in a biological science, and receive training in the responsible conduct of research. Each semester they rotate into a different mentor's lab and thereby become directly engaged in active biomedical problems. During the summers trainees interact with undergraduates doing projects in the Computational Biology and Biostatistics program, which develops mentoring skills. Ultimately, each student pursues original scholarly research that is presented in a PhD thesis.

Emphasis in the interdisciplinary rotations will be on understanding the basic elements of a specific experiment or study, aspects of the biomedical context that motivates the experiment, the methods used, and sources of variation affecting whatever measurements are taken. For this interdisciplinary training program in cardiovascular/pulmonary biostatistics, biostatistics faculty mentors are joined by affiliated scientists at UW Madison whose research in the biological, genomic, and medical sciences addresses some of the most challenging biomedical problems we face, and does so using innovative approaches and exciting biotechnologies .

Type of Training: Pre-doctoral

Key Faculty Available as Preceptors:

Richard Chappell, PhD, serves as the senior biostatistician for the Radiation, Pediatric, and Cancer Control Oncology Programs for the UW Comprehensive Cancer Center. Dr. Chappell has general interests in clinical trial design and applications of survival analysis and generalized linear models. He has examined aspects of the design of phase III clinical trials with rare events and phase I trials of chronically administered drugs. He is interested in applying survival analysis to data from panel survey designs which yield complex patterns of missing data, such as truncation and interval censoring. Dr. Chappell has also conducted research in generalized linear models, particularly those with binary outcomes. He has developed and taught a special graduate course on categorical data analysis methods and generalized linear models. He regularly teaches the clinical trials course and coordinates the data analysis course for the Summer Institute for Training in Biostatistics.

Moo Chung, PhD, is working on image analysis and image smoothing problems in neuroanatomy and medical physics. His research concentrates on modeling geometric structures and patterns in images using tensor geometry, random fields, and data smoothing on image manifolds. He developed the first finite element method based estimation of the Laplacian on manifolds. Most recently he has developed a new type of kernel smoother called heat kernel smoothing that directly generalizes kernel smoothing to non-Euclidean space such as graphs and manifolds. His topics course Statistical Methods in Signal and Image Analysis is a possible elective for trainees.

Mark Craven, PhD, studies machine learning and bioinformatics. His current work is focused on developing and applying machine learning methods to the problems of elucidating, modeling and annotating biological networks. Several of Dr. Craven's projects involve analyzing sequence, expression and comparative-genomics data to predict sequence elements involved in gene regulation, and to predict regulatory interactions among genes. Dr. Craven's group also is working on developing algorithms and systems that aid the annotation of genomes, proteomes and high-throughput experiments by automatically extracting information from the scientific literature. Dr. Craven teaches a course on bioinformatics that is an elective for the predoctoral trainees.

David DeMets, PhD, interest focuses on methods for the analyses of clinical trials, sequential methods used for interim analysis, monitoring accumulating data for clinical trials, survival and longitudinal studies.Hehas developed and taught all of the biostatatics emphasis courses that are offered. He currently teaches an introductory course on clinical trials for medical fellows and pharmacy students. He has coauthored three texts on the field of clinical trials a fourth text is being prepared for the statistical methods in clinical trials course. He has made immeasurable contributions to UW and to the discipline of biostatistics.

Jason Fine, PhD, studies semiparametric methodology for time-to-event and longitudinal data. He has developed novel semiparametric models for the cumulative incidence function in the competing risks setting, investigated methods for interim analyses of clinical trials in which reporting delays are present, and studied flexible models for the analysis of quality of life data which arise in medical setting. Other research interests include statistical methods for genetics and imaging research. Dr. Fine has developed a topics course on statistical methods for human genetics. As a result of his interests in imaging, Dr. Fine has an affiliation with the Department of Radiology at the UW-Madison. He has participated as a co-investigator on the functional evaluation of renal artery stenosis using MRI.

Ronald Gangnon, PhD, statistical research interests include clustering, model selection, order-restricted inference, measurement reliability, survival analysis and interim monitoring. Recent research includes the development of a Bayesian approach for detecting and modeling spatial disease clustering; the development of a nonparametric method for assessing measurement reliability using bivariate isotonic regression; and the implementation of a randomization-based scheme for interim monitoring of multiple endpoints in clinical trials. Dr. Gangnon has collaborated with faculty in a wide range of disease areas, in particular asthma and cardiovascular disease. Dr. Gangnon serves as director of the Biostatstics Core for the Childhood Origins of Asthma (COAST) project.

Sunduz Keles, PhD, studies statistical methods for combining multiple sources of genomic data to understand the genetic code that specifies where and when genes are expressed in the cell. She has developed adaptive regression-based and constrained mixture model based approaches for identifying transcription factor binding sites. Dr. Keles worked closely with scientists from a biotech company to develop statistical methods for analyzing data from newly developed high-throughput ChIP-chip experiments. She collaborates with Dr. Emery Bresnick applying and improving her methods to analyze mouse ChIP-chip data of transcription factors GATA-1 and GATA-2. Her other methodological interests include developing methods for analyzing censored data including methods for bivariate survival data and cross-validation based methods for model selection with censored data.

Christina Kendziorski, PhD, research concerns the development and application of statistical methods to address questions arising in genetics and genomics based studies of complex diseases. Two main research areas are quantitative trait loci (QTL) mapping and microarray expression studies. In particular, her group investigated a microarray experimental design question concerning the utility of pooling subject samples. In addition to this work on experimental design, she has developed sensitive methods for time-course expression studies and for ETL mapping studies which combine microarray with QTL mapping data. She teaches an introductory biostatistics course and a graduate level course on Statistical Methods for Microarrays.

KyungMann Kim, PhD, is Director of the Comprehensive Cancer Center's Biostatistics Shared Resource. His areas of research includes group sequential methods and clustered data analysis. He is a member of the National Cancer Institute's Scientific Review Group, Subcommittee E on Cancer Epidemiology, Prevention and Control and of the National Institute of Allergy and Infectious Diseases' Therapeutic Data and Safety Monitoring Board for HIV/AIDS.

Bret Larget, PhD, studies statistical genomics, especially problems arising in the comparison of genomes. His work on the phylogenetic analysis of aligned molecular sequences has been central to the development of Markov chain Monte Carlo (MCMC) computational strategies in this area. He works on methods for genome rearrangement data in basic efforts to understand the history of genomes. He teaches several statistics courses for the biosciences, including a topics course Statistical Phylogenetics, a possible elective for trainees.

Mary Lindstrom, PhD, has 15 years of experience in translational research collaborations. She has developed new methods for analyzing data including a new model for radiation cell survival curves, an improved method for estimating lytic units, a "broken line" model for tumor cell relative movement data, and a non-parametric method for estimating isodose dose contours for seed radiation sources. Her statistical research focuses on semi-parametric shape invariant models for functional data. Functional data arise when the ideal observation for each experimental unit is a curve or function and the observed data consist of a set of noisy observations from each curve. Examples include sets of tumor re-growth curves or dose response curves from each of a group of individuals. Dr. Lindstrom has also developed a penalized estimate for the knots in free-knot splines which provide a flexible model for response curves, and which can be used to describe shape invariance arising in self-modeling.

Michael Newton, PhD, research concerns statistical inference in various problems from the biological sciences, in particular the use of stochastic models and advanced statistical computing. His theoretical analysis of bootstrap sampling in comparative genomics won him the George Snedecor Prize (1997) for a top theoretical paper in biometry awarded by the Committee of Presidents of Statistical Societies (COPSS). He has been at the forefront of methodological developments in several areas: Markov chain Monte Carlo (MCMC) methods for phylogenetic analysis, now widely adopted in comparative genomic calculations; empirical Bayes mixture model computations for assessing differential gene expression from microarray data; and the first use of MCMC to infer oncogenetic trees. In 2003 Dr Newton received the Spiegelman prize from the American Public Health Association recognizing outstanding contributions in health statistics for someone under 40. He recently received the prestigious Presidents' Award from COPSS. Dr. Newton recently served a four year term on the Genome Study Section. He has been serving as director of the Biostatistics Program at UW.

David Page, PhD, is an associate professor in the medical informatics group. He works on biomedical applications of data mining and machine learning, such as the application of data mining techniques to molecular bioassays, gene expression microarrays, single-nucleotide polymorphisms (SNPs), and proteomics (both protein levels and protein interactions). He currently is involved in a study of SNP and gene expression microarray data for multiple myeloma and monoclonal gammopathy of undetermined significance (MGUS). Within data mining, Dr. Page is an international leader in the areas of multi-relational data mining and inductive logic programming.

Mari Palta, PhD, is a Fellow of the American Statistical Association and Professor in the Department of Population Health Sciences and in Biostatistics and Medical Informatics. Dr. Palta's research interests include statistical methods for epidemiology and longitudinal studies and, in particular, methods to analyze longitudinal data with missing or unknown predictors, selection bias and measurement error. Her population health focus areas are in pediatrics, quality of life and respiratory health. She is Pi for two large epidemiologic cohort studies. One study follows a regional cohort of very low birth weight children in Wisconsin. These projects have regularly engaged biostatistics students. Dr. Palta teaches a course in advanced statistical methods for Population Health, based on her book "Quantitative Methods in Population Health: Extensions of Ordinary Regression." Dr. Palta also directs epidemiologic and health outcomes research projects which provide training in the analysis of observational, population based and clinical, data.

Jude Shavlik, PhD, is Professor of Computer Science and of Biostatistics and Medical Informatics; he was integral to the creation of the Medical Informatics Program here at the UW and serves as Director of the Medical Informatics Program. His research is focused on developing and applying machine learning methods for biomedical tasks such as microarray analysis and design, protein-structure determination, and information extraction from on-line biomedical text. Dr. Shavlik is co-director of the NLM-funded training grant for computational biology. He is founding (and current) member of the Board of Directors of the International Society for Machine Learning and was a founding member of the Board of Directors of the International Society for Computational Biology.

Grace Wahba, PhD, research involves the development and exploitation of new/improved flexible function estimation and model building methods, that are applicable to the analysis of data from the kinds of complex data structures that obtain in large demographic studies and clinical trials. Numerous published applications include analysis of a variety of data sets with a variety of complex structures from the Wisconsin Epidemiological Study of Diabetic Retinopathy, but these techniques are applicable more generally to risk factor modeling, selection of influential variables and classification for cancer data from complex demographic studies, clinical trials, microarray data and certain medical imaging applications. These methods are often able to extract information from data sets where the usual methods assuming linear or other parametric relationships fail. She holds the IJ Schoenberg-Hilldale Chair in Statistics and is a member of the National Academy of Sciences and the American Academy of Arts and Sciences and is a Fellow of the American Association for the Advancement of Science.

Brian Yandell, PhD, has research interests that combine statistical theory with biological applications, with particular attention to statistical genetics and microarray gene expression analysis. Dr. Yandell is co-investigator on an awarded NIH grant to use microarrays in a mouse genetic model for diabetes.

Last updated: January, 2007

Back to the top

Skip footer links and go to content
Twitter iconTwitterExternal link Disclaimer         Facebook iconFacebookimage of external link icon         YouTube iconYouTubeimage of external link icon         Google+ iconGoogle+image of external link icon