Conference Schedule

8:15 – 9:00
Registration and Continental Breakfast

9:00 – 9:10
Introductory Remarks
Boris Iglewicz, Temple University

9:10 – 9:55
Models for Biosecurity: A Case Study in Anthrax
Ron Brookmeyer, Johns Hopkins University

9:55 – 10:40
Two-Stage Modeling of Clustered Non-Gaussian Data with Genetic Applications
Inna Chervoneva, Thomas Jefferson University

10:40 – 11:05

11:05 – 11:50
The Impact of Sequential and Adaptive Trial Clinical Designs on the Confidentiality of Interim Data
Susan S. Ellenberg, University of Pennsylvania

11:50 – 12:35
Antiretroviral Immunotherapy Trials: Tackling Missing Data and Non-Normality
Devan V. Mehrotra and Robin Mogg, Merck Research Laboratories

12:35 – 1:40

1:40 – 2:25
(Data) Size Does Matter, But You Might Be In for a Surprise…
Xiao-Li Meng, Harvard University

2:25 – 3:10
On the Joint Analysis of Randomized Controlled Trial and Observational Study Data
Ross L. Prentice, Fred Hutchinson Cancer Research Center

3:10 – 3:25
General Discussion


Models for Biosecurity: A Case Study in Anthrax

Ron Brookmeyer, Department of Biostatistics, Johns Hopkins Bloomberg School of Public Health

Anthrax is of increasing public health concern because of its potential use in acts of bioterrorism. Critical public health questions include the role of antibiotics, vaccine, and disease surveillance in preparing and responding to an anthrax outbreak. The challenges we face are that there are relatively few scientific studies that directly address these questions. In this talk we show how statistical reasoning and models can help address these questions and link together various scientific studies in a unifying model. We develop a competing risks model to account for the dynamics of anthrax spore germination and clearance. The model allows us to predict the effects of antibiotics, vaccines, and the importance of early detection of an outbreak in reducing morbidity and mortality. We used data from the 2001 anthrax outbreak in the United States, an anthrax outbreak in Russia, and studies from primates. We estimate the incubation period of disease which agrees well with empirical findings. The models can be used to determine the duration of times persons exposed to anthrax should remain on antibiotics to prevent disease, and the additional preventative value of vaccines. The talk illustrates how statistical reasoning together with statistical and mathematical models can be used to help develop effective public health response policies concerning disease surveillance, vaccinations and antibiotics.

Two-Stage Modeling of Clustered Non-Gaussian Data with Genetic Applications

Inna Chervoneva, Department of Medicine, Thomas Jefferson University

For clustered data with sufficiently large number of continuous measures per cluster, we consider analysis approaches in situations when conditional distributions do not support the normality assumption and may vary substantially in distributional shape. We propose a general class of hierarchical models that generalizes the global two-stage (GTS) method for non-linear mixed effects models. The second stage model is a standard linear mixed effects model with normal random effects, but the first stage model for cluster-specific distributions, conditional on random effects, is general enough to include completely parametric as well as semi- and non-parametric approaches. This methodology provides a flexible framework for modeling not only a location parameter but also other characteristics of conditional distributions that may be of specific interest. The data that motivates this work and is used to illustrate the proposed methodology comes from genetic studies of collagen fibril development in various organs in mice.

The Impact of Sequential and Adaptive Trial Clinical Designs on the Confidentiality of Interim Data

Susan S. Ellenberg, University of Pennsylvania School of Mediciney

It has long been recognized that maintaining the confidentiality of accumulating data from randomized clinical trials is critical to ensure the integrity of the study and the reliability of the final results. Experience has shown that investigators who are aware of interim results may alter their patterns of patient recruitment; further, although it is harder to document, it is clear that knowledge of interim data may affect judgments about patient outcomes and about considerations for protocol changes.

The practice of sequential analysis and interim decision-making (usually by a data monitoring committee) does provide some interim information on the possible study results, but not enough to allow for reliable guessing as to study outcome. Recently, new types of sequential designs, permitting not just early stopping but re-estimation of sample size based on interim results, have been proposed. One concern that has been raised about these new types of adaptive designs is that their use will provide much more information to investigators, companies and other interested parties about the likely outcome of the trial, and may thereby threaten study integrity. This presentation will address the extent to which various types of sequential designs and approaches to interim decision-making may impact on the confidentiality of emerging trial results.

Antiretroviral Immunotherapy Trials: Tackling Missing Data and Non-Normality

Devan V. Mehrotra and Robin Mogg, Merck Research Laboratoriesy

For many HIV infected patients, use of antiretroviral therapy (ART) results in a sustained suppression of plasma viral load to undetectable levels. However, due to lack of antigenic stimulation, this may also result in a gradual loss of HIV-specific cell mediated immune (CMI) responses that help control HIV infection. In concept, augmenting ART with periodic administrations of an HIV vaccine that boosts CMI responses could enhance control of viral replication. Researchers are now designing “antiretroviral immunotherapy” (ARI) trials to test this novel concept. In a typical ARI trial, HIV-infected patients with sustained viral suppression will receive inoculations of an experimental HIV vaccine or a placebo, and subsequently stop taking all their antiretroviral drugs. The goal is to assess whether the plasma viral loads during the ART interruption phase are generally lower in the vaccine group. Assessment of a vaccine effect will be challenging if some subjects resume ART or drop out before the scheduled end of the treatment interruption phase. To tackle this anticipated “missing” data problem in ARI trials, we propose a two-step approach: multiple imputation of the missing values followed by use of a non-parametric or robust parametric method. We use simulations to illustrate the power advantages of our proposed methods over other “standard” methods currently being considered for ARI trials. Our proposed approach is general enough for the robust analysis of incomplete longitudinal data in other therapeutic areas as well.

(Data) Size Does Matter, But You Might Be In for a Surprise …

Xiao-Li Meng, Department of Statistics, Harvard Universityy

One of the most frequently asked questions in statistical practice, and indeed in general quantitative investigations, is “What is the size of the data?” (i.e., how much information there is in the data?) A common wisdom underlying this question is that the larger the size, the more trustworthy are the results. Although this common wisdom serves well in many practical situations, sometimes it can be devastatingly deceptive. This talk will report two of such situations: a historical epidemic study (McKendrick, 1926) and the most recent debate over the validity of multiple imputation inference for handling incomplete data (Meng and Romero, 2003). McKendrick’s mysterious and ingenious analysis of an epidemic of cholera in an Indian village provides an excellent example of how an apparently large sample study (e.g., n=223), under a naive but common approach, turned out to be a much smaller one (e.g., n<40) because of hidden data contamination. The debate on multiple imputation reveals the importance of the self-efficiency assumption (Meng, 1994) in the context of incomplete-data analysis. This assumption excludes estimation procedures that can produce more efficient results with less data than with more data. Such procedures may sound paradoxical, but they indeed exist even in common practice. For example, the least-squared regression estimator may not be self-efficient when the variances of the observations are not constant. The morale of this talk is that in order for the common wisdom “the larger the better” be trusted, we not only need to assume that data analyst knows what s/he is doing (i.e., an approximately correct analysis), but more importantly that s/he is performing an efficient, or at least self-efficient, analysis.

On the Joint Analysis of Randomized Controlled Trial and Observational Study Data

Ross L. Prentice, Fred Hutchinson Cancer Research Center, University of Washington

Circumstances in which both randomized controlled trial data and cohort study data are both available in a treatment or exposure of interest may provide an opportunity to identify sources of bias or other limitations in either study design. The presence of the randomized trial data may permit a useful adjustment to observational data analysis, thereby allowing reliable assessment of research questions beyond those examined in the trial; for example, questions related to longer term exposures of the treatments studied in the trial, or to different dose levels, schedules, or preparations. Also, the trial and adjusted observational study data may combine to enhance study power related to treatment effects in subsets or for other purposes. These topics will be illustrated through analysis of postmenopausal hormone therapy in relation to cardiovascular disease and breast cancer.


Ron Brookmeyer is a Professor of Biostatistics at the Johns Hopkins Blooomberg School of Public Health, where he also directs the Masters of Public Health Program. Dr. Brookmeyer’s research is focused on statistical methods and models in epidemiology. He has worked extensively on methods in HIV/AIDS. Dr. Brookmeyer is a Fellow of the American Association for the Advancement of Science and the American Statistical Association. He received the Spiegelman Gold Medal in health statistics from the American Public Health Association. He received his Ph.D in statistics from the University of Wisconsin and B.S from Cooper Union.

Inna Chervoneva is an Assistant Professor at the Biostatistics Section, Department of Medicine, Thomas Jefferson University, Philadelphia, PA. She earned a BA in Mathematics from Kharkov State University, Kharkov, Ukraine, in 1991; a MA in Mathematics (1996) and a MS in Statistics (1999) from Temple University; and a PhD in Statistics from Temple University in 2003. Inna joined Jefferson’s Biostatistics Section in February 2000, first as a master level biostatistician, and then became a faculty member in July 2003. Her research interests include nonparametric statistical methods, mixed effects models, and statistical methods for gene expression assays and cancer marker models. She has authored or co-authored 12 publications including articles in Biometrika, the American Journal of Human Genetics, Cancer, and the Journal of Cell Biology. In 2003, she received the ENAR Spring Meeting Student Paper Award and IMS Annual Meeting Laha Award.

Susan S. Ellenberg is Professor of Biostatistics, Center for Clinical Epidemiology and Biostatistics; and Associate Dean for Clinical Research, University of Pennsylvania School of Medicine. From 1993 to 2004 she served as Director, Office of Biostatistics and Epidemiology in the Center for Biologics Evaluation and Research (CBER) at the U.S. Food and Drug Administration. Prior to that she led the Biostatistics Research Branch at the Division of AIDS, National Institute of Allergy and Infectious Diseases, and served in the Biometric Research Branch in the Cancer Therapy Evaluation Program, National Cancer Institute. During Dr. Ellenberg’s tenure at FDA she played a leading role in the development of international standards for design and analysis of clinical trials performed by the pharmaceutical industry, developed productive programs for postmarketing safety surveillance of biological products, and coordinated the development of a guidance document on clinical trial data monitoring committees. Dr. Ellenberg has published extensively in both statistical and medical journals, on topics including surrogate endpoints, data monitoring committees, clinical trial designs, adverse event monitoring, vaccine safety and special issues in cancer and AIDS trials. She is a Fellow of the American Statistical Association and the American Association for the Advancement of Science, and is an elected member of the International Statistical Institute. Her recent book, Data Monitoring Committees in Clinical Trials: A Practical Perspective, co-authored with Drs. Thomas Fleming and David DeMets, was named Wiley Europe Statistics Book of the Year for 2002.

Dr. Ellenberg received an A.B. from Radcliffe College and a Ph.D. in mathematical statistics from the George Washington University.

Devan V. Mehrotra has worked as a biostatistician for almost 15 years, and is currently Director, Clinical Biostatistics, at Merck Research Laboratories, Blue Bell, PA. He received his PhD in Statistics from the University of Delaware. He is a 4-time winner of the best contributed biopharmaceutical paper award at the annual Joint Statistical Meetings, and has published papers on a variety of research topics. He holds adjunct faculty appointments at the University of Pennsylvania and Villanova University, and has served as an Associate Editor for the Journal of Biopharmaceutical Statistics and the Biometrical Journal. His other professional activities have included serving as President of the Philadelphia Chapter of the American Statistical Association, serving on the steering committee of the annual FDA/Industry Statistics Workshop, and serving as a consultant to the Center for Scientific Review at the National Institutes of Health.

Xiao-Li Meng is Professor and Chair of Department of Statistics at Harvard University. He is also the incoming co-editor of Statistica Sinica. He was the recipient of the 2001 COPSS (Committee of Presidents of Statistical Societies) Award for “The outstanding statistician under the age of forty”, and was ranked (by Science Watch) among the world top 25 most cited authors for articles published and cited during 1991-2001 in mathematical sciences. His research interests, as given in the Harvard statistics Web page, are “Statistical inference under complex settings, such as partially observed data, pre-processed data, and simulated data. Quantifying statistical information and efficiency in scientific studies, particularly for scientific computation, genetic studies, and environmental problems. Statistical principles and foundational issues, especially regarding the theory of multi-party inferences, the theory of ignorance, and the interplay between Bayesian and frequentist perspectives. Effective deterministic and stochastic algorithms for Bayesian and likelihood computation. Bayesian model constructions and diagnoses. Elegant mathematical statistics.”

Ross Prentice is a Member of the Public Health Sciences Division of the Fred Hutchinson Cancer Research Center, a division which he headed from 1983-2002, and Professor of Biostatistics, University of Washington. He heads the Clinical Coordinating Center for the NIH-sponsored Women’s Health Initiative (1992 – present) which involves more than 160,000 postmenopausal women in the United States. His statistical research, which includes failure time data methods and a range of topics in epidemiologic and disease prevention research, has been continuously NIH-funded since 1976. Professor Prentice is a Fellow of the American Statistical Association, elected member, Institute of Medicine, National Academy of Sciences (1990) and is recipient of the 1980 Mortimer Spiegelman award (APHA), the 1986 COPSS award, and the 2005 AACR/ACS excellence in epidemiology and prevention award.


General – $80
Merck – $40
Bristol-Myers Squibb – $60
Wyeth – $60
Full time graduate students – $25

Registration includes: Continental Breakfast, Lunch, Break. Parking is free.

8:15AM – 9:00AM
Meeting: 9:00AM – 3:30PM

Seating is limited. Please make checks payable to Temple University (Biostatistics) and send to:

Boris Iglewicz, Director,
Biostatistics Research Center,
Department of Statistics, Temple University,
1810 N. 13th Street,
Philadelphia, PA 19122-6083

Please include your name, the name of your company, and either your email address, fax #, or address. We must receive checks by Wednesday, October 19, 2005. We cannot accept cash or credit card payments.

For additional information, contact Boris Iglewicz, Director, email: or telephone (215) 204-8637.


640 W. Germantown Pike, Plymouth Meeting PA 19462
(610) 834-8300

From Airport: Take 95 South to 476 North to the last exit #20(Germantown Pike-West). Merge with Germantown Pike and follow for 3 lights. Make a right onto Hickory Rd. at the 3rd light. The hotel is the 3rd building on the left.

New York/ New Jersey Turnpike: Take the New Jersey Turnpike to exit #6, which is PA turnpike. Go west to exit #333- Norristown. Follow signs to Plymouth Rd. Go to the 1st light and make a left. Go to the next light and make a right onto Germantown Pike. Go to the second light and make a right on Hickory Rd. The hotel is the second driveway on the left.

Washington D.C., Wilmington, and Delaware: Take I-95 North to Route 476 North. Take Route 476 to the Germantown Pike West exit #20. Go to the third light, Hickory Rd., and make a right. The hotel is the 2nd driveway on the left.

Route 476: Take 476 to the Germantown Pike West exit #20. Go to the third light, Hickory Rd., and make a right. The hotel is the 2nd driveway on the left.

From downtown Philadelphia: I-76 west Plymouth Meeting exit #331B (Route 476). Take Route 476 north to Germantown Pike exit. Go to the third light, Hickory Rd., and make a right. The hotel is the 2nd driveway on the left.