Conference Schedule

8:15 – 9:00
Registration and Continental Breakfast

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

9:10 – 9:55
Use of Biomarkers for Dose Response Modeling
Louise M. Ryan, Harvard University

9:55 – 10:40
Design Considerations in Stability Studies
Damaraju Raghavarao, Temple University

10:40 – 11:05
Break

11:05 – 11:50
The Price of Kaplan-Meier
Paul Meier, Columbia University

11:50 – 12:35
Addressing Multiple Goals In Evaluating Performance Using Bayesian Methods
Thomas A. Louis, Senior Statistical Scientist, The RAND Corporation

12:35 – 1:40
Lunch

1:40 – 2:25
An Overview of Methods for the Analysis of Clustered Ordinal Response Data
J. Richard Landis, University of Pennsylvania

2:25 – 3:10
Evaluating Product Safety Using Post-Marketing Spontaneous Reports
A. Lawrence Gould, Merck Research Laboratories

3:10 – 3:25
General Discussion
Woollcott K. Smith, Temple University

3:25
General Discussion


Abstracts

Use of Biomarkers for Dose Response Modeling

Louise M. Ryan, Department of Biostatistics, Harvard School of Public Health

This talk will discuss the role of biomarkers in helping to elucidate the pathways involved in dose-response relationships. Statistical methods for incorporating biomarker information into a dose-response analysis will be discussed and illustrated with data from a case-control study in lung cancer”.


Design Considerations in Stability Studies

Damaraju Raghavarao, Temple Univeristy

The expiration dating of a drug product is determined by finding its potency at different prespecified time intervals, fitting a linear degradation equating, and intersecting the lower confidence band of the fitted equation with the specification limit. This has to be done with all factors affecting the storage conditions. The number of assays will be reduced by using Matrixing and Bracketing Designs. The speaker will overview his collaborative work in this area.


The Price of Kaplan-Meier

Paul Meier, Columbia University

In 1983 Rupert Miller wrote, in Biometrics, a paper called “What Price Kaplan-Meier?”, saying that the Kaplan-Meier estimator was quite inefficient and suggesting that, if possible, the analysts who might use it,should use instead some parametric form. (He mentioned the Exponential Distribution but he also indicated other distributions as well.) I thought that his analysis was mistaken, but I thought the references to “What Price” would disappear. Unfortunately, I find that it has been quoted at least 37 times since then, and thus it is time that I answer it, and I will do so.


Addressing Multiple Goals In Evaluating Performance Using Bayesian Methods

Thomas A. Louis, Senior Statistical Scientist, The RAND Corporation

Environmental assessments, health services evaluations and school effectiveness stud-ies all depend on comparing and evaluating units (geographic regions, health care pro-viders, schools) based on unit-specific estimates. For example, Standardized Mortality Ratios are used to compare outcomes to regional or national norms. Invalid estimation or inappropriate interpretation can have serious local and national consequences. Ap-propriate comparison requires considering both estimated values and their statistical un-certainty, especially when uncertainty varies over units. In this situation, hypothesis test-ing to identify poor performers unfairly penalizes units with relatively stable estimates; use of direct estimates unfairly penalizes units with relatively unstable estimates. Stabi-lizing estimates to balance variability while retaining sufficient unit-specific focus is very effective in striking a compromise.

Explicit structuring via hierarchical (Bayesian) models and loss functions have proven very effective in accomplishing these goals. Stabilization results from “borrowing infor-mation” across units. The posterior mean is the “natural” estimate and is commonly used to estimate unit-specific values. However, the histogram of the posterior means is under-dispersed and never a valid estimate of the parameter histogram (needed to com-pare distributions and to estimate the number of values above a threshold). And, ranking posterior means can produce sub-optimal ranks. Therefore, for these goals an alterna-tive to the posterior mean is needed.

We outline the models, goals and issues and propose a single set of estimated values that has good performance for estimating unit-specific values, their histogram and their ranks. We show how it operates using environmental and health services examples and identify issues requiring additional research.


An Overview of Methods for the Analysis of Clustered Ordinal Response Data

J. Richard Landis, Thomas R. Ten Have and Robert J. Gallop, University of Pennsylvania

This talk will review statistical methods for the analysis of ordinal categorical response data for study designs involving subject-level clustering. Methods to be illustrated include cluster-specific (CS) models within the generalized Mantel-Haenszel framework and mixed effects models, as well as population averaged (PA) models within a generalized estimating equation framework.

Attention will be given to common problems encountered in practice, such as the impact of non-informative clusters, samples sizes within clusters relative to the number of clusters, confounding by cluster, and informative missing data. A survey of recent methodological research and challenges for the future will be highlighted.


Evaluating Product Safety Using Post-Marketing Spontaneous Reports

A. Lawrence Gould, Merck Research Laboratories

Spontaneous reports of adverse events observed in clinical practice may provide evidence about risks of treatment that are too small to be observed in clinical trials. Until recently, rechniques have not been readily available for providing a practical way to extract a manageable number of potential signals from large spontaneous reporting databases, although the basic principles have long been well understood. Recently proposed Bayesian methods may contribute usefully to pharmacovigilance currently carried out by skilled clinicians and medical epidemiologists by providing a way to identify possible signals sooner. We compare two recently proposed Bayesian methods, one using a conventional Bayes approach, one using an Empirical Bayes approach. We refine the former by providing ‘exact’ formulas in place of asymptotic approximations originally proposed. We refine the latter by providing ways to incorporate uncertainty explicitly into the signal criterion. Refinements to both approaches do not add materially to the computational burden of either. Examples of the calculations are provided for illustration.


Speakers

Dr. Louise Ryan serves as Professor of Biostatistics at the Harvard School of Public Health and the Dana-Farber Cancer Institute. Professor Ryan works on statistical methods for application in cancer and environmental health research. She has served in a variety of professional capacities, including President of the Eastern North American Region of the Biometric Society, Co-Editor of Biometrics, and service on several National Academy Committees. Dr. Ryan has published numerous journal articles and is a Fellow of the American Statistical Association.


Damaraju Raghavaro is a Laura H.Carnell Professor of Statistics at Temple University. He made extensive contributions to all aspects of Experimental Designs and published a reference book in the construction and combinatorics of designs. He is a Fellow of the Institute of Mathematical Statistics, American Statistical Association, and an elected Member of the International Statistical Institute. He received the Paul W. Eberman Research Award from Temple University in 1996 for his research work.


Paul Meier, Ph.D, Levene Professor of Statistics and Professor of Biostatistics at Columbia University (retired). Ph.D. in Mathematics, Princeton University (1951), advisor J.W. Tukey. Spent 5 years in the Department of Biostatistics, Johns Hopkins University, then 35 years at Department of Statistics, University of Chicago. In 1992 he went to Columbia University, serving as Department Chair, retiring in 2000. Professor Meier has been selected as member of the National Institute of Medicine and to the American Academy of Arts and Sciences. He is a Fellow of the American Statistical Institute, Fellow of the Institute of Mathematical Statistics, and member of the International Statistics Institute. Most of his work, both theoretical and applied, was in the area of Clinical Trials. He also studied the Legal system.


Thomas A. Louis, PHD, Senior Statistical Scientist at Rand; Adjunct Professor at Johns Hopkins SPH and George Washington University SPH. Dr. Louis received his Ph.D. degree in Mathematical Statistics from Columbia University and from 1987-1999 headed Biostatistics in the University of Minnesota SPH. Research interests include en-vironmental, health and public policy and development of related statistical procedures. Methodologic work concentrates on Bayesian modeling and the analysis of observational studies. Current applications include assessing the health effects of airborne particulate matter, clinical quality improvement, cardio-pulmonary consequences of AIDS therapies, evaluation of teacher effectiveness, small area estimation and statistical methods to as-sess environmental justice. Dr. Louis has published more than 150 articles and books and delivered more than 200 invited presentations. He is coordinating editor of The Jour-nal of the American Statistical Association, a member of the National Academy of Sciences Committee on National Statistics, on the Board of the Institute of Medicine’s Medi-cal Follow-up Agency and on the executive committee of the National Institute of Statisti-cal Sciences. He is an elected member of the International Statistical Institute, a fellow of the American Statistical Association and of the American Association for the Advance-ment of Science.


J. Richard Landis is Professor of Biostatistics, and serves as Director of the Division of Biostatistics at the University of Pennsylvania, School of Medicine. He also holds a secondary appointment as Professor of Statistics in the Wharton School. Dr. Landis earned the M.S.(1973) and Ph.D.(1975) in Biostatistics from the University of North Carolina at Chapel Hill. He has been actively involved in collaborative biomedical research and the development and evaluation of methods for the analysis of categorical data. Dr. Landis’ publications are in the areas of statistical methods for repeated measurement and longitudinal categorical data, epidemiological studies, complex sample surveys and applications to cardiovascular, pulmonary and urological research. He currently is PI of two NIDDK-funded, multi-institutional Data Coordinating Centers – the Chronic Prostatitis Collaborative Research Network (CPCRN) and the Interstitial Cystitis Clinical Trials Group


Dr. A. Lawrence Gould is Senior Director, Scientific Staff, Biostatistics and Research Data Systems, Merck Research Laboratories, and also Adjunct Professor of Statistics at Temple University. He is a Fellow of the American Statistical Association, has served on a number of grant review panels, and has served in a variety of positions with the Biopharmaceutical Section of the ASA and the Biometric Society, ENAR. He currently is Editor of the Journal of Biopharmaceutical Statistics. Larry’s research interests tend to be driven by problems arising in drug development, and include blinded sample size re-estimation, Bayesian methods, meta-analysis, bioequivalence, analysis of safety data, data mining, clinical trial simulation, and management science.


Registration

This event has passed.


Directions

DOUBLE TREE GUEST SUITES, PLYMOUTH MEETING
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.