BIOMETRICS SECTION NEWS, MAY 2009
Edited by Page Moore, Biometrics Section Publications Officer

Continuing Education Courses at JSM 2009
Jerry Heatley , Continuing Education Chair.
The Biometrics Section is proud to co-sponsor four short courses during the 2009 JSM in Washington DC :

Missing Data in Longitudinal Studies: Strategies for Bayesian Modeling and Sensitivity Analysis (Saturday, August 1, 8:30 a.m. – 5:00 p.m.)

Instructors: Michael Daniels, PhD ( University of Florida ) and Joseph Hogan, PhD ( Brown University )

This course provides a survey of modern model-based approaches to handling dropout in longitudinal studies, and illustrates the use of newly-developed methods for sensitivity analysis and incorporation of prior information. Emphasis is on Bayesian approaches but the models and methods discussed can be implemented in non-Bayesian settings. Course will be divided into three parts: Part 1 includes a brief review of models for longitudinal data and the basics of Bayesian inference; Part 2 focuses on formal classification of dropout and missing data mechanisms, describes classes of models that can be used to adjust for biases caused by dropout, and the logistics of model fitting; Part 3 deals with specification and fitting of models to handle non-ignorable (informative) dropout, with emphasis on the role of sensitivity analysis and informative prior distributions for encoding key assumptions.

Integrated into the course are three case studies that illustrate many of the concepts introduced during the course. We will build on each case study to illustrate progressively more complex analyses (e.g., progressing from analysis under MAR, to analysis under MNAR, to use of informative priors and sensitivity analyses). The case studies will include real-time demonstration of model fitting using WinBUGS software and the R-to-WinBUGS interface.

Longitudinal Data Analysis: Semiparametric and Nonparametric Approaches (Saturday, August 1, 8:30 a.m. – 5:00 p.m.)

Instructor: Annie Qu, PhD ( University of Illinois at Urbana-Campaign) and Peter Song, PhD ( University of Michigan )

This short course will introduce semiparametric and nonparametric approaches that do not require specification of the likelihood function. In particular, the application of the quadratic inference function (QIF) will be illustrated through development of statistical theory and methods in addition to real-world data examples and the use of the newly developed SAS MACRO QIF by Song and Jiang. The QIF proposed by Qu et al. (2000 Biometrika ) can easily take into account correlation within subjects, and deal directly with both continuous and discrete longitudinal data under the framework of generalized linear models. The short course will demonstrate that (1) the QIF yields a more efficient estimator than the currently popular GEE approach; (2) the QIF is analogous to twice the negative log-likelihood, so it naturally provides for tests about model parameters and goodness of fit; (3) the QIF provides for diagnostic tests about the mechanisms underlying any missing data.

Methodology for Competing Outcomes: The Analysis of Multiple Mutually Exclusive Endpoints in a Clinical Trial (Monday, August 3, 8:00 a.m. – 12:00 p.m.)

Instructor: David C. Naftel, PhD ( University of Alabama at Birmingham )

Many clinical trials today have a primary endpoint that is actually a composite of several separate endpoints that can be constructed so that they are mutually exclusive. Medical devices that provide mechanically assisted circulatory support often are evaluated on the basis of multiple endpoints including death, removal of the device due to cardiac transplantation and removal of the device due to recovery of the heart. A fourth “event” is when the patient is alive with the device in place at a specified point in time. Estimating the distribution of time until the occurrence of the composite endpoint is important but is rarely sufficient. Clinicians want to understand the simultaneous probabilities of these events.

This course will define the problem within a clinical and FDA perspective. Trial design and sample size will be presented. Non-parametric methodological approaches will be presented. Parametric methods that stress the shape of the underlying hazard function will be emphasized. Finally the interactive roles of industry, clinicians, FDA and the statistician will be defined, interpreted and presented as a symbiotic structure .

Absolute Risk Prediction (Tuesday, August 4, 8:00 a.m. – 12:00 p.m.)

Instructors: Mitchell H Gail, M.D., Ph.D. (NIH) and Ruth Pfeiffer, Ph.D. (NIH)

The absolute (or “crude”) risk is the probability that an individual who is free of a given disease at an initial age, a , will develop that disease in the subsequent interval ( a, t ]. Absolute risk is reduced by mortality from competing risks. Models of absolute risk that depend on covariates have been used to design intervention studies, to counsel patients regarding their risks of disease and to inform clinical decisions, such as whether to take tamoxifen to prevent breast cancer. This course will define absolute risk and discuss methodological issues relevant to the development and evaluation of risk prediction models. Various study designs and data for model building will be presented, including cohort, nested case-control, case-control data combined with registry data, and kin-cohort data. Issues relating to the evaluation of risk prediction models and the strengths and limitations of risk prediction models for various applications will be discussed. Standard criteria for model assessment will be presented, as well as loss function-based criteria applied to the use of risk models to screen a population and the use of risk models to decide whether to take a preventive intervention that has both beneficial and adverse effects.

Biometrics Section Mixer & Business Meeting at JSM 2009
Please join us at the Biometrics section Mixer & Business Meeting at the 2009 JSM in Washington , DC ! It will be held on Monday, August 3 rd from 5:30-7:00 p.m. Details to follow in the upcoming June issue of the Biometrics newsletter.

ENAR 2010
It is time to think about invited sessions for ENAR 2010, which will be held March 21–24 in New Orleans , LA. Anyone who is interested in organizing an invited session or who has ideas for one, please contact our 2010 Program Chair, Liang Li, at lil2@ccf.org .

A typical session consists of three 30-minute talks followed by a 30-minute  
discussion or four 30-minute talks.  June 11 is the deadline for proposals.  It is best if you have a well-defined topic and have commitments from participants by June 11. The more detailed the proposal, the better the chances that it will be selected as this is a highly competitive process.

JSM 2010
It's also time to start thinking about invited sessions for next year's Joint Statistical Meetings, which will be held August 1–5 in Vancouver , BC , Canada . Anyone who is interested in organizing an invited session or who has ideas for one, please contact our 2010 Program Chair, Hormuzd Katki , at katkih@mail.nih.gov .

A typical invited session consists of three 30-minute talks followed by a 10-minute invited discussion and 10 minutes of floor discussion. However, other formats are possible. The 2009 program is a good source for examples.

Remember, the most mature ideas will have an advantage in competing for the limited number of slots, so it's best to have your ideas in final form by the middle of June. The Biometrics Section will have at least four invited sessions, but if we generate enough good ideas we will be able to compete for additional slots as well.

Please also submit ideas for short courses our 2009-2010 Continuing Education Chair Jerry Heatley at jerry.heatley@thoratec.com .

Vote for Officers of the Biometrics Section
There are two open positions for officers of the Biometrics Section: Chair-Elect and Council on Sections Representative. You can cast your ballot online at the ASA website. The candidates for Chair-elect are J. Jack Lee, MD Anderson Cancer Center and Joan F. Hilton, University of California , San Francisco . The two candidates for council on sections representative are Sowmya R. Rao, Massachusetts General Hospital and Mousumi Banerjee, University of Michigan . Section members can vote on these positions as well as other positions in ASA.