Spring 2007 Meeting
Held on March 22, 2007 at the Deerfield Embassy Suites.
The Program consisted of three presentations:
-
How to Obtain Accurate
Sample Size and Power Using nQuery Advisor
Brian Sullivan, MS, Statistical Solutions
-
Methodological Challenges In Analyzing Patient-Reported Outcomes
Elizabeth A. Hahn, MA, Department of Preventive Medicine and Institute for Healthcare Studies, Northwestern University
-
Practical
Considerations in Applied Nonlinear Regression Modeling: Design,
Estimation and Testing
Timothy E. O’Brien, Ph.D., Department of Mathematics and Statistics, Loyola University of Chicago
How to Obtain Accurate Sample Size and Power Using nQuery Advisor by Brian Sullivan, MS, Statistical Solutions
Biographical Background
Brian Sullivan graduated from the University of Limerick with a Masters
degree (Research) in Statistics. The main focus of his research work was in
applying a Graphical Modeling technique to large sets of data whilst
accounting for missing values. He has been working at Statistical Solutions
as a Customer Support Technician since November 2006. He was previously
graduated with a Bachelor’s degree in Applied Science and Computing from the
Dublin Institute of Technology.
Abstract
This presentation will focus on
the application side of sample size calculations. Several examples will be
presented for different statistical methods. These will include approaches
for clinical trials, χ2 tests, ANOVA and regression. The process
involved in calculating appropriate sample size will be outlined in each
case. The presentation will also include a demonstration of all calculations
using the sample size software nQuery Advisor.
Methodological Challenges In Analyzing Patient-Reported Outcomes by Elizabeth A. Hahn, MA, Department of Preventive Medicine and Institute for Healthcare Studies, Northwestern University
Biographical Background
She is a medical
sociologist and biostatistician with expertise in the design,
implementation, coordination and statistical analysis of international
clinical trials, survey research studies and other health services research
projects. She has expertise in the application and interpretation of
probabilistic measurement models (item response theory) as well as
statistical models based on classical test theory. She participates in
workshops and symposia to discuss research design, measurement and
statistical analysis issues related to self-report data, and provides
consultation to international researchers in the design and analysis of
clinical trials. She co-teaches a course on Survey Design and Methodology in
the Feinberg School of Medicine, Masters of Public Health program.
Her research primarily
involves patient-reported outcomes (PROs) in patients with cancer and other
chronic illnesses, with a focus on underserved populations and health
disparities. She has been principal investigator on research grants funded
by AHRQ/NCI, ACS, Coleman Foundation and NHLBI, and co-investigator on
research grants funded by ACS, American Heart Association, CDC, NCI, NIAMS,
NHLBI and several foundations. Her research includes the development of a
bilingual, multimedia Talking Touchscreen that allows patients with varying
literacy levels and computer skills to self-administer PRO questionnaires.
Abstract
The analysis of patient-reported outcomes (PRO) data in an international
clinical trial presents methodological, statistical and interpretive
challenges. Specific strategies are needed regarding the psychometric
measurement properties of self-report instruments, cross-cultural
measurement equivalence, definitions of clinical significance, missing data,
longitudinal modeling, and descriptions of clinically interpretable results.
Development and implementation of these strategies is illustrated using the
16-country International Randomized IFN vs. STI571 (IRIS) Study of 1106
newly diagnosed patients with chronic phase chronic myeloid leukemia. The
primary endpoint was the duration of progression-free survival; PROs were
secondary endpoints. Crossover to the other treatment was permitted because
of intolerance or lack of efficacy. The Functional Assessment of Cancer
Therapy-Biologic Response Modifiers (FACT-BRM) was completed as a measure of
health-related quality of life at baseline, months 1–6, 9, 12, 18, and 24 in
the patient’s preferred language. The methodological issues and specific
strategies developed to address them are summarized. An item response theory
(IRT) measurement model was used to evaluate psychometrics, including
cross-cultural comparability (three languages), and to aid in interpretation
of treatment differences. A mixed effects model was chosen for the
longitudinal analyses, with a pattern-mixture technique to adjust for
nonignorable missing data. Crossover effects were added as a time-dependent
covariate. To better understand the meaning of the PRO scores, a clinically
significant treatment effect was prespecified, and a modified forest plot
was used to summarize IRT responses. 1049 patients (95%) participated in the
assessments. The patterns of dropout and change were quite different for the
treatment arms. This study presented major methodological challenges to PRO
data analysis, all of which were addressed using state-of the-science
modeling techniques. The analysis plan and results may be useful for
statisticians, researchers and clinicians who analyze and interpret PROs.
Practical Considerations in Applied Nonlinear Regression Modeling: Design, Estimation and Testing by Timothy E. O’Brien, Ph.D., Department of Mathematics and Statistics, Loyola University of Chicago
Biographical Background
Dr. Timothy E. O’Brien is a tenured
associate professor with the graduate faculty in the Department of
Mathematics and Statistics, Loyola University of Chicago. Dr. O’Brien
received his Ph.D. in Statistics from North Carolina State University in
1993. His dissertation topic, “New Design Strategies for Parameter
Estimation and Model Discrimination in Nonlinear Regression Models” focuses
on optimal experimental design, generalized linear and nonlinear modeling,
and computer intensive methods, with applications to drug synergy research.
Dr. O’Brien also received an M.A. in Statistics from the University of
Rochester (1987), an M.A. in Mathematics from Syracuse University (1985),
and a B.A. in Mathematics and Economics from Pace University (1978). He is
a member of ASA, ENAR, IASC, IASE, and ISI.
Dr. O’Brien has made
several contributions to the theory and methods of optimal experimental
design, particularly regarding nonlinear modeling. Some of his publications
appear (or will appear) in Biometrika, Statistica Sinica, Journal of
Statistical Planning and Inference, The American Statistician, Journal of
Agricultural, Biological, and Environmental Statistics, the Journal
of Chemical Ecology, Computational Statistics and Data Analysis,
and the Journal of Data Science. Dr. O’Brien also published three
book chapters on optimal design, robust design, and lack of fit, for
nonlinear regression models, as well as several refereed conference
proceedings (e.g., Proceedings of the 15th Conference on
Applied Statistics in Agriculture, Proceedings of Agro-Industrie et Methodes
Statistiques) and collaborative papers in refereed biomedical journals
(e.g., Development, Annals of Neurology, Cell and Tissue Research),
which illustrate the immediate application of his theoretical work. Dr.
O’Brien has served numerous times as a referee for top tier statistical
journals, and he is frequently invited to both domestic and international
conferences and universities to speak on his theoretical developments and
their applications to pharmacology and pharmacokinetics. Dr. O’Brien won a
SUGI Best Contributed Paper Award for demonstrating how some of his
metholdogical work on optimal designs for nonlinear regression models can be
implemented with SAS®
Dr. O’Brien's previous
industrial and academic work experience contributed greatly to both the
direction and applications targeted for his current research activities.
For example, Dr. O’Brien spent two years as a biostatistical consultant at
Janssen Pharmaceutics NV, two years as an internal statistical consultant
and biostatistician at Novartis Pharma AG, and three years as an assistant
statistician at Glaxo. In addition, Dr. O’Brien also provided statistical
consulting services to SmithKline, Bristol Myers Squibb, Chiron, and Amgen.
Dr. O’Brien’s previous domestic academic experience includes Assistant
Professor positions at Loyola University of Chicago, the University of
Georgia, and Washington State University. Internationally, Dr. O’Brien has
been a Visiting Professor at both Limburgs Universitair Centrum (Belgium)
and Katholieke Universiteit Leuven (Belgium), a Visiting Lecturer at the
University of Natal at Pietermaritzburg (South Africa), and he was awarded
two postdoctoral fellowships, one at the Universität Augsburg (Germany) and
the other at INRA, Laboratorie de Biometrie (France).
Dr. O’Brien has a strong interest in and commitment to statistical and mathematical education. He has developed and taught a wide range of theoretical and applied statistics, statistical computing, statistical programming, statistical consulting, and mathematics courses, at both the graduate and undergraduate levels, at both domestic and universities abroad. In addition, Dr. O’Brien has taken the time to supervise more than a dozen directed reading courses with graduate students in nonlinear mixed modeling, generalized linear models, nonlinear regression, differential geometry, optimal design, multivariate statistics, survival analysis, advanced statistical inference, and drug synergy, some of which led to the students’ dissertation research. Dr. O’Brien has been invited to conferences on teaching statistics and also to universities to share his ideas on successful teaching, and he has recently written an invited book chapter on “Innovative Methods in Undergraduate Courses Following Calculus” which is to appear in the MAA Notes series. Perhaps the clearest demonstration of Dr. O’Brien’s commitment to education was his two-year tour with the Peace Corps in French West Africa where he taught several mathematics courses preparing lycée students for the French Boccaleureate exam and entrance into university
Abstract
Researchers often find that
nonlinear regression models are more applicable for modeling various
biological, physical and chemical processes than are linear ones since they
tend to fit the data well and since these models (and model parameters) are
more scientifically meaningful. These researchers are thus often in a
position of requiring optimal or near-optimal designs for a given nonlinear
model. A common shortcoming of most optimal designs for nonlinear models
used in practical settings, however, is that these designs typically focus
only on (first-order) parameter variance or predicted variance, and thus
ignore the inherent nonlinear of the assumed model function. Another
shortcoming of optimal designs is that they often have only p support
points, where p is the number of model parameters. Measures of marginal
curvature, first introduced in Clarke (1987) and further developed in Haines
et al (2004), provide a useful means of assessing this nonlinearity.
This talk examines the reliability of
Clarke’s marginal curvature measures (vis-à-vis other curvature and
nonlinearity measures) in practical settings, and introduces a design
criterion that combines variance minimization with nonlinearity
minimization. These techniques, coded in the SAS® software packages (NLIN
and NLMIXED procedures), are illustrated in the context of a dissolution
model given in Weiss and Lansky (1998) as well as other examples.
Last updated: 03/31/09 by: Clint Lovell
