Statistics Calendar of Events 2007-2008
|May 6, 2008||1:30 pm SC 170||Ron Brookmeyer, PhD, Johns Hopkins School of Public Health||Modeling and Bio-security: A Case Study in Anthrax
About the talk: The threat of bioterrorism is of increasing public health concern. Yet we have very limited modern experience with diseases such as anthrax and smallpox which are considered some of the major biological weapons threats. In the face of this uncertainty, the problem is how to devise rational public health policy in the event of a bio-security breach. About the talk:In this talk, I discuss the case of anthrax and how statistical models, novel data sources, and efficient use of information can be brought together to address some critical public health questions. The work was motivated by the 2001 anthrax outbreak in the United States that occurred because letters contaminated with anthrax were mailed through the postal system. A huge public health intervention was initiated that involved the mass distribution of antibiotics to thousands of persons in the hopes of preventing disease. Two questions are addressed in this talk: Was the public health intervention successful? Are there lessons to learn about antibiotic use in the event of another bio-security breach?
We develop statistical models to address these questions. For example, one model is a mechanistic model that accounts for the pathogenesis of infection. We show that these mechanistic considerations lead to a relatively simple competing risks model that provides insight into how long persons infected with anthrax should remain on antibiotics.In order to calibrate the models, we use data and information from multiple sources including the 2001 anthrax outbreak in the United States, an anthrax outbreak in Russia, and studies from primates. The talk illustrates how statistical reasoning, models, and novel data sources can help develop effective public health response policies in the event of future intentional or naturally occurring outbreaks.
|May 6, 2008||1:30 pm SC 188||Carlos Castillo-Chavez, PhD,
Mathematical Biologist,U of Arizona
|Mathematics and Disease: The case of Tuberculosis About the presenter: In this talk, the role of mathematics in the study of disease dynamics over the past century will be highlighted. The role of mathematics in epidemiology through the introduction of mathematical models used to enhance our understanding of the evolution and the transmission dynamics of tuberculosis will be illustrated with emphasis on finding ways of reducing the disease’s burden. The talk will be accessible to undergraduates and to biology majors interested in the use of models in the life and social sciences.|
|April 29, 2007||7:00 pm||Carlos Castillo-Chavez, PhD,
Mathematical BiologistU of Arizona
|The Dynamics of Disease in the 21st Century: From Vaccines to Homeland Security
About the talk: The precision of the military fatality count in the current war is notable. In contrast, we know little about the number of Iraqi civilians who have died because of the invasion. Estimates vary from tens to hundreds of thousands. This talk will discuss scientific issues related to estimating and communicating about the number of Iraqi civilians who have died due to the U.S. invasion. It will consider the definition of “cause”, the influence of prior beliefs on estimation, the value of expert opinion, and contending with significant uncertainty. Data from two mortality surveys published in the Lancet (2004, 2006) will be relied upon..
In this talk, the role of mathematics in the study of disease dynamics over the past century will be highlighted. The role of mathematics in epidemiology through the introduction of mathematical models used to enhance our understanding of the evolution and the transmission dynamics of tuberculosis will be illustrated with emphasis on finding ways of reducing the disease’s burden. The talk will be accessible to undergraduates and to biology majors interested in the use of models in the life and social sciences.
|April 1, 2008||1:30 pm||Katie St. Clair, Carleton College||Two Perspectives on Finite Population Sampling
Abstract: Sample surveys give us a current picture of a population of interest. They can be used to provide a snapshot of current public opinion, the state of the economy, or to evaluate the condition of natural resources. The objective of a sample survey is to make conclusions about an entire finite population with information obtained from just a sample of population members. In my talk I will discuss two competing methods for making these conclusions. The first and most common approach to finite population sampling constructs estimates and confidence intervals based on how the sample was selected (i.e. the sampling design). A second approach, called the Bayesian approach, constructs estimates and confidence intervals based on a statistical model that relates the unobserved population members to the sampled members. I will describe a specific Bayesian model which uses a “Polya Urn” to relate the unobserved to the observed. I will discuss how both of these approaches can be used to estimate the density of wintering waterfowl in central Florida.
Bio: Katie St. Clair grew up in Hastings, MN and received my Ph.D. in statistics from University of Minnesota in 2004. From 2004-2207 she was an assistant professor at Colby College in Maine. Katie moved back to Minnesota in August 2007 to join the mathdepartment at Carleton College
|March 7, 2008||5:30 pm SC 182||Pam Arroway, NCState||Graduate School at North Carolina State|
|December 4, 2007||1:30 pm SC 182||Andie Bykerk Christopherson, FSA, MAAA Reden & Anders||Being a Health Actuary: Cool Models, Cool Career
Abstract: Andie will discuss the actuarial modeling challenge of understanding and analyzing the impact of calendar patterns and content on expected health care claims costs, as well as the trend patterns these values produce. She will also discuss skills most important to being a successful actuary and a little about the preliminary exam and education requirements.Bio: Andie Bykerk Christopherson, FSA, MAAA is a consulting actuary at Reden & Anders, focused on health care trend analytics. An ‘02 Ole grad with a mathematics major and statistics concentration, she finished her actuarial exams this fall, becoming a Fellow in the Society of Actuaries. She resides in the Twin Cities, and has been very involved with the recruitment of new students into the actuarial field, focused especially on making candidates with broad backgrounds aware of the possibilities the field has to offer.
|November 6, 2007||1:30 pm SC 182||Karl W. Broman, PhD.,
Biostatistics & Medical Informatics,University of Wisconsin-Madison
|Mapping multiple QTL in experimental crosses
Abstract: Many diseases are inherently quantitative (e.g. hypertension). Others are generally viewed as binary (e.g. diabetes), but are closely associated with intermediate quantitative phenotypes (e.g. glucose tolerance). Quantitative traits are generally influenced by multiple genetic loci (called quantitative trait loci, QTLs) as well as the environment. The number, location, and effects of the genetic loci that contribute to a disease-related phenotype help us to understand the biochemical basis of the disease and can lead to the development of improved treatments for human disease. We consider the problem of identifying QTL in an experimental cross (such as with mice). In the traditional approach to QTL mapping, one considers each genomic position, one at a time, and tests for association between genotype and the quantitative phenotype. Great attention has been placed on the adjustment for multiple hypothesis tests. The simultaneous consideration of multiple QTL can provide greater power, can better separate linked QTL, and allows the investigation of interactions between loci. The problem is best viewed as one of model selection. We describe the key issues and propose a penalized likelihood approach for model selection. Our approach provides an automated procedure that can enable biologists with limited statistical training to obtaina more complete understanding of the set of genetic loci contributing to variation in a quantitative trait.
|September 25, 2007||1:30 pm SC 170||Andrea MacKay, MSPH,
Office of Analysis and Epidemiology, National Center for Health Statistics,Centers for Disease Control and Prevention
|Adolescent Health in the United States, 2007 Adolescent Health in the United States, 2007 presents the most recent data available on adolescent health in the United States, using data from nationally representative surveys and vital statistics. The differences in health status between younger and older adolescents are documented. The health of the adolescent population also varies by gender, race and ethnicity, and socioeconomic status. Many of the measures of health status are shown by single year of age or by two- or three-year age intervals to highlight the changes that occur in health status as adolescents move through this important developmental period. Understanding patterns of health among adolescents requires attention to differences in the population and recognition of the economic and racial disparities that exist.|