Biostatistics Student Academics Club 2005-2006

A short Guide to SAS Macros
4/17/06
Patrick Breheny

Description: The SAS macro facility is a powerful tool for extending and customizing SAS code. This workshop will explain how the SAS macro processor works and how macro code is compiled, and go over how to define and use macros and macro variables, and how to pass macro parameters. Once these macro basics are established, I will go over some additional topics that come up often in practical situations, such as conditional and iterative program control using macros, using macro functions to parse and modify macro variables, and how to interface macros with data. Throughout the workshop, I will give examples to illustrate these concepts in which the use of macros greatly simplifies a SAS coding problem. Although this is a large topic, I hope this short guide will give everyone who comes a greater understanding of how you can use macros to write cleaner, more generalizable, and more powerful SAS code. Notes available for download (SAS_Macro.pdf).

Bayesian Sequential Analysis in Phase IIB Clinical Trials
4/10/06
Lili Zhao (Department of Statistics)

Abstract: Sequential designs are widely used in clinical trials. In some cases, it is scientifically inappropriate to characterize patient outcome as a binary variable. In such settings, it often is more natural to construct stopping rules based on time-to-event variables. However, Bayesian decision-theoretic methods for designing group sequential monitoring with time-to-event endpoints have not been described in the literature. In this study, we propose a simulation-based solution for a single-armed Phase IIB study to decide whether a new treatment is non-inferior or inferior, relative to a standard. For a conjugate prior distribution, exponential failure times, and linear and threshold loss structures, we obtain the optimal Bayes rule by backward induction. Once the decision boundaries are obtained, we compute Bayesian and frequentist operating characteristics including Type I error, power, and expected run length; we further explore the optimal number of interim looks and the effect of sample size on operating characteristics. Finally, we extend this idea to piecewise exponential models.

The Concordance Correlation Coefficient and HS Application in Radon Measurements
3/20/06
Zugui Zhang

A Literature Review of the Surrogate Endpoints - Knowledge addition to your clinical trial class
3/20/06
Qian(Cicci) Shi

Estimation of the Mean Function of Panel Count Data Using Monotone Polynomial Splines
2/9/06
Minggen Lu

Abstract: We study the nonparametric pseudo-likelihood and full likelihood estimators of the mean function of a counting process based on panel count data using the monotone polynomial splines. The setting for panel count data is one in which n independent subjects, each with a counting process with common mean function, are observed at several possibly different times during a study. Generalized Rosen algorithm was used to compute the estimators. We show the proposed spline estimators are asymptotically consistent and the rate of convergence is higher than 1/3. The simulation study show the spline-based estimators have smaller variances and mean square errors than nonparametric pseudo and full likelihood estimators proposed in Wellner and Zhang (2000). A real Bladder cancer trial example is used to illustrate the method.

An Adaptive Dose-finding Design Incorporating Both Toxicity and Efficacy
11/7/05
Wei Zhang

Abstract: Novel therapies are challenging the standards of drug development. Agents with specific biologic targets and limited toxicity require novel designs to determine doses to be taken forward into larger studies. In this paper, we describe an approach that incorporates both toxicity and efficacy data into the estimation of the biologically optimal dose of an agent in a phase I trial. The approach is based on the flexible continuation-ratio model, and uses straightforward optimal dose selection criteria. Dose selection is based on all patients treated up until that time point, using a continual reassessment method approach. Dose-outcome curves considered include monotonically increasing, monotonically decreasing, and unimodal curves. Our simulation studies demonstrate that the proposed design, which we call TriCRM, has favorable operating characteristics.

Expression as Phenotype eQTL Mapping in the Mammalian Eye
11/7/05
Kwang-Youn Kim

Abstract: Recent advances in microarray technology and bioinformatics have made it possible to perform experiments that examine the expression of thousands of genes in related individuals and to use these data to identify the chromosomal locations of the genetic elements that are responsible for gene expression variation among individuals. This technique is known as expression quantitative trait locus (eQTL) mapping. In order to gain a broad perspective of mechanisms of gene regulation in the mammalian eye, and to use this perspective to find new genes that cause human eye disease, we performed eQTL mapping in the mammalian eye. Two inbred strains of laboratory rats were crossed and the resultant F_1 animals were used to generate 120 healthy F_2 male animals. The F_2 animals were euthanized at 12 weeks of age, and tissues were harvested. RNA was harvested from the whole eyes and ocular gene expression of each animal was determined using Affymetrix microarrays containing approximately 31,000 probes. DNA was obtained from the liver of each animal and used for genotyping with microsatellite markers across the whole rat genome. Two types of analyses were performed with the data. First, genetic analysis was performed to look for relationships between marker locus and expression. Second, pair-wise analysis of each gene with all other genes was performed to identify correlated variations in gene expression. Of the ~31,000 probes on the array, ~19,000 exhibited sufficient signal and variation in expression among the 120 F_2 animals for reliable analysis. Significant linkage to at least one genetic marker was detected for thousands of probes. In all, more than 4,000 marker-probe linkages were detected at an / a/ of 0.001. Both cis- and trans-acting loci were identified. The pair-wise correlated expression analysis make it possible to infer gene expression networks. In a novel application of these data, we used predicted expression networks to prioritize candidate genes within a linkage interval to aid in the identification of a gene that causes a human retinal degeneration syndrome.

Lorem ipsum dolor

Lorem ipsum dolor sit amet, consectetaur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur.