Communications in Statistics: Theory and Methods

Ahn, Hongshik; Zhu, Wei; Yang, Joonsook; Kodell, Ralph L.

doi: 10.1080/03610929808832158pmid: N/A

In a typical carcinogenicity study, animals, usually rats or mice. are divided into a control and two to three dose groups of 50 or more by randomization. A chemical is administered at a constant daily dose rate for a major portion of the lifetime of the test animals, for example, two years. In general, such an experiment is expensive and time consuming In this paper, we propose an efficient design with reduced sample size and/or shortened study duration. An equal number of animals per dose group is considered in this study. A power study of the age-adjusted trend test, for the turnor incidence rate for single-sacrifice experiments proposed by Kodell et al. (Drug Information Journal, 1997) is conducted. A Monte Carlo simulation study is performed to compare the performance of the trend test for the standard design and various reduced designs. Based on the Kodell et al. test, the 21-month study duration with sample size 50 per group is recommended as the best, reduced design over the traditional 2-year study design with the same sample size.

Balkin, Sandy D.; Lin, Dennis K.J.

doi: 10.1080/03610929808832159pmid: N/A

The search of construction methods for a “good” supersaturated design has received a great deal of attention in recent years. Many single-valued criteria have been proposed for comparison purposes. In this paper, we propose/review three criteria based on the information matrix of a projective design and provide two graphical methods for comparing different types of supersaturated designs: the line plot and the boxplot methods. Four different construction methods are compared using the proposed graphical techniques. These construction methods were proposed by Booth and Cox (1962), Lin (1993), Wu (1993) and Lin (1995). A thorough discussion on the proposed graphical techniques is given.

Chan, Ivan S.F.; Bohidar, Norman R.

doi: 10.1080/03610929808832160pmid: N/A

In vaccine efficacy studies the goal is to show that the vaccine reduces the incidence of the disease compared to placebo. In this report we describe two procedures for calculating sample size and powei based on exact distributions In small studies where the disease incidence and the anticipated vaccine efficacy are both high, an unconditional exact procedure is desirable because it guarantees the level of the test and loses little sensitivity. In large studies where the disease incidence is rare, a Poisson approximation to the number of events is reasonable and an exact test is simple to construct conditional on the total number of events. We compare the power and type I error rate of these two exact methods to the method based on the normal approximation for varying disease incidence and sample size.

Chang, Myron N.; Hwang, Irving K.; Shin, Weichung J.

doi: 10.1080/03610929808832161pmid: N/A

We propose flexible group sequential designs using type I and type II error probability spending functions. The proposed designs preserve the overall significance level and power and allow the repeated testing to be perloimed at a flexible schedule. Computational methods are described. An example on a mega clinical trial is provided.

Chen, Ming-Hui

doi: 10.1080/03610929808832162pmid: N/A

In this article, we develop statistical models for analysis of correlated mixed categorical (binary and ordinal) response data arising in medical and epidemi-ologic studies. There is evidence in the literature to suggest that models including correlation structure can lead to substantial improvement in precision of estimation or are more appropriate (accurate). We use a very rich class of scale mixture of multivariate normal (SMMVN) iink functions to accommodate heavy tailed distributions. In order to incorporate available historical information, we propose a unified prior elicitation scheme based on SMMVN-link models. Further, simulation-based techniques are developed to assess model adequacy. Finally, a real data example from prostate cancer studies is used to illustrate the proposed methodologies.

Timothy Chen, T.; Simon, Richard M.; Kom, Edward L.; Anderson, Stewart J.; Lindblad, Anne S.; Wieand, Harry Sam; Douglass, Harold O.; Fisher, Bernard; Michael Hamilton, J.; Friedman, Michael A.

Cheng, K.F.; Hsueh, H.M.; Chien, T.H.

doi: 10.1080/03610929808832164pmid: N/A

The most popular goodness of fit test for a multinomial distribution is the chi-square test. But this test is generally biased if observations are subject to misclassification, In this paper we shall discuss how to define a new test procedure when we have double sample data obtained from the true and fallible devices. An adjusted chi-square test based on the imputation method and the likelihood ratio test are considered, Asymptotically, these two procedures are equivalent. However, an example and simulation results show that the former procedure is not only computationally simpler but also more powerful under finite sample situations.

Lau, Tai-shing

doi: 10.1080/03610929808832165pmid: N/A

We propose to use mixture of geometries to model the duration Ot a spell in discrete time. The discrete hazard rate can be determined by the moments of the mixing distribution through the qd(quotient difference) algorithm, By introducing a sequence of parameters called canonical moments, we can ensure the resulting estimate of the mixing distribution is supported on [0,1]. Using the bootstrap method, we can construct, instantaneous confidence intervals for the hazard rates. For the continuous time, a mixture of exponentials is proposed. By a simple transformation, the problem can be reduced to the case for mixture of geometries. The instantaneous confidence intervals for the hazard rates can be constructed.

Li, Lung-An

doi: 10.1080/03610929808832166pmid: N/A

Stochastic curtailing method is used in monitoring slowly accruing clinical data from a fixed sample size design for any possible early stopping of the trial. This paper refines the results of Lan and Wittes (1988) by providing reasonable estimator of the conditional power at any given information time for better decision making. Some comparisons, examples are provided, and consequences are-discussed.

Lin, Stan C.; Whipple, David M.; Ho, Charles S.

doi: 10.1080/03610929808832167pmid: N/A

In this article, the Food and Drug Administration's (FDA) medical device substantial equivalence provision will be briefly introduced and some statistical methods useful for evaluating device equivalence are discussed. A sample size formula is derived for limits of agreement, which may be used to assess statistical equivalence in certain medical device situations. Simulation findings on the formula are presented, which can be used to guide sample size selections in common practical situations. Examples of the sample size procedure are given.