In this paper, We propose a ridge combined principal Component estimator for the regression coefficients. And the admissibility, restricted admissibility and consiste-pcy of the estimator are discussed. We show the regions preferable to OLS, OER, under mean square error and Pitman's clossness respectively, and also that it can improve the resistance and sensitivity to data.
After glancing at the book title, we initially thought the focus would be on mathematical formulas and statistical methodologies that are often used in clinical trials. To our pleasant surprise, th...
Statistical methods to address disease-modifying (DM) clinical data for agents like BACE for Alzheimer's disease require either randomized start or withdrawal designs. For ethical reasons randomized start design for AD are gaining more interest in the research communities and by regulatory agencies. Statistical methods to address disease modifying clinical trials are not fully developed yet. Statistical methods proposed for DM include work by C, Xiong and colleagues. These researchers proposed a trivariate distribution whose means and covariance matrix exist. The DM efficacy hypothesis is tested by comparing the rate of change of an efficacy outcome measures (e.g. ADAS-Cog) between treatment arms with and without a treatment switch. In their method, they considered the simplest longitudinal design of DM trials with 3 observations per subject and assumed a linear growth or decline pattern for the time 1 to 2 and 2 to 3 in the tested groups. Misspecification of the underlying models poses a danger in the conclusions of these trials. We plan extending this simple assumption to non-linear models and to investigate the effect of dropouts for long term trials of about 2 years or longer through simulations. Simulation results will be presented at the meeting.
Fairness-related decision-making often involves a conflict between egoistic and prosocial motives. Previous research based on Terror Management Theory (TMT) indicates that mortality salience can promote both selfish and prosocial behaviors, leaving its effect on fairness-related decision-making uncertain. By integrating TMT with the strength model of self-control, we propose that managing death-related thoughts depletes self-control strength, thereby impairing individuals' ability to resist selfish impulses during fairness-related decision-making tasks. Additionally, this effect is moderated by dispositional self-control. We tested these hypotheses in two studies. Participants were primed with either mortality salience or negative affect and then asked to made a series of binary choices (equal allocation vs. unequal allocation favoring themselves) to distribute monetary resources. In both studies, mortality salience heightened selfish tendencies, leading allocate resources less equitably. Study 2 further revealed that this effect occurred among participants with low, but not high, self-control. These findings indicate that mortality salience promotes selfishness and inequitable resource allocation, but that self-control can buffer these effects.
Historical information can be used to improve the efficiency of the development process for a medical product (drug, vaccine, or device) and to provide an approach to address indirectly clinically important questions that cannot be addressed directly, e.g., for ethical reasons or because of operational difficulties. Information about control agents such as placebo or active control often can be obtained from previous clinical trials. The historical information may be used to strengthen the conclusions about the treatment-control difference based on previous trials that share the same control with the current trial or to indirectly contribute to the knowledge (e.g., distributional attributes) of the difference between the treatment in the current trial and a control agent available from previous trials but not included in the current trial. Some adjustment to the influence of the historical information is usually required in practice because of heterogeneity between the source of the historical information and the current trial. Considerable research has been conducted towards methods for incorporating historical information, including methods such as power priors, commensurate priors, and meta-analytic predictive priors to adjust for heterogeneity between the historical information and a current trial. The chapter includes a review of the rationale for incorporating historical information, and previous work for attenuating undue influence of historical information. The various approaches are illustrated with real clinical trial examples, including comparison of strategies for employing historical information and down-weighting its influence to account for heterogeneity between the previous trials and the current trial.
Summary Test‐then‐pool is a simple statistical method that borrows historical information to improve efficiency of the drug development process. The original test‐then‐pool method examines the difference between the historical and current information and then pools the information if there is no significant difference. One drawback of this method is that a nonsignificant difference may not always imply consistency between the historical and current information. As a result, the original test‐then‐pool method is more likely to incorrectly borrow information from the historical control when the current trial has a small sample size. Statistically, it is more natural to use an equivalence test for examining the consistency. This manuscript develops an equivalence‐based test‐then‐pool method for a continuous endpoint, explains the relationship between the two test‐then‐pool methods, explores the choice of an equivalence margin through the overlap probability, and proposes an adjustment to the nominal testing level for controlling type I error under the true consistency scenario. Furthermore, the analytical forms of the type I error and power for the two test‐then‐pool methods are derived, and practical considerations for using them are presented.
Traditional oncology trials are usually expensive, take a long time, and suffer from high failure rates. With cancer being one of the leading causes of death worldwide, there is an increasingly urgent need to improve current oncology drug development so that more innovative therapies can become available to cancer patients much sooner. Adaptive seamless Phase II/III designs hold great promise as they can accelerate the decision-making by eliminating the white space between Phase II and Phase III. Most literature discusses trial adaptions by using the same endpoint for both trial stages. However, in oncology drug development, the Phase III endpoint is usually a clinical endpoint, that is, overall survival, which takes long time to observe. Use of the clinical endpoint for adaptive decision may delay the trial adaptation and make seamless Phase II/III designs less appealing. This is one of the main reasons why such designs are less used in practice in oncology drug development. In this article, we would like to address the following two issues: (1) how to incorporate intermediate endpoint (e.g., progression-free survival or objective response rate) data into the decision criteria; (2) how to derive objective adaption criteria from a benefit-cost ratio perspective to streamline the decision-making process. This work is based on two real design examples in the oncology therapeutic area: an operationally seamless design with dose-selection and a statistically seamless 2-in-1 design. However, the general approach is applicable to various other therapeutic areas.
AD endpoints are translated and culturally adapted for multiregional trials with the intent of preserving their cognitive and functional concepts. However, data supporting performance of AD endpoints across countries and cultures are sparse. To examine the potential influence of culture on the measurement of cognition and function, we evaluated regional variability in the ADAS-Cog and the ADCS-ADL using baseline data from a global trial of subjects with mild-to-moderate AD. Mild-to-moderate AD subjects (n=2,221) were recruited from 21 countries for a double-blind, placebo-controlled randomized clinical trial. Countries were grouped into one of six geographic regions. To assess ADAS-Cog and ADCS-ADL item scores across regions, data were stratified according to disease severity (mild, MMSE= 21-26; moderate, MMSE= 15-20). To present the data graphically, standardized z-scores were calculated within each stratum. These standardized z-scores were then pooled across disease severity and re-standardized. ADAS-Cog: The largest absolute z-scores were observed in the Middle East, South America and the Far East, primarily in the Commands, Praxis, and Naming items. The Middle East and South American z-scores indicate subjects tended to perform worse than the overall study population mean, while Far East and Oceania performed better than average. Absolute z-scores for North America and the European Union were the smallest, though this observation may be due, in part, to the sample sizes of these regions comprising a large proportion of the total on which the standardized scores were based. ADCS-ADL: The Middle East and South American z-scores were consistently lower than the mean (suggesting more functional impairment) in basic activities and activities related to communication and engagement. Oceania z-scores were consistently larger than average indicating reduced functional impairment. Assuming no true underlying interaction between geographic region and cognitive and functional performance, these data suggest there may be residual variance in the measurement of cognition and function related to cultural differences across regions. Future evaluation of the psychometric characteristics of the endpoints by region may provide further insight into the influence of cultural variability on measurement error.
