Showing papers by "William F. Rosenberger published in 2013"

On Recent Advances in Optimal Allocation Designs in Clinical Trials

Abstract: One of the fundamental questions in the design of clinical trials is how to optimally allocate treatments to study subjects to achieve selected experimental objectives. In this article we review major recent methodological advances in optimal allocation for clinical trials. A literature review shows that starting from the 2000s many new allocation methods have been proposed to enhance the design of multiobjective and multiarm clinical trials. These allocation designs can provide improvements over traditional balanced allocation designs both in terms of statistical efficiency and ethical criteria. We also discuss state-of-the art response-adaptive randomization procedures for implementing optimal allocation designs in practice.

Utility of Covariate-Adjusted Response-Adaptive Randomization in Survival Trials

Abstract: Covariate-adjusted response-adaptive (CARA) randomization is used in clinical trials to balance the competing goals of assigning a greater number of study subjects to the better treatment and achieving high statistical efficiency in estimating treatment effects in the presence of covariates, while maintaining randomness in treatment assignments. In this article, we propose CARA randomization procedures for survival intervention trials with two treatment arms when the primary outcomes follow an exponential regression model. We demonstrate, through extensive simulation, that our CARA procedures can achieve some reduction in the number of events in the study without compromising power and Type I error, compared with balanced randomization designs, including in situations when the exponential model is misspecified. We conclude that the proposed CARA procedures can be suitable alternatives to the traditional balanced randomization designs in survival trials, provided response data are available during the recr...

A Graphical Comparison of Response-Adaptive Randomization Procedures

Abstract: Response-adaptive randomization procedures have a dual goal of estimating the treatment effect and randomizing patients with a higher probability of receiving the superior treatment. These are competing objectives, and no procedure in the literature is “perfect” with respect to both objectives. For clinical trials of two treatments, we discuss metrics for comparing response-adaptive randomization procedures that can be represented graphically to compare designs. These metrics involve the simulated distribution of the set of jointly sufficient statistics for estimating functions of the unknown parameters. We explore the binary response and normal cases, and compare numerous procedures found in the literature. We distinguish between metrics of efficiency and metrics that measure ethical cost. Each of these is a function of the joint sufficient statistics. When graphed against each other, we can gauge competing designs in obtaining these competing objectives. We find that, contrary to asymptotic results, tun...

Randomization in clinical trials: can we eliminate bias?

Abstract: Randomization plays a fundamental role in clinical trials. While many modern clinical trials employ restricted, stratified or covariate-adaptive randomization designs that pursue balance in treatment assignments and balance across important covariates, some clinical trials call for response-adaptive or covariate-adjusted response-adaptive (CARA) randomization designs to address multiple experimental objectives primarily related to statistical efficiency and ethical considerations. In this paper, we elicit key principles of the well-conducted randomized clinical trial and explore the role of randomization and other important design tools in achieving valid and credible results. We give special attention to response-adaptive and CARA randomization designs, which have a firm theoretical basis, but are more complex and more vulnerable to operational biases than traditional randomization designs. We conclude that modern advances in information technology, rigorous planning, and adherence to the key principles ...

Adaptive Bayesian Design with Penalty Based on Toxicity-Efficacy Response

Abstract: The penalized local D-optimal design is introduced by Dragalin and Fedorov (J. Stat. Plan. Inference 136:1800–1823, 2006). We extend the method to the Bayesian realm for the bivariate Gumbel model. Then we conduct a simulation study to compare our method with the trade-off methods of Thall and Cook (Biometrics 60:684–693, 2004). Various measures are employed to present a thorough understanding of both the methods. Our method is more favorable in terms of consistency across simulations and information gain.

Dry-Needling is Effective in Treating Active Myofascial Trigger Points in Subjects with Chronic Myofascial Pain: A Prospective Trial

Abstract: Disclosures: A. Reddy, No Disclosures: I Have Nothing To Disclose. Objective: Development of functional assessment tool for chronic pain patients that provides practitioners with a common outcome evaluation to quantify patients’ progress during a pain management program. It was hypothesized that the functional assessment tool developed will show objective evidence of decreased physical capabilities by patients with chronic pain relative to healthy participants’ outcomes. Design: Case-control study. Setting: Outpatient pain treatment center. Participants: 31 healthy control participants and 104 chronic pain patients. Interventions: Literature review performed initially to assimilate various individual functional tests into a single comprehensive assessment tool which was named, Marianjoy Pain Functional Assessment Tool (MPFAT). Collection of data was obtained utilizing MPFAT on healthy control subjects and initial assessments of chronic pain patients enrolled in a comprehensive pain management program. Main Outcome Measures: The MPFAT outcomes were based upon subject’s assessment of initial self physical ability, visual analog pain score, 6 minute walk test, fast walking 10 meters, handgrip strength, sit to stand repetitions, lifting adjustable weighted object, step-ups, subject’s estimate of perceived exertion, and assessment of pain behavior using Modified University of Alabama-Birmingham Pain Behavior Scale. T-tests and Chi-Square were used for comparison between healthy control subjects and chronic pain patients’ initial assessments. Results or Clinical Course: All performance outcomes measured by the MPFAT showed significant (p<.05) decrease in baseline physical function of chronic pain patients compared to healthy control subjects, with exception of hand grip strength. Most significant mean differences were noted by chronic pain patients’ lower 6 minute walk test, fast walking 10 meters, sit to stand repetitions, and lifting adjustable weighted object compared to healthy controls. Conclusions: Framework provided by MPFAT showed objective evidence of physical decline in the chronic pain patient study group as hypothesized. Thus, it may be plausible for the MPFAT to be further used in other pain management program settings to objectively track patient progress.