A Graphical Comparison of Response-Adaptive Randomization Procedures

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, tuning parameters that affect the variability of the procedure do not have much impact in the finite case. In the binary response case, we find that procedures that target an optimal allocation based on ethical and efficiency considerations generally provide a better compromise design than procedures that do not. In the normal response case, a randomly reinforced urn tends to provide a good compromise procedure.

[1]  Connie Page,et al.  Estimation after adaptive allocation , 2000 .

[2]  William F Rosenberger,et al.  Response‐Adaptive Randomization for Clinical Trials with Continuous Outcomes , 2006, Biometrics.

[3]  Jeffrey R. Eisele.  The doubly adaptive biased coin design for sequential clinical trials , 1994 .

[4]  Feifang Hu,et al.  Asymptotic theorems of sequential estimation-adjusted urn models , 2006 .

[5]  Piercesare Secchi,et al.  Asymptotically Optimal Response‐Adaptive Designs for Allocating the Best Treatment: An Overview , 2012 .

[6]  William F. Rosenberger,et al.  Asymptotically best response-adaptive randomization procedures , 2006 .

[7]  Feifang Hu,et al.  Immigrated urn models—theoretical properties and applications , 2008, 0812.3698.

[8]  Uttam Bandyopadhyay,et al.  Adaptive designs for normal responses with prognostic factors , 2001 .

[9]  N. Flournoy,et al.  Asymptotics in response-adaptive designs generated by a two-color, randomly reinforced urn , 2009, 0904.0350.

[10]  Wenle Zhao,et al.  Quantitative comparison of randomization designs in sequential clinical trials based on treatment balance and allocation randomness , 2012, Pharmaceutical statistics.

[11]  Feifang Hu,et al.  Efficient randomized-adaptive designs , 2009, 0908.3435.

[12]  Atanu Biswas,et al.  Optimal Adaptive Designs in Phase III Clinical Trials for Continuous Responses with Covariates , 2004 .

[13]  L. J. Wei,et al.  The Randomized Play-the-Winner Rule in Medical Trials , 1978 .

[14]  Richard L. Smith.  Properties of Biased Coin Designs in Sequential Clinical Trials , 1984 .

[15]  C. Assaid,et al.  The Theory of Response-Adaptive Randomization in Clinical Trials , 2007 .

[16]  P. Thall,et al.  Practical Bayesian adaptive randomisation in clinical trials. , 2007, European journal of cancer.

[17]  S. D. Durham,et al.  Randomized Play-the-Leader Rules for Sequential Sampling from Two Populations , 1990, Probability in the engineering and informational sciences (Print).

[18]  W. R. Thompson.  ON THE LIKELIHOOD THAT ONE UNKNOWN PROBABILITY EXCEEDS ANOTHER IN VIEW OF THE EVIDENCE OF TWO SAMPLES , 1933 .

[19]  W. Rosenberger,et al.  Randomization in Clinical Trials: Theory and Practice , 2002 .

[20]  Feifang Hu,et al.  Asymptotic properties of doubly adaptive biased coin designs for multitreatment clinical trials , 2003 .

[21]  Anastasia Ivanova,et al.  A play-the-winner-type urn design with reduced variability , 2003 .

[22]  R Core Team,et al.  R: A language and environment for statistical computing. , 2014 .

[23]  Anna Maria Paganoni,et al.  A numerical study for comparing two response-adaptive designs for continuous treatment effects , 2007, Stat. Methods Appl..

[24]  N Stallard,et al.  Optimal Adaptive Designs for Binary Response Trials , 2001, Biometrics.

[25]  L. J. Wei,et al.  The Adaptive Biased Coin Design for Sequential Experiments , 1978 .

[26]  B. Efron.  Forcing a sequential experiment to be balanced , 1971 .

[27]  Nancy Flournoy,et al.  On Testing Hypotheses in Response-Adaptive Designs Targeting the Best Treatment , 2010 .

[28]  D. Robinson,et al.  A comparison of sequential treatment allocation rules , 1983 .

[29]  T. A. Derouen,et al.  G1-minimax procedures for the case of prior distributions in discriminant analysis , 1975 .

[30]  S. D. Durham,et al.  A sequential design for maximizing the probability of a favourable response , 1998 .

[31]  S. Haneuse,et al.  On the Assessment of Monte Carlo Error in Simulation-Based Statistical Analyses , 2009, The American statistician.

[32]  Alessandra Giovagnoli,et al.  On the Large Sample Optimality of Sequential Designs for Comparing Two or More Treatments , 2005 .

[33]  A. Atkinson.  Optimum biased coin designs for sequential clinical trials with prognostic factors , 1982 .

[34]  Alan W. Beggs,et al.  On the convergence of reinforcement learning , 2005, J. Econ. Theory.

[35]  Anna Maria Paganoni,et al.  A randomly reinforced urn , 2006 .

[36]  W. Rosenberger,et al.  The theory of response-adaptive randomization in clinical trials , 2006 .

[37]  William F. Rosenberger,et al.  Asymptotic normality of maximum likelihood estimators from multiparameter response-driven designs , 1997 .