Platform-of-1: A Bayesian Adaptive N-of-1 Trial Design for Identifying an Optimal Treatment Among Multiple Candidates

SC28.04 Adaptive Clinical Trial Designs

We are honored to be guest editors for this special issue of Applied Stochastic Models in Business and Industry, dedicated to Kathryn Chaloner's life and achievements. Contributions from her colleagues and friends in this volume are devoted to Kathryn's early research interest in Bayesian optimal design. She was a pioneer in this field and had a broad impact on the subsequent literature on experimental design. Kathryn was internationally known for her research in Bayesian Statistics, which included design of experiments, outlier detection, prior elicitation, and clinical trials, and for her advancements in the study of HIV/AIDS. She was an accomplished researcher, teacher, and mentor. She actively encouraged women and under-represented minority students to join the field of biostatistics and provided mentorship, support, and encouragement. The special issue contains seven papers. The first three papers present some of the recent work in Bayesian optimal designs for nonlinear models, binary responses, and experiments with adversarial components. This is followed by two articles on Bayesian designs in clinical trials. The last two articles of the special issue involve engineering design problems. The paper by Giovagnoli considers the binary response models of Chaloner and Larntz (1989, Journal of Statistical Planning and Inference, 21:191–208) and develops an adaptive version of A-optimal Bayesian designs that are revised as additional data become available during the experiment. Adaptive Bayesian compound designs is also discussed as another extension. Konstantinou and Dette develop approximate Bayesian D-optimal designs for nonlinear regression models that involve covariates that cannot be observed directly. Applications to specific nonlinear models including the exponential regression model are considered and characterizations of the D-optimal saturated designs are presented. Optimal sample size selection for experiments is considered by De Santis and Gubbiotti in an adversarial setting where the decision makers have different prior opinions while having common utility functions and data. Using a Bayesian decision approach with a quadratic loss function, the authors present results for the one-parameter exponential family and investigate effect of the priors on the optimal sample size. Müller, Xu, and Thall discuss advantages and limitations of Bayesian decision theoretic approach in design of clinical trials. In doing so, the authors present a specific case study and point out the limitations such as the computational difficulties involved in solving sequential problems as well as selection of utility functions capturing different stakeholder interests. The article by Ventz, Parmigiani, and Trippa also consider limitations of the Bayesian framework in design of clinical trials and propose a strategy to alleviate these. The proposed strategy involves combining Bayesian designs with frequentist metrics such as error rates, confidence intervals, etc. that are commonly used by medical community and regulatory agencies. The authors show that this can be achieved by an inclusion of frequentist constraints into the Bayesian formulation. The constrained decision theoretic framework is illustrated via several applications and computational algorithms are presented for implementing the proposed approach. Nakamura, Seepaul, Kadane, and Reeja-Jayan discuss design of experiments in materials science. The authors consider use of a batch-sequential experiment and a regression model to determine optimal levels of input variables for synthesis of ceramic materials. It is illustrated that the optimal settings produce more reliable results than other experiments. The final article by Polson and Soyer considers Bayesian design of accelerated life tests for reliability assessment. An augmented probability simulation (APS) approach using Lindley's (1976, Annals of Statistics, 4:1–10) conjugate utility functions is proposed to compute optimal designs in an efficient manner. The proposed approach is illustrated using single and multiple point designs. Use of particle based methods for APS is also discussed as an alternative to Markov chain Monte Carlo methods.

Current tuberculosis (TB) regimen development pathways are slow and in urgent need of innovation. We investigated novel phase IIc and seamless phase II/III trials utilizing multi-arm multi-stage and Bayesian response adaptive randomization trial designs to select promising combination regimens in a platform adaptive trial. Clinical trial simulation tools were built using predictive and validated parametric survival models of time to culture conversion (intermediate endpoint) and time to TB-related unfavorable outcome (final endpoint). This integrative clinical trial simulation tool was used to explore and optimize design parameters for aforementioned trial designs. Both multi-arm multi-stage and Bayesian response adaptive randomization designs were able to reliably graduate desirable regimens in ≥ 95% of trial simulations and reliably stop suboptimal regimens in ≥ 90% of trial simulations. Overall, adaptive phase IIc designs reduced patient enrollment by 17% and 25% with multi-arm multi-stage and Bayesian response adaptive randomization designs respectively compared to the conventional sequential approach, while seamless designs reduced study duration by 2.6 and 3.5years respectively (typically ≥ 8.5years for standard sequential approach). In this study, we demonstrate that adaptive trial designs are suitable for TB regimen development, and we provide plausible design parameters for a platform adaptive trial. Ultimately trial design and specification of design parameters will depend on clinical trial objectives. To support decision-making for clinical trial designs in contemporary TB regimen development, we provide a flexible clinical trial simulation tool that can be used to explore and optimize design features and parameters.

There has been much development in Bayesian adaptive designs in clinical trials. In the Bayesian paradigm, the posterior predictive distribution characterizes the future possible outcomes given the currently observed data. Based on the interim time-to-event data, we develop a new phase II trial design by combining the strength of both Bayesian adaptive randomization and the predictive probability. By comparing the mean survival times between patients assigned to two treatment arms, more patients are assigned to the better treatment on the basis of adaptive randomization. We continuously monitor the trial using the predictive probability for early termination in the case of superiority or futility. We conduct extensive simulation studies to examine the operating characteristics of four designs: the proposed predictive probability adaptive randomization design, the predictive probability equal randomization design, the posterior probability adaptive randomization design, and the group sequential design. Adaptive randomization designs using predictive probability and posterior probability yield a longer overall median survival time than the group sequential design, but at the cost of a slightly larger sample size. The average sample size using the predictive probability method is generally smaller than that of the posterior probability design.

Clinical trial design efficiency is often defined as a design using the fewest subjects to achieve the trial goal with high probability. Multi-arm trials are more efficient than two-arm trials as they compare multiple therapies in a single trial. There has been an increase in the use of adaptive designs, particularly in response adaptive randomization (RAR) and early stopping, to further improve efficiencies in trial design. Designs with a higher probability of participants receiving better treatment and/or the ability to end early can be more efficient and attractive to investigators, participants, and stakeholders, but come with their criticisms. We set out to find the most efficient design for a future study. The purpose of this motivating study is to identify the application of docosahexaenoic acid (DHA) (control, chew, or patient choice of capsule or chew) that results in the highest adherence among pregnant women. The primary goal of this research was to find the most efficient trial design by comparing the performance of fixed and Bayesian adaptive designs (BAD) through operating characteristics using simulations. To evaluate the performance of competing designs, ‘efficiency’ was defined through a novel optimality assessment using a data envelopment analysis (DEA). Several fixed and adaptive trial designs with response adaptive randomization (RAR) and early stopping criteria across different accrual rates were considered. The DEA incorporated a balance in power, total sample size, and expected response outcome. The BAD utilizing RAR and early stopping for success and/or futility was determined to be the most efficient study design.

A landscape of methodology and implementation of adaptive designs in cancer clinical trials

Preface.- The Need for and the Future of Adaptive Designs in Clinical Development.- Regulatory Guidance Documents on Adaptive Designs: an Industry Perspective.- A Commentary on the U.S. FDA Adaptive Design Draft Guidance and EMA Reflection Paper from a Regulatory Perspective and Regulatory Experiences.- Considerations and optimization of adaptive trial design in clinical development programs.- Optimal Cost-effective Go-No Go Decisions in Clinical Development.- Timing and frequency of interim analyses in confirmatory trials.- Approaches for optimal dose selection for adaptive design trials.- A Review of Available Software and Capabilities for Adaptive Designs.- Randomization Challenges in Adaptive Design Studies.- Response-adaptive randomization for clinical trials.- Implementing Adaptive Designs: Operational Considerations, Putting it all together.- Implementation Issues in Adaptive Design Trials.- Implementing Adaptive Designs Using Technology to Protect Trial Integrity, Reduce Operational Bias, and Build Regulatory Trust.- Considerations for Interim Analyses in Adaptive Trials, and Perspectives on the Use of DMCs.- Approaches for Clinical Supply Modelling and Simulation.- Approaches for Patient Recruitment Modeling and Simulation.- A case study for adaptive trial design consideration and implementation.- Design Considerations for a Phase Ib Randomized, Placebo-Controlled, 4-Period Cross-over Adaptive Dose-Finding Clinical Trial.- Continual Reassessment Method for a First-In-Human Trial: From Design to Trial Implementation.- Practical Considerations for a Two-Stage Confirmatory Adaptive Clinical Trial Design and its Implementation: ADVENT Trial.

The utilization of adaptive trial designs can increase the probability of success, reduce the cost, reduce the time to market, and deliver the right drug to the right patient. Two popular definitions of adaptive design are: (1) Chow and Chang’s definition that an adaptive design is a clinical trial design that allows adaptations or modification to aspects of the trial after its initiation without undermining the validity and integrity of the trial (Chow and Chang 2005), and (2) The PhRMA Working Group’s definition that adaptive design refers to a clinical trial design that uses accumulating data to decide on how to modify aspects of the study as it continues, without undermining the validity and integrity of the trial (Gallo et al. 2006). Most recently, the FDA released the draft guidance on Adaptive Design for Clinical Drugs and Biologics, where the definition of adaptive trial design is stated as “a study that includes a prospectively planned opportunity for modification of one or more specified aspects of the study design and hypotheses based on analysis of data (usually interim data) from subjects in the study. Analyses of the accumulating study data are performed at prospectively planned timepoints within the study, can be performed in a fully blinded manner or in an unblinded manner, and can occur with or without formal statistical hypothesis-testing.”

With the advancement in biomedicine, many biologically targeted therapies have been developed. These targeted agents, however, may not work for everyone. Biomarker profiles can be used to identify effective targeted therapies. Our goals are to characterize the molecular signature of individual tumors, offer the best-fit targeted therapies to patients in a study, and identify promising agents for future development. We propose an outcome-based adaptive randomization trial design for patients with advanced stage non-small cell lung cancer. All patients have baseline biopsy samples taken for biomarker assessment prior to randomization to treatments. The primary endpoint of this study is the disease control rate at 8 weeks after randomization. The Bayesian probit model is used to characterize the disease control rate. Patients are adaptively randomized to one of four treatments with the randomization rate based on the updated disease control rate from the accumulated data in the trial. For each biomarker profile, high-performing treatments have higher randomization rates, and vice versa. An early stopping rule is implemented to suspend low-performing treatments from randomization. Based on extensive simulation studies, with a total of 200 evaluable patients, our trial has desirable operating characteristics to: (1) identify effective agents with a high probability; (2) suspend ineffective agents; and (3) treat more patients with effective agents that correspond to their biomarker profiles. Our trial design continues to update and refine the estimates as the trial progresses. This biomarker-based trial requires biopsible tumors and a two-week turn around time for biomarker profiling before randomization. Additionally, in order to learn from the interim data and adjust the randomization rate accordingly, the outcome-based adaptive randomization design is applicable only for trials when the endpoint can be assessed in a relative short period of time. Bayesian adaptive randomization trial design is a smart, novel, and ethical design. In conjunction with an early stopping rule, it can be used to efficiently identify effective agents, eliminate ineffective ones, and match effective treatments with patients' biomarker profiles. The proposed design is suitable for the development of targeted therapies and provides a rational design for personalized medicine.

With better understanding of the disease's etiology and mechanism, many targeted agents are being developed to tackle the root cause of problems, hoping to offer more effective and less toxic therapies. Targeted agents, however, do not work for everyone. Hence, the development of target agents requires the evaluation of prognostic and predictive markers. In addition, upon the identification of each patient's marker profile, it is desirable to treat patients with best available treatments in the clinical trial accordingly. Many designs have recently been proposed for the development of targeted agents. These include the simple randomization design, marker stratified design, marker strategy design, efficient targeted design, etc. In contrast to the frequentist designs with equal randomization, we propose novel Bayesian adaptive randomization designs that allow evaluating treatments and markers simultaneously, while providing more patients with effective treatments according to the patients' marker profiles. Early stopping rules can be implemented to increase the efficiency of the designs. Through simulations, the operating characteristics of different designs are compared and contrasted. By carefully choosing the design parameters, types I and II errors can be controlled for Bayesian designs. By incorporating adaptive randomization and early stopping rules, the proposed designs incorporate rational learning from the interim data to make informed decisions. Bayesian design also provides a formal way to incorporate relevant prior information. Compared with previously published designs, the proposed design can be more efficient, more ethical, and is also more flexible in the study conduct. Response adaptive randomization requires the response to be assessed in a relatively short time period. The infrastructure must be set up to allow timely and more frequent monitoring of interim results. Bayesian adaptive randomization designs are distinctively suitable for the development of multiple targeted agents with multiple biomarkers.

The aim of this study was to identify and characterize all European Medicines Agency (EMA) approvals derived from adaptive designs in clinical trials and to provide an update of the current status of these drugs. Relevant files were identified in the EMA database for annual reports for the period between 2008 and 2020 using a list of suitable keywords related to adaptive designs. We recorded trial characteristics from drug approvals and used Fisher exact test to compare the characteristics. A total of 1,054 EMA approvals were identified, and the percentage of EMA approvals planned with adaptive trial designs increased from 1.85% in the period 2008-2012 to 6.19% in between 2017-2020. A total of 41 approvals were identified among 91 original EMA files that contained adaptive designs. The types of adaptive designs used in clinical trials increased after 2017 where the most common type used was the most common (17/41). Most approvals (32/41) comprised pivotal trials, and most assessments had not been accelerated (38/41). Of 32 confirmatory trials planned with adaptive designs, the proportion of those with additional monitoring (AM) increased significantly (p<0.0001) from 0% in the 2008-2012 period to 90.48% in the 2017-2020 period. The percentage of approved antitumor drugs in approved drugs in ongoing clinical trials was 82.35%, compared to 20.83% in trials that were completed (p=0.0001). The proportion of drug approved but where clinical trials were still ongoing in companies requiring post-authorization safety studies (PASSs) or post-authorization efficacy studies (PAESs) or who were granted conditional marketing authorization (CMA) significantly differed from the group of drugs approved where clinical trials were completed (p=0.0230). A trend showing an increased number of EMA approvals related to adaptive designs was observed for the period from 2008 to 2020. Different types of adaptive trial designs could be encouraged for the designation of clinical trials, especially for antitumor drugs; meanwhile, more stringent monitoring regulations seemed to be conducted for ongoing trials of antitumor drugs with adaptive design.

Cardiovascular diseases are diseases of the heart and blood vessels constituting a major cause of death and disability worldwide. Cardiovascular drug development aims to deliver efficacious drugs to address the public health burden of cardiovascular diseases. However, the high costs associated with cardiovascular drug development, for example due to long-running clinical trials, sometimes including thousands of patients, place a high burden on the development of new efficacious treatments for cardiovascular diseases. Proposals for improving the efficiency of cardiovascular drug development include better disease characterization, more defined target populations, and the use of adaptive clinical trial designs. This dissertation focuses on adaptive clinical trial designs for cardiovascular research. &#13;\nAdaptive clinical trial designs, commonly referred to as adaptive designs, are clinical trial designs with a preplanned modification of design aspects, under some constraints such as preserving integrity and validity of the trials, based on interim data of the ongoing trial. Design aspects which are commonly modified include the sample size, number of doses or treatments, or endpoints. Adaptive designs offer flexibility compared to traditional clinical trials with a fixed design to accommodate newly gained information. However, with the flexibility comes an increased statistical complexity, as adaptive designs require an increased effort to control the probability that the clinical trial declares efficacy of an inefficacious treatment, that is the type I error rate, and to plan the number of patients required such that an efficacious treatment is detected with a high statistical power. &#13;\nThe focus of this dissertation is on two types of adaptive designs: group sequential designs and designs with a nuisance parameter based sample size re-estimation. In group sequential designs, the efficacy of a treatment is tested repeatedly during the conduct of the trial and the trial is stopped early if efficacy of the treatment can be shown with statistical significance. Thus, an efficacious treatment can be detected early in clinical trials with a group sequential design. In designs with a nuisance parameter based sample size re-estimation, the final sample size is adjusted using estimates of the potentially several nuisance parameters based on interim data. Nuisance parameters are for example the outcome variance in trials with continuous outcomes and the overall event rate in trials with count outcomes. The nuisance parameter based sample size re-estimation aims to assure that a clinical trial achieves the target power independently of the initially planned sample size. &#13;\nThe first objective of this dissertation is to study group sequential designs with recurrent events, motivated by clinical trials with patients suffering from chronic heart failure. In clinical trials with patients suffering from chronic heart failure, a common clinical relevant recurrent event outcome is the number of heart failure hospitalizations, which can also be part of a composite endpoint in combination with cardiovascular death. To model heart failure hospitalizations and the respective composite, a negative binomial model and a more robust semiparametric model have been proposed in the literature. However, group sequential designs have not been studied for these models. Therefore, I propose statistical methods for planning and analyzing group sequential designs for negative binomial models and more robust semiparametric models and study their asymptotic properties. Moreover, I show that the proposed planning and analysis methods result in an appropriate power and type I error rate, respectively, for parameter combinations common in clinical trials with patients suffering from chronic heart failure. I put a particular focus on the longitudinal nature of the recurrent events, i.e., a single subject can experience new events throughout the trial, and its consequential on the group sequential designs. The longitudinal natures of the outcomes distinguishes group sequential designs with recurrent events from group sequential designs for other common models, such a continuous, binary, or survival data. &#13;\nA second objective of this dissertation is to study nuisance parameter based sample size re-estimation in three-arm trials with normal outcomes; an investigation motivated by clinical trials with patients suffering from hypertension. A common endpoint in these trials modeled as normally distributed is the change of blood pressure between the baseline measurement and the end of the trial. I show that the ideas for nuisance parameter based sample size re-estimation in two-arm trials can be adapted to three-arm trials and highlight that the corresponding approaches do not result in the desired target power. Furthermore, I modify one of the sample size re-estimation procedures such that it results in appropriately powered three-arm clinical trials. &#13;\nThe third objective of this dissertation is to study incorporating prior information on the variance into the nuisance parameter based sample size re-estimation in two-arm trials with normal outcomes. This objective, too, is motivated by clinical trials with patients suffering from hypertension. I propose several ad hoc rules for incorporating prior information into the sample size re-estimation and by means of Monte Carlo simulation studies I show that the incorporation of prior information can reduce the variability of the final sample size when no prior-data conflict is present. However, I illustrate that in the presence of a prior-data conflict, the designs with a sample size re-estimation incorporating prior information do not convey the target power. I also highlight that common approaches of robustifying the prior information cannot completely mitigate the negative effects of a prior-data conflict without also nullifying the benefits of incorporating prior information on the nuisance parameter into the sample size re-estimation.

Using Bayesian pre-trial simulations to optimize the design of adaptive clinical trials in childhood nephrotic syndrome.

Bayesian and adaptive clinical trial designs offer the potential for more efficient processes that result in lower sample sizes and shorter trial durations than traditional designs. To explore the use and potential benefits of Bayesian adaptive clinical trial designs in comparative effectiveness research. Virtual execution of ALLHAT (Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial) as if it had been done according to a Bayesian adaptive trial design. Comparative effectiveness trial of antihypertensive medications. Patient data sampled from the more than 42000 patients enrolled in ALLHAT with publicly available data. Number of patients randomly assigned between groups, trial duration, observed numbers of events, and overall trial results and conclusions. The Bayesian adaptive approach and original design yielded similar overall trial conclusions. The Bayesian adaptive trial randomly assigned more patients to the better-performing group and would probably have ended slightly earlier. This virtual trial execution required limited resampling of ALLHAT patients for inclusion in RE-ADAPT (REsearch in ADAptive methods for Pragmatic Trials). Involvement of a data monitoring committee and other trial logistics were not considered. In a comparative effectiveness research trial, Bayesian adaptive trial designs are a feasible approach and potentially generate earlier results and allocate more patients to better-performing groups. National Heart, Lung, and Blood Institute.

BackgroundIn recent years, experience on the application of adaptive designs in confirmatory clinical trials has accumulated. Although planning such trials comes at the cost of additional operational complexity, adaptive designs offer the benefit of flexibility to update trial design and objectives as data accrue. In 2007, the European Medicines Agency (EMA) provided guidance on confirmatory clinical trials with adaptive (or flexible) designs. In order to better understand how adaptive trials are implemented in practice and how they may impact medicine approval within the EMA centralised procedure, we followed on 59 medicines for which an adaptive clinical trial had been submitted to the EMA Scientific Advice (SA) and analysed previously in a dedicated EMA survey of scientific advice letters. We scrutinized in particular the submission of the corresponding medicines for a marketing authorisation application (MAA). We also discuss the current regulatory perspective as regards the implementation of adaptive designs in confirmatory clinical trials.MethodsUsing the internal EMA MAA database, the AdisInsight database and related trial registries, we analysed how many of these 59 trials actually started, the completion status, results, the time to trial start, the adaptive elements finally implemented after SA, their possible influence on the success of the trial and corresponding product approval.ResultsOverall 31 trials out of 59 (53%) were retrieved. Thirty of them (97%) have been started and 23 (74%) concluded. Nine of these trials (39% out of 23) demonstrated a significant treatment effect on their primary endpoint and 4 (17% out of 23) supported a marketing authorisation (MA). An additional two trials were stopped using pre-defined criteria for futility, efficiently identifying trials on which further resources should not be spent. Median time to trial start after SA letter was given by EMA was 5 months. In the investigated trial registries, at least 18 trial (58% of 31 retrieved trials) designs were implemented with adaptive elements, which were predominantly dose selection, sample size reassessment (SSR) and stopping for futility (SFF). Among the 11 completed trials including adaptive elements, 6 demonstrated a significant treatment effect on their primary endpoint (55%).ConclusionsAdaptive designs are now well established in the drug development landscape. If properly pre-planned, adaptations can play a key role in the success of some of these trials, for example to help successfully select the most promising dose regimens for phase II/III trials. Interim analyses can also enable stopping of trials for futility when they do not hold their promises. Type I error rate control, trial integrity and results consistency between the different stages of the analyses are fundamental aspects to be discussed thoroughly. Engaging early dialogue with regulators and implementing the scientific advice received is strongly recommended, since much experience in discussing adaptive designs and assessing their results has been accumulated.