Adaptive by Design

Contributed by:

Denise Myshko

NOTE: The content below contains the first few paragraphs of the printed article and the titles of the sidebars and boxes, if applicable.

Adaptive trial designs have the potential to speed development and reduce overall costs — that is, if sponsors invest in up-front planning and can implement these designs correctly. Interest in adaptive designs has risen in recent years. These types of studies involve a broad class of methods and can have a positive impact on any program across the ­different stages of ­development. Experts say the flexibility of an adaptive design allows sponsors to make course corrections during a trial based on interim data. But experts stress these changes should be preplanned decisions based on simulations and ­­well-thought out, up-front planning. The Potential Our experts discuss how adaptive designs address the challenge of getting new products to the market faster and with… Sidebar: The FDA Draft Guidance The Adaptive Process THOUGHT LEADERS Randy Anderson, Ph.D. VP, Global ­Product Development, PPD Inc., a global contract research organization. For more information, visit Molly Blake-Michaels. Director of Clinical Services, ClearTrial, a provider of clinical trial operations software. For more information, visit, or e-mail Susan Bornstein. Executive VP, eClinical Solutions, which provides clinical data management and consulting services. For more information, visit, or e-mail Michael Boyd, Ph.D. Senior Director, Global Biostatistics, INC Research, a therapeutically focused contract research ­organization. For more information, visit Sylva Collins, Ph.D. VP, E-Clinical Solutions, Aris Global, which provides drug safety monitoring, pharma­covigilance, regulatory compliance, and clinical trials management software. For more information, visit ­ Jennifer Dudinak, Pharm.D. Global Head, Inflammation, Regulatory Affairs, Roche, a research-focused healthcare company, which has combined strengths in pharmaceuticals and ­diagnostics. For more information, visit Brenda Gaydos, Ph.D. Senior Research ­Advisor, Eli Lilly and Company, which is developing pharmaceutical products by applying the latest research from its own ­worldwide laboratories and from collaborations. For more information, visit Mary F. Johnson, Ph.D. Executive VP, ­Biostatistics, PharmaNet Development Group Inc., a global drug development services company that ­provides a comprehensive range of services. For more information, visit Bryan Luce, Ph.D. Senior VP, United BioSource Corp., a global medical affairs and research ­organization. For more information, visit ­ Olga Marchenko. Global Head of Data Services Therapeutic Consulting, i3 ­Statprobe, which is the clinical data services ­division of i3. For more information, visit Judith Quinlan. VP, Adaptive Clinical Trials, Cytel Inc., which is a provider of adaptive clinical trial design services and specialized statistical ­software. For more information, visit Bill Sietsema, Ph.D. VP, Regulatory Consulting and Submissions, Kendle, a global full-service ­clinical research organization. For more information, visit Interest in adaptive designs has risen in recent years. These types of studies involve a broad class of methods and can have a positive impact on any program across the ­different stages of ­development. Experts say the flexibility of an adaptive design allows sponsors to make course corrections during a trial based on interim data. But experts stress these changes should be preplanned decisions based on simulations and ­­well-thought out, up-front planning. The Potential Our experts discuss how adaptive designs address the challenge of getting new products to the market faster and with greater efficiency. Collins. Aris Global. Adaptive designs, when implemented fully and successfully, will make clinical development faster. They will reduce development costs, and patients will not be exposed unnecessarily to unsafe drugs or drugs that are not efficacious. If done right, adaptive designs can be profound. Blake-Michaels. ClearTrial. There is a realization that adaptive designs can be a more efficient way of gaining information about a drug. Adaptive designs increase the likelihood that a compound that actually works will make it to market. They can also be used to identify drugs that don’t work earlier; this is an enormous advantage, as determining if a drug doesn’t work can take a considerable amount of time. Johnson. PharmaNet. Stopping trials early or reducing treatment group sizes save both time and money. We can also check study design assumptions earlier. If the assumptions were wrong, we can modify the ongoing study instead of wastefully abandoning it. Adaptive designs have the potential to accelerate almost every phase of drug development. They are most commonly used during the early exploration of drug dose effects, in Phase I. But adaptive designs have even greater potential value in later-phase trials, where they make it possible to adjust treatment group sizes, terminate one or more treatment groups, drop ineffective doses, select the best dose for future studies, prepare early for regulatory filing, or terminate the study of an ineffective or dangerous drug. Sietsema. Kendle. Adaptive trials can result in speedier, more efficient trials. A good example of that is the seamless move between Phase II and Phase III; the dosing range can be determined in Phase II and the product moved into a Phase III program without having to do an extensive analysis of the Phase II results before proceeding. This can save at least six months or more during development. Boyd. INC Research. An adaptive design is not a shortcut nor is it a mechanism for avoiding the requirements for adequate and well-controlled trials. Adaptive trials do, however, have several potential advantages, including the ability to check design assumptions for future studies and transition from one phase to the next with a higher chance of success. Key to this process is ensuring that there is a factor of trust with the FDA. Clear documentation of the process for unblinding and any modifications must be provided to ensure the validity and integrity of the trial. Luce. UBC. An adaptive design creates a continuous learning process that allows the clinical team to be much more flexible in determining what works best. The experience to date is that adaptive designs have led to a more efficient means for making decisions. An adaptive design means that the dosing, the types of patients, or treatment settings can be adjusted during the course of the trial as evidence accumulates, as long as the midcourse decision rules were pre-specified. Boyd. INC Research. If properly designed, any number of adaptations can be made to increase efficiency. Examples include sample size re-estimation to maintain power, adjusting eligibility criteria, dropping ineffective treatment arms, and stopping the trial early because of futility or efficacy. The goal is clearly to make good decisions more quickly, without assuming too much risk. This needs to be a joint effort among clinical, regulatory, and statistical colleagues. The FDA Draft Guidance The FDA released a draft guidance in February 2010 for pharmaceutical and biologics developers on adaptive design clinical trials. The guidance covers the clinical, statistical, and regulatory aspects of adaptive design clinical studies. (For more on the draft guidance, see the digital edition of PharmaVOICE.) Gaydos. Lilly. The guidance provides a path forward, a blueprint to successfully implement these new methods. Regulators have taken the tack of covering the full range of adaptations at a principle level instead of focusing on specific methodology. This is going to enable the guidance to be relevant over time since new methods will continue to be developed. This also enables sponsors to determine where the opportunities may be and, once determined, how to engage the regulators to implement successfully. Sietsema. Kendle. The draft guidance provides clarity; up until now there was confusion. When people heard the term adaptive designs, they heard faster and cheaper. But the improper use of adaptive designs can introduce bias within a clinical study and can result in making incorrect conclusions and that’s a major concern for regulatory agencies. The adaptive design guidance does a nice job of explaining the permissible limits under which adaptive designs can be used; the factors that cause concern when introducing bias; and how to organize the data to be acceptable to regulatory agencies. Blake-Michaels. ClearTrial. The agency’s concern is to make sure that they still receive adequate characterization of the dose response and safety profile of the drug in the marketing application. We really needed this guidance. A lot of companies that would have liked to start using adaptive designs in their drug development programs have been waiting for the guidance because there is risk involved in trying a new approach. Anderson. PPD. The guidance has really been needed. It sets high standards for rigor. It may have an impact of slowing the adoption of adaptive designs in adequate and well-controlled studies, but there are going to be instances where an adaptive design is sorely needed because of the cost of the confirmatory trials. Johnson. PharmaNet. The FDA’s draft guidance distinguishes between adaptive designs that are less risky and are already commonly used in Phase III trials for example group-sequential studies, internal pilot studies with blinded interim analyses to verify assumptions, conditional power calculations to predict trial success, and designs whose properties are not as well understood, such as Bayesian techniques, sample size re-estimation, and response-adaptive methods. Marchenko. i3 StatProbe. We have one concern about the guidance. Regulators separated adaptive designs into two big types of designs: well-understood adaptive designs and less well-understood designs. They are concerned about confidentiality and integrity of the ­trials, especially when less well-understood adaptive methods are used. There are several examples of successful less well-understood adaptive designs used in pivotal trials that improved and accelerated the drug development. Bigger companies are willing to take a risk to make these designs well-understood ones, while smaller companies may not be able to afford to take such a risk. I think we need to hear that regulators are open to considering various kinds of adaptive designs including less well-understood designs in adequate and well-controlled studies when the designs are planned and implemented properly. The Challenges Our experts discuss the importance of addressing issues, such as ensuring integrity of the study and controlling against bias. Sietsema. Kendle. In many adaptive designs, the data have to be unblinded in some way so that decisions can be made on how to proceed with the trial. The unblinding of the data creates a potential for bias in how the rest of the study is conducted or how the data are processed or interpreted. If bias is introduced, this can reduce the credibility of the study. Quinlan. Cytel. From a practical aspect, there are more complex logistical issues that need to be addressed. For example, one such issue is the timely availability of data for the interim analysis. Additionally, for adaptive dose ranging studies, we can’t predict at the start of the trial what will be needed, so more thought needs to be given to how to manage drug supply and distribution during the course of the trial. Luce. UBC. Another challenge is that most trialists and most biostatisticians are not trained to do this type of work. The traditional double-blind frequentist hypothesis testing randomized controlled trial is a tried-and-true method that people are very comfortable with and many people are not ready to accept a new paradigm, even when there is good evidence that it may lead to a more efficient process. There is a lot of inertia because of that. The challenges come when we get beyond Phase II and into Phase III or Phase IV. On the face of it, Phase IV trials would seem to be particularly well-suited for adaptive designs. The concept of adapting a comparative trial or registry for the real world in a systematic way just makes sense, but this really hasn’t happened yet. Best Practices Our experts discuss the important factors to consider when planning an adaptive study. Blake-Michaels. ClearTrial. Adaptive trials involve a lot of up-front planning. It can take twice as long to plan an adaptive trial as a traditional trial. Simulations are required not only for patient outcomes, such as efficacy, but also to understand the drug supply, head count, and budget requirements. This really represents a paradigm shift in planning. Teams need to think in terms of ranges of possible outcomes and plan a variety of different scenarios to make sure they understand the impact each would have on the program. A lot of the benefits of this type of study — such as time savings or earlier identification of doses — can be lost if adequate long-term planning doesn’t occur up front. Johnson. PharmaNet. The adaptive methodology to be used and the modifications to be made must be specified at the outset of a trial, not devised ad hoc during the course of the trial. Similarly, interim analysis procedures, group-sequential stopping boundaries, risk/benefit determinations, and criteria for maintaining the blind should be specified at the outset. In a Phase III trial, it’s best to plan for just one modification during the study, not several. The FDA is very clear in urging drug sponsors to interact with the agency during the early planning of adaptive designs and gain agreement on the methodology before launching any studies. These include methods for controlling bias, maintaining blinding, and providing documentation of statistical procedures, including simulations or other calculations necessary to support study results. Gaydos. Lilly. It is important to have clarity on the rationale for doing an adaptive design over a conventional design. These aren’t something to do because they are neat or new. In addition, more up-front planning is required. A team needs to appropriately adjust its project management plan. A lot of clarity on the data flow is needed to implement these trials; they need to run efficiently. Teams need to ensure they have good firewalls to protect data integrity. Quinlan. Cytel. Not every trial has to be an adaptive trial. Like any trial it has to make sense within the clinical development plan and it has to add value. Unfortunately because adaptive trials have become a buzz term, there are a lot of people who say they can do adaptive trials. So it’s very important to make sure the right people are working on these trials. These should be people who have experience and know the planning and execution issues involved. Dudinak. Roche. Adaptive trials should be evaluated in the scope of all other trial designs. Companies should consider how this approach fits into the overall development program and life cycle of the product. Team discussions and the strategic evaluation should be initiated as early as possible. There are times when an adaptation may not be appropriate. There isn’t a one-size-fits-all approach as it relates to adaptive designs. Generally, a very quick enrollment rate isn’t advantageous to an adaptive design. Companies need to look at adaptive studies from a cross-functional perspective. There are multiple stakeholders that need to be involved: clinical, regulatory, operations, technical, safety, biostatistics, and data monitoring. It also requires support from trial simulation to justify the design and control for a type 1 error. Marchenko. i3 StatProbe. Implementation of adaptive designs involves integration of data capture, drug supply management, data analysis, and interactive communication systems. A key operational aspect for resources means more planning is needed up front if additional resources and monetary investment might be needed for such trials. Clean data are definitely desirable, but quality needs to be balanced against timelines. If the decision is made during the interim analysis on primary endpoints only, then primary endpoints should be clean. Bornstein. eClinical Solutions. A key practice is to use technology to have cleaner data sooner. Leveraging electronic data capture and having the data in a central repository are important. It’s key to interact with the data ongoing and have a proactive data management plan. In an adaptive design or in any clinical trial, the data are vital. Updating SOPs up front is one best practice that needs to occur to incorporate adaptive trial design into the process; for example this would allow for the protection of the study blinding and eliminate the introduction of bias. Collins. Aris Global. Management and leadership are key. Hundreds of people have to do things more quickly than they would normally be accustomed. They are not doing more work, but they have to do it faster and in an environment with a lot of pressure and a lot of visibility from higher levels. EDC and standards provide the brains to be fast. Management is the spine. If the sponsors are going to seek approval based on trials that involve multiple looks at the data, they must show efficacy with greater statistical certainty. If sponsors are not fast, what happens is that they would have done a less efficient conventional study that is burdened by a statistical penalty. Even if the database is locked fast and the analysis completed faster, there is one more thing that is needed: a sponsor has to act on the information quickly. If it takes weeks to make a decision on how to change the course of the study, nothing much was really accomplished. In fact, one might pay a penalty for the multiple looks at the data. Boyd. INC Research. The FDA’s major concern is bias. It is imperative, therefore, to have clear documentation about SOPs and firewalls to protect against bias. It is also important to recognize that this is a new process and each adaptive design is different. It cannot be stressed too much that early communication with the agency is vital. The FDA will need to review and understand any proposed design. F PharmaVOICE welcomes comments about this article. E-mail us at What is an Adaptive Design? Adaptive trial design refers to a clinical trial methodology that allows trial design ­modifications to be made after patients have been enrolled in a study, without ­compromising the scientific method. In order to maintain the integrity of the trial, these ­modifications should be clearly defined in the protocol. Melissa Hammond is Managing Director of Snowfish, which provides insights to healthcare, life-sciences, and biotechnology companies. For more information, visit “From a marketing perspective, adaptive trial design appears to make excellent sense, ­particularly with respect to large trials with ­extended follow-up. Ideally, analysis of the overall clinical data and practice landscape for an area is ongoing, even while a large-scale study is being conducted. Such analysis may indicate subtle changes in the clinical and marketing landscapes, which if not addressed, can pose major marketing barriers moving forward. An adaptive approach can therefore allow for appropriate modification of the trial in order to proactively address key issues, thereby obviating a need for multiple post-hoc analyses (which many times are poorly powered) and additional ­ follow-up studies.” Andrea Perrone, M.D., is VP, Clinical ­Operations and Medical Director at BioClinica Inc., a global provider of integrated, technology-enhanced clinical trial management services. For more information, visit “In February 2010, the FDA, CDER, and CBER released a draft guidance on adaptive design in clinical trials for drugs and biologics. Their mission is to provide further clarity on how trials can be designed more efficiently and to maximize the knowledge accumulated. The metrics for efficiency can be monetized by the pharma and biotech industry in three main areas: by decreasing the number of patients needed to reach statistical ­significance; decreasing the length of the study; and potentially expanding the data gathered based on dose and response. Specific modifications that may occur throughout an adaptive trial and are defined a priority include study eligibility, randomization, treatment regimens, ­concomitant treatment, schedule of assessments, ­primary endpoint, including outcome assessments, order of secondary endpoints and/or analytic ­methods to evaluate endpoints. A critical factor to the success of an adaptive design is the prospective nature of the modifications that are proposed before any unblinding of the data.” Jason Rock is Chief Technical Officer at GlobalSubmit, a developer of software designed exclusively for the review and validation of electronic common technical document (eCTD) global submissions. For more information, visit “The main reason why drug development is so ­costly is the time needed for clinical trials. Even when a drug is not working, standard trial design forces sponsors to run a trial to the end. If trial data were made available earlier, sponsors would know that the trial will be ineffective earlier. Accordingly, the cost of bringing drugs to the market would drop significantly. Adaptive trials will greatly reduce the cost of drug development by allowing a sponsor to quickly learn that a drug is not working and stop the trial. The real sticking point between the FDA and industry with adaptive trials is around how well the protocol needs to define recruitment. Clearly allowing the sponsor to unblind the study and add one patient at time is not sound but waiting until the trail is ­complete is too costly.” PharmaVOICE asked industry experts how pharma/biotech companies can use adaptive designs for more efficient drug development.

Posted in:

Post a Comment

You must be logged in to post a Comment.