Complex Innovative Trial Designs

Clinical research is ever-changing and increasing in complexity as a result of novel therapies and technologies. In order to address this challenge and maximise the chances of research success, more innovative and creative trial designs are being developed and researched.

The FDA's Complex Innovative Trial Design (CID) Program is a modern initiative used to highlight successful studies which have adopted an innovative trial design. The program has demonstrated the need for increasingly sophisticated and complex trials. The CID initiative allows the showcasing of new trial methodologies, demonstrating how they can improve the efficiency and flexibility of study designs, as well as personalise clinical research, through the use of adaptive and Bayesian frameworks.

What is a Complex Innovative Trial Design (CID)?

Complex Innovative Design (CID) are new and complex methodologies in clinical research, selected to advance and facilitate the use of novel and efficient clinical trial designs in drug development. The FDA provides no fixed definition of what constitutes a Complex Innovative Design, citing that the definitions of “complex” and “innovative” are ever-changing. However, study designs regarded as CID examples include Bayesian frameworks and complex adaptive designs.

The goal of the CID incentive is to facilitate the use of CID methodologies with a focus on later-stage drug development, as well as promoting and advertising successful trial designs which used a CID approach. These successful studies can then be used as case-studies to inspire new designs and future research (see below). Therefore, CIDs are part of the FDA's broader effort to advance modern drug development as a worldwide public health initiative.

Types of Complex Innovative Trial Designs (CIDs)

Many clinical trial designs have successfully adopted the CID framework. The designs below can be used as examples:

Traditional Adaptive Designs:

One example is Response-Adaptive Randomization (RAR): A clinical trial design strategy where the probability of assigning patients to different treatment arms changes dynamically during the trial based on accumulating data about the efficacy or safety of those treatments.
Another example is model-based dose escalation studies: determine the optimal dose of a drug during early-phase clinical trials (typically Phase I), using mathematical or statistical models to guide decisions about dose adjustments.

Biomarker-Guided Adaptive Designs:

Biomarker-guided adaptive designs incorporate biomarkers (biological indicators of disease, treatment response, or patient characteristics) to dynamically adjust trial elements, such as patient enrollment, treatment allocation, or endpoint evaluation.

Bayesian Borrowing Designs:

Single arm trial with external control arm, where all enrolled patients receive the experimental treatment, and the outcomes are compared to an external dataset (the "external control arm") rather than a concurrent control group randomized within the trial.
Dynamic Bayesian borrowing to incorporate external or historical data (e.g., data from previous studies, real-world evidence, or registries) into the analysis of a current study, where the "dynamic" aspect relates to how similar the external data are to the current trial data.

Platform Trial Designs for Vaccines:

Vaccine trials at the time of a pandemic require optimised speed and flexibility, facing challenges such as limited prior data and rapidly evolving pathogens.
The WHO Solidarity Vaccine Trial used a platform trial design which had the ability to evaluate multiple vaccines simultaneously under a shared protocol, which optimised the speed and flexibility of the study and helped combat the challenges of limited prior data and a rapidly evolving pathogen.

Why Use Complex Innovative Trial Designs?

Complex Innovative Designs are increasingly popular for drug and vaccine development in contexts such as pandemics, rare diseases, and precision medicine. Evidence backed by the FDA promotes the use of CID research, suggesting the CID approach to maximize the chance of research success while reducing costs and time. Below are some key reasons to utilize CIDs:

Increased Trial Efficiency:

A smaller sample size is required since CIDs are able to incorporate external data, reducing the number of participants required at study enrollment.
For studies using a platform or master protocol design, there is an opportunity for shared protocols and resources, saving both time and resources.

Faster Development:

Studies with adaptive designs can make real-time adjustments following interim analyses, accelerating decision making and reducing overall study length.
Clinical trials planned over multiple study phases can be combined into one continuous study to eliminate any gaps between phases.

Personalisation and Precision:

Research adopting a biomarker-guided design are able to focus on patient populations who are most likely to benefit from the treatment, therefore increasing the likelihood of success.
Basket and umbrella trials can tailor treatments to a specific subgroup, defined by genetic or molecular markers.

Improved Statistical Power:

Bayesian methods can be used to incorporate prior knowledge and real-time data, improving the effect size estimates and statistical power of the trial.
Studies utilising adaptive sample size techniques can adjust the sample size dynamically, ensuring trials are neither under-powered or over-enrolled, protecting the statistical power as a result.

What are the Challenges of CIDs?

Complex Innovative Designs (CIDs) offer significant advantages for clinical trials, but they also come with unique challenges. The planning required for a trial adopting a CID approach in comparison to a more simplistic trial design is more rigorous, and careful communication with regulatory agencies must be maintained throughout the duration of the trial. Additional time and effort may have to be spent on promotion of the novel trial design, in order for it to be accepted by regulatory bodies or a Data Monitoring Committee (DMC), with a high degree of transparency expected.

Furthermore, CIDs rely on advanced statistical methodologies which have an increased risk of misinterpretation when not monitored and assessed closely. For studies adopting a Bayesian approach specifically, borrowed data must be selected carefully, ensuring the data is of high quality and appropriate for the trial design. In addition, any statistical models adopted for analysis must be appropriately validated prior to being used in a clinical trial, to maintain result validity and avoid any misinterpretation.

Complex Innovative Trial Design Case Studies

Now let's review five FDA case studies which illustrate the practical application of these advanced trial designs. The case studies being reviewed are:

1) EMAS (Epilepsy and Multiple-Arm Screening) Case Study

The EMAS trial design focuses on efficiently evaluating multiple treatments for epilepsy within a single study framework. Unlike traditional trial designs, where different treatments would require separate studies, the EMAS design employs a multiple-arm screening approach. This method allows for the concurrent testing of several therapies, which significantly reduces the overall time and resources needed. This approach is appealing for clinical research in rare diseases, which often encounter difficulties during the enrollment period.

One of the most alluring features of the EMAS design is its ability to stop ineffective or poor performing treatment arms earlier in the study duration. This decision can be made using interim analyses, where data collected during the trial is analyzed at multiple points prior to database lock. During an interim analysis, if a particular treatment poses safety concerns or is lacking in efficacy, that arm can be discontinued; allowing resources to be redirected toward more promising treatment groups. This not only improves the efficiency of the trial, but also accelerates the development of effective therapies for patients.

Moreover, the EMAS design also allows for the integration of new treatment arms during the trial. If new therapies are developed while the trial is ongoing, they can be added without the need to initiate a new study, providing a flexible and adaptive framework that responds to the evolving landscape of epilepsy treatment.

2) Multiple Sclerosis (MS) Case Study

The Multiple Sclerosis (MS) case study highlights the use of a Bayesian adaptive design; a sophisticated approach which leverages accumulating data to modify trial parameters in real time. More traditional trial designs often require pre-specified sample sizes and fixed treatment allocations, which can be limiting if new information emerges throughout the study lifespan. The Bayesian approach, however, allows for continuous learning from the data, enabling adjustments to be made as the trial progresses.

The MS case study investigates an active treatment in comparison to an active control in pediatric patients with MS. The Bayesian component of this trial design involves integrating external research relating to MS into the statistical model, by incorporating previous relevant data as informative meta-analytic predictive (MAP) priors. For this case study specifically, external data relating to MS in adults and children is integrated into the model, to analyse the annualized relapse rate (ARR).

During the MS trial, the adaptive design is used to evaluate multiple treatment regimens simultaneously. As data is collected, the trial design allows for the probability of treatment success to be updated, which informs decisions about treatment allocation. If a particular treatment produces promising results, more participants can be assigned to that regimen and as a result there is an increased likelihood of demonstrating its efficacy.

This approach not only enhances the efficiency of the trial but also increases the ethical appeal. Participants are more likely to receive a potentially effective treatment, and fewer patients are exposed to less effective therapies. Additionally, the Bayesian design facilitates quicker decision-making, potentially shortening the time required to bring effective MS treatments to market.

3) Master Protocol Case Study

The Master Protocol case study illustrates another novel approach in clinical trial design within chronic pain and oncology research. The Master Protocol design allows multiple therapies to be tested under a single, unified protocol, rather than focusing on one single treatment for a disease. This approach is particularly valuable in diseases like cancer, where it may be advantageous to pursue multiple treatment strategies simultaneously, or to combine multiple investigational products as a method of treatment.

The Master Protocol case study uses multiple sub-studies, where the primary endpoint for each sub-study measures the change in pain from the start of the study compared to the end of the study. Analysis for each sub-study then incorporates available data from other relevant sub-studies, such as studies using the same investigational product or treating the same source of chronic pain. This is advantageous as it reduces the number of patients required in each sub-study, which is beneficial for rare chronic pain diseases affecting small proportions of the population. However, investigators must exert caution when selecting appropriate data to incorporate into the study design, as the selection of irrelevant data may reduce the efficacy of this study design.

A key advantage of the Master Protocol is its ability to test several hypotheses within the same trial. For example, a single trial can evaluate different drugs targeting various mutations in a particular type of cancer. This not only accelerates the research process but also provides a more comprehensive understanding of the disease and its potential treatments.

It also allows for a more efficient use of patient populations, particularly in rare diseases or subtypes of cancer where recruiting enough participants for a traditional trial can be challenging.

4) Systemic Lupus Erythematosus (SLE) Case Study

The Systemic Lupus Erythematosus (SLE) trial demonstrates the application of Bayesian approaches in targeting specific patient subgroups; a concept which aligns with the principles of personalized medicine. SLE is a rare autoimmune disease with highly variable symptoms and responses to treatment, thus presenting unique challenges in clinical trial design. Traditional trial designs often struggle to account for this variability, leading to inconclusive results.

For the SLE case study, patients are randomised to receive either 3 doses of investigational product, or placebo. The primary endpoint for this trial assessed an improvement in SLE Responder Index 4 (SRI-4) response at 52 weeks between the two treatment groups.

Under the CID, this trial incorporated an adaptive rule to change the primary endpoint at Week 52 to Lupus Low Disease Activity State (LLDAS) or BILAG-Based Composite Lupus Assessment (BICLA) depending on the results obtained during a planned interim analysis. Furthermore, the adaptive design allows data obtained from all dose levels to be pooled and compared to the data obtained from the placebo group, or allows increased recruitment in a dose group suggested to be clinically effective during interim analyses.

Adopting this approach not only improves the likelihood of demonstrating treatment efficacy, but also minimizes exposure to ineffective therapies, from which patients are less likely to experience clinical improvement. By concentrating resources on the most promising treatment arms, the Bayesian design enhances the trial’s efficiency and ethical standards.

The SLE case study exemplifies how adaptive designs can bring the principles of personalized medicine into clinical trials, making them more responsive to individual patient needs and more likely to produce meaningful results.

5) DLBCL (Diffuse Large B-Cell Lymphoma) Case Study

The Diffuse Large B-Cell Lymphoma (DLBCL) case study showcases a Bayesian adaptive design applied to the evaluation of multiple treatment combinations for a type of non-Hodgkin lymphoma. This approach is particularly useful in oncology, which often features complex treatment regimens and where the stakes are high in terms of patient outcomes.

Similair to the Master Protocol case study, the DLBCL case study features a Bayesian design which allows for the simultaneous testing of multiple hypotheses. This is a significant departure from traditional trial designs, where each hypothesis would typically require a separate study. By testing several treatment combinations within the same trial, the DLBCL study was able to efficiently compare the effectiveness of different therapies in one clinical trial.

A key secondary endpoint for the DLBCL case study is Overall Survival (OS), which was analysed utilizing a Bayesian commensurate prior with a Weibull model, allowing data from external control arms to be incorporated into the study. Similar to other CID studies, the adaptive nature of the trial allowed non-effective treatment arms to be modified or dropped based on interim analyses, sparing resources and patients from potentially ineffective treatments.

This design not only accelerates the identification of effective treatment combinations but also optimises the use of available resources. The ability to test multiple combinations within the same trial framework provides a comprehensive understanding of the disease and its treatment, offering hope for more effective therapies in the future. This is especially important for oncology research, which still has a strong need for improved treatments.

The Value of Complex Innovative Trial Designs

These 5 case studies highlight the transformative potential of Complex Innovative Trial Designs in modern clinical research. From the adaptive flexibility seen in the MS, SLE, and DLBCL studies to the resource efficiency of the EMAS and Master Protocol designs, these innovative approaches represent the future of drug development. They offer significant advantages in optimising resources, accelerating timelines, and enhancing the personalisation of treatments.

Are you considering implementing a Complex Innovative Trial Design in your research? Quanticate’s Statistical Consultancy experts are here to help. With our vast expertise in advanced clinical trial designs, we can help you navigate these complex methodologies and apply them effectively to your next clinical trial. Schedule a consultation and learn more about how we can support your clinical trial design needs.

A Guide to Complex Innovative Trial Designs

What is a Complex Innovative Trial Design (CID)?