Written by Sydney Ringold, Customer Success Manager, and Kevin Trimm, Chief Product Officer
In an ever-changing clinical development environment, sponsors face many challenges when designing clinical trials. These challenges can range from shifting regulations to growing pressure for a rapid time-to-market, and the variety of challenges can complicate trial design requirements. A thorough evaluation of numerous parameters is required in order to choose the best fit or optimal study, and this has led to more complex and innovative clinical trial designs.
When designing trials, biostatisticians have a choice between two main tools: commercial software or open-source development. Both options can be utilized to effectively design a trial, but which approach is best? Here, we will discuss both options and make a case for why this does not need to be an either-or question, but rather a “combined capabilities” approach that integrates open-source and commercial software.
Commercial software tools: Peace of mind, with limitations
Commercial software remain a common and popular choice for clinical trial design, with many sponsors opting for these tools. This choice allows for confident and quick design through validated workflows and pre-coded and verified design types. As a validated choice with a wealth of trial design options, biostatisticians are able to easily and quickly design and compare a variety of trials. Furthermore, utilizing commercial software allows for easy access to expert support in addition to frequent software releases that ensure updates to methodologies and design types.
Although commercial software provides numerous benefits to biostatisticians, there are also drawbacks to this choice. By choosing to design solely in commercial software, biostatisticians can only select from the available design options and analysis methods. While there are numerous designs readily available, it is possible that these may not be fit for the purpose of the specific questions your trial aims to resolve. With the speed and confidence that the use of validated commercial software provides, there is also a degree of inflexibility in terms of the methods available, a trade-off that should be carefully considered when choosing this option.
Opting for open-source code: Flexibility, at a cost
The alternative to commercial software is often seen as open-source development. Due to both the rigidity of commercial solutions and the popularity of open-source languages, developing custom software has become increasingly attractive to many biostatisticians. This option provides almost limitless flexibility in terms of study design options and can often be tailor-made to each unique situation. This flexibility can address gaps in trial designs and functionality that may not be readily available in current commercial software.
However, with this near-limitless flexibility come several potential drawbacks. Building a bespoke solution can be complicated and resource intensive. Time is required for writing as well as validating a custom open-source design option. Additionally, due to the complexity of this approach, it can be difficult to acquire the needed expertise in-house. A biostatistician must have deep programming and statistical expertise to efficiently implement this approach. Open-source studies also require additional resources for both design selection and communication of results, placing an additional burden on the already complex and high-pressure process of clinical development. A biostatistician’s time will be diverted from providing valuable strategic input to the clinical development team towards the numerous necessary software development tasks. The benefits of an open-source solution are therefore not materialized in every case.
A new approach: Software synergy
Rather than approaching the nuanced question of whether commercial software or open-source development is the better approach to clinical trial design, we propose an alternative: a combined-capabilities solution that integrates commercial software with open-source code. Supplementing commercial software with the ability to integrate open-source code harnesses the benefits of both individual approaches and greatly reduces the drawbacks often experienced when an either-or approach is taken. Currently, high-quality commercial software such as Cytel’s East® is available with the ability to integrate custom R code.
There are numerous benefits to this proposed alternative. With commercial software and open-source integration, the biostatistician experiences the same high-quality, validated commercial software with the added benefit of increased flexibility. This allows for the exploration of different designs not readily available and the capability to add custom methods and functionality as needed. Furthermore, with this approach, it is no longer necessary to develop and validate an entire software solution de-novo, saving significant time and resources and placing the focus of the biostatistician back on strategic input and expertise. Teams can better explore and analyze clinical trial designs to find the optimal fit solution faster and with less effort than required with either of the tools separately. This further allows the team to select the design that provides the highest likelihood of success, leading to more efficient clinical development cycles.
There are several identified use cases of integrating R in commercial trial design and analysis software, all of which harness the trusted efficiency of commercial software with the added benefit of increased flexibility that custom code allows. For East and R integration examples specifically, Cytel’s innovation team, led by J. Kyle Wathen, has developed a continuously updating R package called CyneRgy, with sample code and documentation for specific design needs, intended for use with Cytel’s software solutions. Below are some examples:
-
- New analysis methods not native to Cytel’s software, including Bayesian analysis approaches, can now be integrated seamlessly to a design. In one CyneRgy example, a go/no-go decision rule can be added using confidence intervals for a binary endpoint.
- Custom futility rules can be incorporated for a wide variety of study designs, beyond those currently available in our software offerings. CyneRgy contains an example of code that allows just such flexibility, including the ability to assess multiple futility rules side-by-side.
- Custom R code can also be deployed for a unique approach to patient simulation. CyneRgy offers several examples for Bayesian methods for patient simulation ready to use in a trial design. In one example, the software-native patient-simulation capability is replaced with a bimodal distribution allowing the user to specify a certain proportion of non-responders in a particular sample.
- New analysis methods not native to Cytel’s software, including Bayesian analysis approaches, can now be integrated seamlessly to a design. In one CyneRgy example, a go/no-go decision rule can be added using confidence intervals for a binary endpoint.
The above are just a few examples illustrating the promise of this new integrated approach. With increased sponsor interest and use of this feature, the CyneRgy package will evolve over time to provide additional examples and sample code.
The future is integrated: Commercial software & open-source code
As the clinical development environment continues to change and increased pressure is felt for unique, complex, and innovative trial designs, an integrated, combined-capabilities approach will fill the gaps in commercial software while not requiring the intensive time and expertise needed to develop a bespoke design. The future of clinical trial design is integrated with biostatisticians able to harness the benefits of both commercial software and open-source solutions while limiting the drawbacks experienced with each approach individually. We envision innovative designs that draw upon the established, validated backbone of commercial software with the added creativity encouraged by open-source code.
Thank you to Boaz N. Adler and J. Kyle Wathen for their insights on this piece.
Interested in learning more? Sydney Ringold, Customer Success Manager, and J. Kyle Wathen, Vice President, Scientific Strategy and Innovation, discuss the integration of East® and R, including best practices and examples, in their recent webinar. Click to watch on demand:
About Sydney Ringold
Sydney Ringold is Customer Success Manager at Cytel. She has experience in both commercial and research aspects of clinical trials through co-founding a biotechnology start-up during her undergraduate years. Additionally, Sydney has coauthored numerus publications and conference presentations ranging from health communication to diagnostic techniques for Alzheimer’s disease and breast cancer. Sydney has experience in multiple programming languages including R, Python, and MATLAB.
About Kevin Trimm
Kevin Trimm is Chief Product Officer at Cytel. Kevin has 20 years of experience at the intersection of data science technology and drug development. He is responsible for Cytel’s software product strategy, development, and commercialization. Under Kevin’s leadership, Cytel is delivering a comprehensive and transformative software as a service (SaaS) clinical trial design platform. The platform includes over 30 years of proprietary innovation, coupled with open-source technology and artificial intelligence.
Read more from Cytel Perspectives:
Sorry no results please clear the filters and try again
Negative Binomial Distribution in Group Sequential Designs
Metadata Repositories: Overcoming Challenges with Automation
Quantitative Strategies for Rare Disease Clinical Trials
