Fixed-sample clinical trial designs are a type of clinical trial in which the patient population and number of patients are set prior to the beginning of the trial. These traditional designs do not include adaptive elements, but their relative simplicity in approach does not imply they require any less rigor or attention to the statistical design.

Here, we discuss the value of fixed-sample designs as well as the role of a simulation-driven approach in avoiding inaccurate estimations of study outcomes and probability of success.

With the advances in adaptive trial design, why use fixed-sample studies?

In practice, many clinical studies are still based on fixed-sample designs, and do not include adaptations based on an interim analysis.

There are many reasons why this is the case. For some sponsors, the administrative and operational burdens of inserting an interim analysis may outweigh the potential benefits of such an approach. Fixed designs may also be seen as a “safe choice” for faster internal and external approvals for larger sponsors. In some studies, regulatory advice may emphasize the collection of data for all patients to make an informed decision on either efficacy or safety endpoints. Finally, misconceptions about adaptive trials being “complex” or unwieldy may lead drug development teams to err on the side of conservative design. Whatever the reason underlying these choices, fixed study designs are still quite prevalent in today’s drug development pipelines of many sponsors.

A simulation-driven approach to fixed-sample clinical trial designs

The relative simplicity in approach of fixed-sample clinical trials does not imply less rigor or attention should be given to their statistical design and implementation. Relying on computational approaches alone may lead biostatisticians to inaccurate estimations of study outcomes and probability of success. This is particularly true in unbalanced designs, in which patient allocation between the study arms is not equal. In such cases, extensive simulation work is required to accurately assess study power and other key design outputs. This is also true in the case of smaller sample studies, relying on Exact Statistics for analysis.

A simulation-driven approach to clinical trial design is just as relevant for fixed designs as for the more complex variety.

The launch of East Horizon^TM, now with a new Fixed Sample module

Cytel’s newly launched East Horizon platform now includes a Fixed Sample module. This module, as the name suggests, offers statisticians the ability to compute and simulate single-arm and two-arm study designs with no interim analyses. In addition, the module contains several Exact tests for small-sample study design and simulation. These capabilities support the three main endpoints: continuous, binary, and time-to-event.

The module also offers four completely novel tests never before included in any East product:

1. Logrank test of the Hazard Rate assuming an Exponential Distribution with Accrual Duration and Study Duration. Event calculation can be implemented with either the Lachin and Foulkes (1986) or the Lawless (2002) method.
2. Logrank test of the Hazard Rate assuming a Weibull Distribution with Accrual Duration and Study Duration.
3. Parametric test of the Hazard Rate assuming a Weibull distribution with Accrual Duration and Study Duration.
4. Win ratio with probability of a tie.

Final takeaways

Despite the less complex nature of fixed-sample clinical trial designs, it remains crucial that sponsors employing these designs approach them with the same level of rigor and attention to statistical design and implementation. A simulation-driven approach can help sponsors avoid common pitfalls.

Cytel’s East Horizon Fixed Sample Module, the first in a series of six revamped cornerstone components of Cytel’s new cloud-based trial design platform, provides statisticians with the tools needed for design optimization and selection of such clinical studies.

Share Via