Clinical trials methods;
Clinical trials methods; and the sample size reassessment permits sample size adjustments to achieve a desired power. According to a review of Bothwell and colleagues (2018) which reviewed 142 adaptive design clinical trials published between 1978 and 2014, the most popular adaptive designs used are the seamless designs (mainly combining phase II and III; i.e., Seamless IIb/III), followed by the group sequential design, and finally the enrichment or subgroups design. Could you describe the three most popular adaptive designs, highlighting the advantages and challenges? Seamless design II/III Fig. 2: Seamless phase II/III adaptive design, also called “drop the losers or pick the winners design” The seamless phase II/III design is one of the most famous adaptive designs, also called “drop the losers or pick the winners design”. This design addresses the early (phase II) and confirmatory (Phase III) objectives, which are traditionally investigated in separated trials, in one trial. As shown in figure 2, during the first interim analysis, the Group C (the “loser”) is not judged effective enough (according to a previous efficacy threshold clearly defined in the protocol). Similarly, Group A (the “loser”) is dropped during the second interim analysis. This leads to a final phase with only two groups, the Group B (the “winner”) and placebo. All the patients included in Group B during the entire trial are included in the final data analysis for efficacy assessment. The advantage of this adaptive design is that there is no stop between the two phases, and doses are selected or stopped during all the trial according to interim results. The seamless design is a popular adaptive design because it allows testing of a larger number of doses with a smaller number of patients at the beginning of the trial. In the middle of the trial, more patients can be allocated to the more superior dose because the inferior ones were stopped before the end of the trial. This design is usually approved by the authorities for important and innovative therapeutic treatments. It should be noted that at the end of the first stage, meaning the first interim analyses, the lack of statistical power may increase the risk of concluding inefficiency of test products. Indeed, the efficacy threshold needs to be prospectively planned and it can be difficult to define it at the time of the protocol development. There is also an extensive logistics requirement as sponsors needs to adapt quickly the products production, and formulations and galenic forms tested in the trial need to be the definitive ones. Finally, because the primary outcome has to be assessed during the interim analyses, this design is not applicable if the primary outcome can only be assessed at the later stage of trials. Group sequential design The group sequential design allows for prematurely stopping a trial due to safety, efficacy (if the treatment is very promising) or futility (if the treatment is not efficient) with the options of additional types of adaptations (as seen above) based on interim results. This design is advantageous in that it can help reduce cost and time as well as minimize the risk of study participants if the test product appears unsafe. However, it can be challenging to pre-define the safety, efficacy, and futility limits for prematurely stopping the trial. Sample size readjustment Usually, the sample size is prospectively determined using a targeted treatment effect size, an assumption about the variability of the primary endpoint, the desired Type I error rate, and the statistical power. Inaccurate estimates of the parameters lead to an underpowered or overpowered study, both of which have negative consequences. Sample size re-estimation designs allow the parameter estimates to be updated during an ongoing trial, which can then be used to adjust the sample size. In theory, the sample size re-estimation can lead to an increase or a decrease in sample size, though in practice this is done only to increase the sample size. One reason is that trials often have other objectives in addition to their primary objective, including safety objectives for which there is a need of a certain amount of exposure data. Another reason is for practical consideration when the recruitment can be completed by the time of sample size re-estimation, particularly when the primary endpoint is disease outcomes and observed late in trials. Finally, the potential to decrease the sample size is best achieved through a group sequential design with an opportunity to decrease the study sample size by early termination at the time of the interim analysis. Blinded methods for sample size adjustment should generally be considered for most studies. They are based on nuisance parameters such as variance. Examining the data in this blinded analysis does not introduce statistical bias, thereby no statistical adjustments are required. In contrast, sample size adjustment based on knowledge of the unblinded treatment-effect sizes at an interim stage of the study can cause an increase in the Type I error rate that can increase risk of a false positive conclusion. To protect against such an increase, a statistical adjustment of theType I error rate is necessary for the final study analysis. What is your opinion about conducting adaptive design? Do you think it’s a real opportunity for industry? Adaptive designs have become popular due to the promise of greater flexibility and efficiency. When adaptive designs are used properly, advantages include a smaller sample size, a more efficient treatment development process (seamless or enrichment designs), 30 No. 3 November/December 2020
and an increased likelihood of correctly answering the clinical question of interest. Thus, it is a real opportunity for industry to save time and money. Additionally, a key benefit of adaptive designs is the ability to reduce the number of patients exposed to unnecessary risk with an inefficient treatment. However, there are some constraints, depending on the type of adaptive designs, such as the possible need of independent DMC, the increase of products logistic, and statistical concerns. Are there some regulatory challenges? Are the authorities supportive of this kind of design? While some ethics committees can be reluctant to authorize some adaptive designs, such as seamless designs with a lack of feedback between phases, regulatory agencies as the FDA or the European Medicines Agency (EMA) are promoting the use of adaptive designs. As an example, among the 142 studies reviewed by Bothwell et al., 9 % and 12 % were used for FDA and EMA product approval considerations, respectively. Additionally, adaptive designs are accepted by the European Food Safety Authority (EFSA) for health claims submission. If the design is built with strong scientific justifications and sound ethical practices and is technically feasible, there is no reason for the project to be rejected. Conclusion The use of adaptive design is still limited even in pharma/drug studies. Regulatory agencies are familiar with adaptive designs but it is still used by only a few sponsors. At first, as it was new, it was used more for innovative treatments which had to be developed very quickly because no solution was available for the patient. So, it was more used for heavy therapeutics. Now, it is becoming more and more popular so we can reasonably think that it will also be considered in the field of nutrition and food! References Guidance for Industry published by the FDA in November 2019 (Adaptive Design Clinical Trials for Drugs and Biologics). Several publications that can help with recommendations for reporting, or systematic reviews with examples: Adaptive Designs in Clinical Trials: Why Use Them, and How to Run and Report Them. BMC Med. 2018 Feb 28;16(1):29. doi: 10.1186/s12916-018-1017-7 Bothwell LE, Avorn J, Khan NF, et alAdaptive design clinical trials: a review of the literature and ClinicalTrials.govBMJ Open 2018;8:e018320. doi: 10.1136/bmjopen-2017-018320. Cerqueira, F. P., Jesus, A. M. C., & Cotrim, M. D. (2019). Adaptive Design: A Review of the Technical, Statistical, and Regulatory Aspects of Implementation in a Clinical Trial. Therapeutic Innovation & Regulatory Science. https://doi. org/10.1177/2168479019831240 Flight L, Arshad F, Barnsley R, et al. A Review of Clinical Trials With an Adaptive Design and Health Economic Analysis. Value Health. 2019;22(4):391-398. doi:10.1016/j.jval.2018.11.008 For more information, please contact Biofortis Mérieux NutriSciences - Europe 3, route de la Chatterie 44800 Saint Herblain, France Phone +33 (0)2 40 20 57 99 biofortis-contact@mxns.com biofortis.merieuxnutrisciences.com Biofortis Research – US 800-A South Rohlwing Road Addison, IL 60101 Phone (630) 617-2000 biofortisresearch@mxns.com www.biofortisresearch.com No. 3 November/December 2020
