How can ECAs be used to accelerate the development of better medicines?

This blog is part of The Regulatory Navigator series, where we explore the evolving regulatory landscape with actionable insight from Parexel's experts, sharing their experience to maximize success for clinical development and patient access.

EMA recently released guidance on using single-arm trials (SATs) as pivotal evidence in drug development.¹ The reflection paper emphasizes that SATs are accepted as pivotal evidence for safety and/or efficacy only in exceptional circumstances and where justified and refers to Part 4 of Directive 2001/83/EU, which states that: “In general, clinical trials shall be done as ‘controlled clinical trials’ and if possible, randomized; any other design shall be justified”.² This guidance, discussed in an earlier blog notes that many of the principles for SATs also apply to external control arm (ECA) studies. 

Here, we investigate how sponsors might use ECAs to expand opportunities for optimizing available treatments and accelerate the development of novel therapies; this issue is more extensively addressed in a white paper I authored with my colleagues Katja Hakkarainen and Amy Pace entitled “Harnessing the Benefits of Externally Controlled Clinical Trials (ECTs) to Accelerate Development of Better Medicines.”

ECA data sources

The optimal data sources for ECAs are completed placebo-controlled trials where the data are regulatory-compliant and, ideally, owned by or accessible to the sponsor. This approach is exemplified in the studies supporting the extension of the ravulizumab indication to neuromyelitis optical spectrum disorder (NMOSD). Here, the sponsor not only had access to placebo arm data from an earlier trial with eculizumab but could provide strong justification for an ECA in that this was for a rare disease, the treatment effect size was large, and treatment paradigms had not changed.³ 

Another robust approach to ECA trials is collaboration between sponsors to create a database using data from past studies; the PRO-ACT platform is a prime example, housing ALS patient records from the placebo and treatment arms of 29 Phase II/III clinical trials that multiple companies contributed to.⁴ 

Data from previous trials could, in theory, be used to optimize dosage, support new routes of administration, or develop novel delivery platforms and combination therapies, in addition to supporting the development of next-generation therapies. However, strong regulatory concerns exist when deviating from a randomized controlled trial approach, particularly relating to potential bias and strong persuasive justifications will be required for ECAs to be accepted in these circumstances. 

Often, control data from prior studies are not available leaving real-world data (RWD) as the only data source to construct an ECA. Unfortunately, only in very exceptional circumstances are RWD-based ECAs accepted by regulators as the primary evidence supporting therapeutic effect due to concerns on bias and data robustness. Vaghela et. al. reported that in rare disease, in 55% of applications, FDA expressed concerns about the implementation of RWD such as differences in baseline characteristics, differences in inclusion criteria, imprecise population matching techniques, insufficient information on key input elements, the presence of potentially subjective elements in defining study endpoints, and the possibility of selection bias and measurement error.⁶ Wang et.al reported that EMA, too, frequently has raised concerns about use of ECA’s which primarily related to the Pocock criteria⁷, participant heterogeneity –  particularly due to differences in the standard of care, and limited information on baseline characteristics, especially on significant prognostic factors. EU regulators also expressed concerns over externally derived data, such as missing or different assessments of endpoints, unclear study design, and selection bias.⁷ 

Designing an ECA is a stepwise process:

1. Assessing the feasibility and potential for acceptability of an ECA: Many questions need to be addressed, such as whether robust prior evidence exists, if treatment effect is expected to be relatively large, and ability to justify using an ECA. The validity of an external control depends on how closely it matches that of a randomized controlled trial (RCT). When reliant on RWD, justification becomes more challenging with the need to address issues such as matching of populations including baseline characteristics, accounting for concomitant therapies, ensuring methods of assessment are comparable, dealing with data gaps and data integrity.
2. Assessing the role of technology: AI could play a key role in interrogating data to assess suitability for incorporation into an ECA. For example, RWD could be used to analyze if there are any trends over time e.g. in terms of response rates, standard of care or relationship between response and baseline characteristics. In this way, it can be possible to identify factors that are important to control and those that are not and integrate this information into the design of the externally controlled trial. Another AI approach gaining traction is the use of digital twins which comprises a comprehensive forecast of the participants’ future health based on data from previous studies. This methodology has been qualified by the EMA for use for primary analysis in Phase II/III clinical trials with continuous outcomes, while FDA has noted that this approach does not deviate from their current guidance.⁸ 
3. Focus on data quality and processes: When reliant on RWD, these need to be collected and analyzed following the principles of good data practice and good clinical practice (GCP). Regulators will require documented evidence for absence of bias and FDA may require patient-level data, requiring written agreements with data-base owners. If confidentiality issues are a concern, such data can be provided to FDA directly by the third-party provider either by opening a pre-investigational new drug application (pre-IND) or supplying data via drug master file (DMF). 

Sponsors should discuss approaches with the regulators early, for transforming the data to approved data standards. A range of approaches may be used but with adequate documentation data can be transformed to study data tabulation model (SDTM) datasets and submitted to FDA. There are now some ongoing efforts to connect RWD to CDISC standards.⁹ 

In general, RWD and trial-generated data may differ concerning several variables relevant to the study outcome, and balancing of these variables is required. There are several different statistical methods available to address cohort balancing with estimate propensity score (ePS) methods being a popular choice, although other statistical cohort balancing methods can be incorporated into the protocol. 
Utilization of different statistical teams separated by a firewall is a key element for eliminating bias; one team for developing the ECA and the other for trial data analysis. Finally, once the ECA study is complete, sensitivity analyses can be applied to support the robustness of the conclusions, e.g. ‘tipping point’ approach based on the E-value.¹⁰ 

Proposing an ECA Steering Group  

The application of ECA in pivotal clinical studies supporting regulatory submissions will be subject to intense scrutiny and requires a case‐by‐case analysis to ensure the external data are fit-for-purpose in the specific settings.¹¹ 

Prior to discussions with the regulatory agencies, multiple issues need to be considered. Based on Parexel experience, we propose that an ECA Steering Group be established, with workshops held to evaluate the data and options for ECA application in a structured way and to identify and thoroughly address design challenges and confounding issues. The Steering Group would explore, firstly, the feasibility of an ECA and then develop a study design concept with sufficient detail to engage the regulatory agencies for scientific advice and/or approval. The plan should be informed by assessment of available data, coupled with knowledge of regulatory and scientific precedent, including specific regulatory requirements and recommendations, guidance and publications. 

The ECA Steering Group would best include experts covering the requisite range of disciplines, ideally including epidemiologists, medical and regulatory specialists, statisticians, pharmacologists, and AI experts. Epidemiologists and statisticians can advise and support the adoption of appropriate methods to replicate, as far as possible, the controls and statistical methods that would be applied in a comparable randomized controlled clinical trial.12 

Conclusion

To shift regulatory thinking toward acceptance of an ECA, a robust and detailed justification is required. Such a justification requires multi-disciplinary expertise to address all the factors discussed above. In addition, a case for the criticality of rapid development and approval needs to be made; often, this will be to serve unmet clinical needs or enhance patient care. In circumstances where patient pools are limited, an ECA could allow for better utilization of the limited patient pool by reducing or eliminating the placebo control group, thus facilitating more rapid development of promising therapies and offering potentially effective therapies to more patients earlier.

As a leading global phase I-IV clinical research organization, we leverage the breadth of our clinical, regulatory and therapeutic expertise to optimize clinical development programs. Our team of ex-regulators and industry experts are always available to help you navigate the regulatory requirements of the major regulatory agencies, including an assessment of the applicability of an ECA for your clinical development program. Parexel’s Regulatory Consulting group interprets evolving regulatory requirements, prepares robust submissions, and effectively manages interactions with regulatory agencies, ultimately helping to bring innovative therapies to patients more efficiently and effectively. Please contact us, we are always available for a conversation.

References
[1]     European Medicines Agency: Establishing efficacy based on single-arm trials submitted as pivotal evidence in a marketing authorization. 16 September 2024. EMA/CHMP/430688/2024. Establishing efficacy based on sin-gle-arm trials submitted as pivotal evidence in a marketing authorisation | European Medicines Agency (EMA) (europa.eu) [accessed September 20, 2024] 
[2]    Directive 2001/83/EC of the European Parliament and of the Counci
[3]    Pittock SJ, Barnett M, Bennett JL, Berthele A, de Sèze J, Levy M, Nakashima I, Oreja-Guevara C, Palace J, Paul F, Pozzilli C, Yountz M, Allen K, Mashhoon Y, Kim HJ. Ravulizumab in Aquaporin-4-Positive Neuromyeli-tis Optica Spectrum Disorder. Ann Neurol. 2023 Jun;93(6):1053-1068. doi: 10.1002/ana.26626. Epub 2023 Apr 5. PMID: 
[4]    Neurological Clinical Research Institute. Pooled Resource Open-Access ALS Clinical Trials Database (PRO-ACT). https://ncri1.partners.org/ProACT]    
[5]    Vaghela S, Tanni KA, Banerjee G, et al. A systematic review of real-world evidence (RWE) supportive of new drug and biologic license application approvals in rare diseases. Orphanet J Rare Dis 19, 117 (2024). https://doi.org/10.1186/s13023-024-03111-2
[6]    Pocock SJ. The combination of randomized and historical controls in clinical trials. J Chronic Dis [Internet]. 1976[cited 2018 Apr 9];29(3):175–188. Available at: http://linkinghub.elsevier.com/retrieve/pii/0021968176900448 Accessed 14 Aug 2018.
[7]    European Medicines Agency: Opinions and letters of support on the qualification of novel methodologies for medicine development | European Medicines Agency (EMA) (europa.eu) qualification-opinion-prognostic-covariate-adjustment-procovatm_en.pdf (europa.eu) https://www.linkedin.com/pulse/us-fda-comments-unlearns-procova-methodology-unlearn-ai-l31ic/
[8]    Wang X, Dormont F, Lorenzato C, Latouche A, Hernandez R, Rouzier R. Current perspectives for external control arms in oncology clinical trials: Analysis of EMA approvals 2016–2021. Journal of Cancer Policy: 35,2023, 100403, ISSN 2213-5383, https://doi.org/10.1016/j.jcpo.2023.100403. (https://www.sciencedirect.com/science/article/pii/S2213538323000206)
[9]    Clinical Data Interchange Standards Consortium: Real World Data. Available at:   https://www.cdisc.org/standards/real-world-data Accessed 5 October 2024.
[10]     VanderWeele TJ, Ding P. Sensitivity Analysis in Observational Research: Introducing the E-Value. Ann Intern Med. 2017 Aug 15;167(4):268-274. doi: 10.7326/M16-2607. Epub 2017 Jul 11. PMID: 28693043.
[11]    FDA:  Guidance Document. Considerations for the Design and Conduct of Externally Controlled Trials for Drug and Biological Products. February 2023. Available at:  https://www.fda.gov/regulatory-information/search-fda-guidance-documents/considerations-design-and-conduct-externally-controlled-trials-drug-and-biological-products. Accessed 19 November 2023.
 

Return to Insights Center