Beyond randomised controlled trials: How external control arms could change the game

In this article, we unpack the promise of external control arms, covering benefits, design considerations, challenges, and success stories.

The confines of randomised controlled trials

Randomised controlled trials (RCTs), are widely regarded as the “gold standard” in clinical research. Praised for their scientific rigor and statistical robustness, RCTs facilitate unbiased evaluation of efficacy and safety by randomising patients to either a treatment arm or a control arm with placebo or standard of care. By randomly allocation patients to one of these groups, clinical researchers aim to neutralise the effect of any confounding factors that arise out of the patient’s enormously complex and varied clinical experiences and outcomes. This design feature thus allows researchers to evaluate whether differences in observed outcomes can truly be attributed to the treatment of interest.

Despite its strengths, bringing an RCT to fruition can be a daunting process for clinical trial sponsors, especially if the target patient population is small and hard to find. Depending on study phase, sponsors are often required to recruit hundreds of participants, of which only a part will receive treatment. This is generally regarded as a painstaking task and indeed research shows that 9 out of 10 clinical trials fail to meet their recruitment timeline^[1]. Moreover, there is also evidence suggesting patients are less willing to participate in placebo-controlled RCTs, imposing even greater constraints on recruitment and retention^[2].

Fortunately, recent advancements in data science and technology offer a possible solution for cases where RCTs are simply not feasible. By drawing from the ever-growing pool of data sources outside of the clinical trial, sponsors can generate an external control arm (ECA), also called a ‘virtual’ or ‘synthetic’ control arm. This new method promises to unlock a next wave of innovation in treatments that target rare and life-threatening disease as well as precision medicine. In this article, we’ll unpack both the promises and limitations of this innovative approach, covering key benefits, design considerations, implementation challenges, and success stories.

Introducing externally controlled trials

Externally controlled trials use pre-existing data as a comparator for evaluating the efficacy and safety of a new intervention, rather than relying solely on a concurrent control arm. This design offers a solution to clinical trial settings where recruiting a control arm is not feasible or ethical—for instance in rare, life-threatening, or severely debilitating conditions with no or inadequate treatment options.

An externally controlled trial draws patient-level data from historical clinical trials or real-world data (RWD) sources, such as electronic health records (EHRs), registries, or medical claims. For an ECA to be valid, the data compilation and patient selection criteria must match and closely resemble the clinical trial under investigation. Sponsors should therefore take adequate measures to ensure treatment and control arm are as comparable as possible to mitigate potential bias and confounding factors.

In practice, ECAs can be used to supplement and strengthen both single-arm trials and RCTs. When applied appropriately, they can provide invaluable information at all stages of clinical development—from early go/no-go decisions to approvals and regulatory submissions. Let’s dive into some of the key benefits of this novel methodology.

The benefits of the external control arm

Externally controlled trials bring a number of benefits to modern clinical trial design and conduct (see Figure 1). First, ECAs can drastically reduce the number of clinical trial participants that need to be recruited. Estimates are that an ECA can lower patient recruitment needs by up to 20-50%^[3]. Considering costs per patient average $40k^[4] and clinical trials typically recruit anywhere from 65 patients (Phase 1) to 638 patients (Phase 3) on average^[5], cost-savings can range into the millions.

Next to saving on the resources associated with recruiting and catering to control arm participants, a lower recruitment target also speeds up the recruitment timeline. This alleviates a major bottleneck for clinical trial sponsors, especially those that evaluate treatments for rare diseases where the patient population is small and hard to come by. From the perspective of the patient, externally controlled trials are also more attractive since the odds of receiving treatment as opposed to placebo or standard of care is higher, thus potentially providing a further boost to recruitment.

Given these cost and time-saving benefits, externally controlled trials show much promise for accelerating the evaluation of novel treatments, especially in rapidly evolving fields such as oncology. Moreover, sponsors can use an ECA for generating preliminary insights into the potential efficacy and safety of their intervention—insights crucial for planning and guiding further research and determining whether a more rigorous RCT is needed.

Lastly, externally controlled trials bring ethical benefits since they minimise the number of patients that will receive a placebo or standard-of-care treatment that may be less effective. This is especially holds true when new evidence seems to suggest that the investigational treatment is beneficial.

Figure 1. The benefits of the external control arm in clinical research

Design considerations & challenges

Despite their potential, externally controlled trials can raise serious questions around validity and potential bias. The core challenge lies in ensuring that treatment and control arm are as similar as possible, especially as it relates to known and unknown factors that can affect the outcome being measured. The FDA therefore recently published draft guidance on key considerations for the design and conduct of externally controlled trials, including how to communicate and submit these studies with FDA officials^[6].

For starters, the FDA guideline emphasises that the suitability of an ECA warrants a case-by-case assessment and is highly dependent on key factors such as the heterogeneity of the disease, any preliminary evidence regarding the treatment under investigation as well as the approach for assessing the outcome of interest. Sponsors should investigate whether it is possible to distinguish treatment effect from outcomes that can be attributed to the disease’s natural history, prognostic differences in the study population, lack of blinding, or other factors such as differences in concomitant therapies. To illustrate, externally controlled trials are not suitable for cases where the natural history of a disease is not clearly understood or where the disease is known to improve in the absence of an intervention or with standard of care.

In terms of timing, sponsors should not initiate an ECA after completion of a single-arm trial, but rather should start this process after finalising the study protocol. This is because the protocol should pre-specify design elements pertaining to the ECA, including suitable data sources, baseline eligibility criteria, appropriate exposure definitions and windows, well-defined and clinically meaningful endpoints, cogent analytic plans, and approaches to minimise missing data and sources of bias.

As mentioned, the primary challenge for externally controlled trials is the variability of external data and its fitness for use. Differences in study design, patient population, or treatment standards may undermine the comparability between treatment and control arm, leading to biased observations. Moreover, the FDA highlights the potential impact of disparities in healthcare outcomes due to variability in population demographics, socio-economic factors, and healthcare systems across different regions and time periods.

Concerns around variability relate to both historical clinical trial data and RWD, where the latter warrants even more caution since missing and misclassified data seems to be more potent in RWD. For example, EHR data collected during routine clinical care may include information on lifestyle characteristics, such as alcohol use, where healthcare providers may use different quantitative or qualitative descriptions. As a result, two patients with a similar alcohol intake may be assigned to different categories, posing a serious problem if alcohol use is a potentially important confounding factor (covariate) in the analysis of treatment effect.

To answer this issue, the FDA guideline provides a summary overview of important considerations in assessing the comparability of treatment and control arm data (see Table 1). This overview supports sponsors in their endeavour of finding comparable data and in proactively managing potential threats to the validity of their trials.

Communicating with regulators

Given the novelty of this approach, regulatory requirements are still diverse and evolving. This was confirmed by a recent research study which concluded that regulatory acceptance of the same data package varied across jurisdictions^[7]. Sponsors are therefore recommended to keep an ongoing dialogue and consultation with applicable regulators. The FDA recommends sponsors to consult with the relevant FDA review division early in their drug development program to assess whether it is reasonable to include an ECA.

As part of these discussions, sponsors should provide a detailed description of the (1) reasons why the proposed study design is appropriate, (2) proposed data sources for the ECA and an explanation of why these are fit for use, (3) planned statistical analyses, and (4) plans to address FDA’s expectations for the submission of data.

Real-world applications: Case studies & success stories‍

Although there are numerous cases where an ECA has been used to establish natural history of disease^[8], using externally controlled trials for generating primary evidence is yet to pick up pace. At the time of writing, there is a handful of success stories that, though limited in number, clearly demonstrate potential and regulatory openness to externally controlled trials.

One company that is spearheading the use of ECAs is Medicenna, which recently received FDA approval to design a Phase 3 study that combines both a concurrent control arm and ECA to evaluate its drug MDNA55 for recurrent glioblastoma multiforme (rGMB). The approval was granted after Medicenna successfully conducted a Phase 2 study that included an ECA with patients from rGMB registries going back 5 years. This Phase 2 study marked the first-ever registration trial approved by the FDA that included an ECA, applying the same inclusion/exclusion criteria in both the treatment and control arm by looking at 11 different prognostic factors such as age, tumour size, tumour location, and genetic makeup of the tumour^[9].

Another case study is biopharma company Imunon, which compared data from a Phase 1B study on its drug for advanced ovarian cancer IMNN-001 (formerly GEN-1) with historical clinical trial patients who received standard neoadjuvant chemotherapy. The early information about IMNN-01’s potential comparative treatment advantage over standard of care informed the decision to continue its clinical development under the FDA’s Fast Track designation while also providing practical information on study design, including the appropriate number of patients to recruit for the subsequent phase 2 study^[10].

Historical clinical trial data or RWD?
An externally controlled trial can draw patient-level data from historical clinical trials or RWD sources (e.g., EHRs, registries, etc.). Clinical trial data is generally regarded to be more rigorous, robust, and complete compared to RWD. For example, clinical trial protocols generally include a plan for collecting data on concomitant medications that could impact the observed outcomes, including detailed data on the characteristics and administration of such medications. In contrast, such data is more likely to be unavailable, incomplete, or inaccurate in data collected during routine clinical care.

Introducing the Clinical Insights Exchange‍

Triall is building the Clinical Insights Exchange (CIX), a platform that enables analysis of historical clinical trial data to inform future research in its planning and design. The CIX platform applies advanced cryptographic techniques such as Compute-to-Data to allow for privacy-friendly analysis over aggregated data from clinical datasets and eClinical systems connected to the platform. It therefore enables biopharma companies, clinical CROs, and medical research institutes to provide and consume clinical trial data without compromising data privacy or confidentiality. This allows these companies to generate data-informed insights that promote the speed, resource-efficiency, and predictability of their clinical development activities.

Future outlook: Where do we go from here?‍

Externally controlled trials will undoubtably play an important role in the future clinical trial arena. Today’s success stories demonstrate how fully external and hybrid control designs can accelerate the evaluation of new treatments as well as inform early go/no-go decision moments that guide further research. As clinical trials continue to become increasingly digital, data-driven, and patient-centric, the ECA is likely to gain a strong foothold with sponsors and regulators alike. Ongoing dialogue, capacity-building, and the development of robust guidelines are essential for navigating the challenges and realising the full potential of this innovative approach.