External control arms and debunking real-world data myths

For the healthcare ecosystem to drive research and care, clarity is needed about the opportunities and challenges of ECAs and their value in bringing therapies to patients faster, say COTA Viraj Narayanan and Karla Feghali from ConvergeHEALTH by Deloitte.

External control arms (ECAs) – also known as synthetic control arms – are based on real world data (RWD) pulled from electronic health records, medical claims, wearable devices and other sources of patient data. ECAs are being used to support the primary approval, label expansion of their assets and even go/no-go decisions for trials.

However, like many other emerging healthcare technologies, they have been slow to gain traction and there is still confusion and scepticism around ECAs, stemming from a lack of precedence, step-by-step guidance and data to quantify their impact on drug development costs and clinical outcomes. This uncertainty has understandably given rise to several myths around ECAs that deserve to be dispelled.

Myth 1: There are too many challenges with adopting an ECA

For the innovators and early adopters of the life sciences industry, engineering a regulatory-grade ECA can be daunting. To be clear, there will always be challenges in drug development, and stakeholders are no stranger to this. Still, in 2021, we have seen the industry evolve and develop strategies on how to establish a viable ECA.

Christopher Boone, global head of health economics and outcomes research at AbbVie, advises: “Evidence generation in drug development is ripe for innovation in the digital era and the expanded use of real-world evidence presents a clear enabler to saving time and money. However, if history is any indication, pharma companies have been risk-averse when it comes to fully immersing itself into the expanded use of RWE in our development efforts.

“External control arms tend to de-risk some of the major concerns when applied in the right context. As industry, we need to have a mindset of innovation and an appetite for risk-taking to successfully implement new methodologies like ECAs.”

The industry has looked to regulators for guidance, and the Food and Drug Administration (FDA) has been quick to respond. The US regulator has not only taken a proactive role in exploring the current and future applications of RWD, but also published the framework for its Real-World Evidence Program in 2018, which launched through the 21st Century Cures Act. Within this framework, the FDA seeks to answer outstanding questions on how it will ultimately assess RWD, as well as address any prospective issues, such as cohort bias and data quality.

Beyond the FDA, other institutions have documented what industry standards should look like in assessing RWD in a regulatory context. For example, Washington, DC-based Duke-Margolis Center for Health Policy has established a Real-World Evidence Collaborative think tank with the goal of guiding “high-priority efforts aimed at improving the development and use of RWE” – including the development of master RWE protocols for research purposes.

Lastly, we look to our life science partners themselves, who have undoubtedly spearheaded ECAs for their own pipelines. Their efforts pave the way for others, who can learn from the successes and failures to inform their own ECAs. Many of our innovative life sciences partners pre-socialise their ECA strategy in early discussions with the FDA before finalising the clinical trial to get feedback on the approach from the ultimate decision maker.

Myth 2: Implementing an ECA requires too much money, time and resource

Pushing molecules through the pipeline is already a monumental task, requiring multiple inputs from multiple stakeholders over a period of months, along with significant funding and resources. On average, it takes between 10 and 15 years to develop a new molecule from conception to FDA approval at a cost of approximately $2.4 billion.

In constructing an ECA, there are additional processes involved, such as planning the best study design, selecting the highest-quality clinical RWD source, defining the most appropriate inclusion/exclusion criteria, and conducting all necessary analyses to ensure the synthetic cohort is well matched to the experimental cohort. Each step is crucial and can be costly, but the ultimate gains of an ECA far outweigh these costs.

“Evidence generation in drug development is ripe for innovation in the digital era and the expanded use of real-world evidence presents a clear enabler”

Take patient recruitment as an example: enrolling patients into a clinical trial generally takes up 30% of the trial process, longer than any other step of the trial. With the rise of targeted therapies, complex trial designs, and increasingly niche cohorts, finding and recruiting eligible patients are becoming more and more difficult. Time is not the only factor for drug companies; each recruited patient generally carries a slew of costs – from labs and tests to clinical visits and travel to compensation anywhere from $50-$300 per day. It is no surprise that total per patient costsexceed $100,000 in most oncology trials, with haematology patients costing multiples of that. In sum, with sponsors looking to prove statistically significant treatment effects, hundreds of eligible patients are oftentimes needed, bringing the total direct patient costs over $10 million on the low end. In addition to direct patient costs, there is the cost from CRO resources running recruitment and day-to-day efforts.

According to Jeffrey Morgan, managing director and head of real-world evidence at ConvergeHEALTH by Deloitte: “The use of ECAs has the potential to transform drug development by improving decision-making, decreasing overall development costs, and accelerating approval timelines. We have seen some select examples of this across the industry, but in order to get to state where there is more widespread application of them, a more systematic approach to ECAs must be embedded into the overall development process. This can require new processes, technologies and an organisation mind shift.”

Pending study design, with an external control capable of replacing up to half of these costs (or potentially more), it is no surprise that to an increasing number of pharmaceutical companies, the ECA investment is well worth the money, time, and resources.

Myth 3: RWD is too imperfect to be used to as a comparator arm to a clinical trial

Clinical trial researchers are trained to differentiate signals from noise in their datasets, which is easier to do when the data is very clean and structured from the beginning.

Because physicians in the real world are not following a clinical trial protocol, there will always be less consistency in RWD relative to clinical trial data. It is true that RWD contains more potential bias as a result. To overcome the potential bias, biostatisticians and clinical researchers have developed robust methodology and statistical approaches. Such approaches include inclusion/exclusion criteria for RWD to avoid cherry picking and augmenting RWD to closely mimic the information in a clinical trial.

“While RWD is indeed imperfect today, it is a hurdle that can be overcome”

ECAs are being considered because patient cohorts are becoming so narrow that enrolling patients in a traditional control arm is increasingly difficult (from a cost, time, and ethical standpoint). As a result, in such circumstances where regulators are willing to accept an ECA approach, they too are mindful of the limitations of RWD in the assessment of the ECA. While RWD is indeed imperfect today, it is a hurdle that can be overcome.

Myth 4: ECAs are only used to support regulatory filings

We talk a lot about ECAs in a regulatory context, however ECAs have also been used successfully for go/no-go trial decisions and extensively in post-marketing studies. ECA’s provide actionable information by describing patterns of response and adverse events associated with the drug’s use in the general population.

In a drug development lifecycle, activities that occur from the discovery/translational medicine stage to phase 2A seldom are submitted to the FDA for regulatory decisions. These items – including mechanism of action models, safety signal monitoring, pharmacokinetic research – are often supplemented by external controls consisting of RWD, and are used almost entirely to inform internal go/no-go decisions, which form the basis for whether a drug manufacturer will invest millions if not billions of dollars to initiate or continue clinical trials for an agent.

Furthermore, ECAs are well-documented in the post-approval setting. In the US healthcare system, drug makers seek reimbursement from payers by conducting database and post-marketing studies in which ECAs are often leveraged for price negotiations and coverage. Similarly, in Europe, manufacturers submit lengthy dossiers to each country’s HTA agencies also to seek insurance coverage and ensure patient access. In both situations, ECAs provide the foundation of comparison between an active treatment population and a control cohort.

While ECAs are still in the early stages of adoption, life science stakeholders can be certain about their long-term potential for clinical research. With buy-in, leadership and creativity from across the industry, clinical trial sponsors will quickly start to see the benefits of augmenting traditional strategies with ECAs and curated, comprehensive RWD.

About the authors

Karla Feghali is senior manager at ConvergeHEALTH by Deloitte and Viraj Narayanan is general manager and SVP of life sciences at COTA