Interview: EMA’s Peter Arlett reflects on RWE approvals, early learnings from DARWIN

BASEL, Switzerland – Having a common data structure has been a “major enabler” of the EU’s DARWIN network, allowing real-world evidence (RWE) studies to be conducted rapidly. This has facilitated a massive scale up of data collection and increased the power of these studies to generate reliable evidence, Peter Arlett, head of the European Medicine Agency’s (EMA) data analytics and methods task force, told Focus in an interview at DIA Europe last month.

He also spoke about how RWE has evolved in recent years and its use as a tool to support marketing authorization applications (MAAs). The interview has been edited for clarity.

The Data Analysis and Real-World Interrogation Network (DARWIN EU) aims to provide EMA, the European Commission and national competent authorities in the 27 EU member states access to the results of analyses of real-world healthcare databases; it was unveiled last year. Last month, EMA announced that DARWIN EU has completed its first studies and is seeking new data partners. (RELATED: EMA, HMA outline evolution of DARWIN EU real-world database, Regulatory Focus 2 August 2022)

Focus: How has real-world evidence collection evolved in the EU over the past couple of years?

Arlett: Historically, companies set up a registry specific to their product, and that can be fine if it’s the first treatment for a particular disease, but then the next product comes along, and that company sets up another registry, and then another. One of the of the fundamental problems of this is that you can’t compare because its only collecting data on one product.

The approach that we have been promoting is to have a registry based on a particular population of patients such as patients with multiple sclerosis rather than patients treated with a particular product, which had been the long-time approach, because you can see the clinical events and the clinical progression and their journey of different medicines.

Patient groups are very interested in building registry cohorts based on a common group of patients with a particular disease and then collecting useful data. Our position as regulators is that this approach is very useful.

Focus: What do you see as the biggest change in real-world evidence collection?

Arlett: We have used real-world evidence in the industry and the regulatory field for decades, particularly for drug safety and particularly for drug development, and most successful companies have been looking at real-world evidence to understand if there are treatments gaps for products and to inform the dosing of their clinical trials. For example, if you know that 80% of all diabetic patients are between 30 and 60, that helps you design the recruitment of your clinical trials. I made up that number, but you get the point.

What has been changing and changing rapidly are a couple of things. First, access to data is getting better. With the digitization of health care records, we can have access to far, far more data. Also, we have access to insurance records. So that is a game changer. Going back ten years, I would say electronic health records were a minority of health records, it now constitutes the majority of health records. By having it digitalized you can start to analyze it. Also, the methodology has come a long way over the last couple of years.

Focus: What are there the learnings from the DARWIN pilot?

Arlett: First, we are using a common data model – the OMOP [Observational Medical Outcomes Partnership] common data model – meaning that the data has been transformed into a common structure. It also means that the terms used in the data have been mapped … So you know that a heart attack in that database is [the same thing as] an attack of the heart, the slightly different wording has been mapped so you know medical concepts and you can do queries across different data sets using computer algorithms rather than manually pulling the data back. A common data model is a major enabler of DARWIN. It allows studies to be done rapidly and it also allows you to scale up so you can do a study not just on one dataset but when we onboard them, 30 or 40 or 50 data sets, and that increases the power because big is good when it comes to analyzing data, and that is one of the learnings.

The next learning is that we need to go step by step …The point is that we need to bring stakeholders with us, we need to build the network, we need to onboard trading partners and we need to learn as we go. We need to scale up learning as we go. We also need to bring in internal and external stakeholders with us. By internal stakeholders I mean the regulators. Let’s not see clinical trials and real-world evidence as being in opposition. It is really important, there are still some regulators who have a clinical trial background and do not fully trust real-world evidence. So we need to go step by step.

Focus: What are some products that have been approved using RWE?

There are two examples: There was a product authorized in 2020 for spinal muscular atrophy, a horrible disabling disease. The drug is [Novartis’s] Zolgensma. Real-world evidence was an important part of the dossier. There was a Phase 3 study without a control, a single-arm study done under [good clinical practice] GCP and they used historical control data from RWE like a registry. That was considered a very important part of the dossier.

The other one is [Adienne Pharma & Biotech’s] Phelinun for various blood cancers. The evidence was derived entirely from the published literature, it was a combination of clinical trials and observational studies. They added substantively to the safety and effectiveness, from the [European public assessment report] EPAR. You can see that both of these examples are for very rare diseases. If it’s a common disease, there is no reason not to have a randomized controlled trials from a logistics point of view, and randomized controlled trials are superior because it deals with confounding bias through randomization.

Focus: I keep hearing about these patient registries in Denmark, can you discuss these?

Arlett: There is a whole other story in the Nordic countries… In Denmark the healthcare system itself collects health records into registries. These are state run health registries so rather than being run by a patient organization you will find a registry for the whole Danish population then there are sub registries of cancer patients. They have very high-quality electronic health registries, so as a result they can do very good studies.

I gave an example [at the meeting] of the effectiveness study of the COVID-19 study that the EMA commissioned the Danish agency to do a year ago. That was one of the highest quality studies on effectiveness done on COVID-19 and the reason it was possible was because of these high-quality registries in the Nordic countries.

Focus: Any advice of sponsor on making use of better use of RWD?

Arlett: I will give you one stock answer: Seek scientific advice from EMA, and I know you probably read about this before, but getting science advice is proven to help a good study, and if it is followed, can help lead to a marketing authorization, particularly in areas that are evolving rapidly like real-world evidence.

Interview: EMA’s Peter Arlett reflects on RWE approvals, early learnings from DARWIN

Related topics

Recommended content

Welcome to the new RAPS Digital Experience