LONDON – The timing seems counterintuitive, but in the midst of its Brexit-enforced removal from London to Amsterdam, the EMA has embarked on the task of shaping up its regulatory science to ensure it can properly support development of increasingly complex products.

Earlier this month, the agency published a strategy for regulatory science for 2020-2025, opening the document for public consultation over the first half of 2019, to give stakeholders a say in future priorities and how resources should be allocated.

Regulators are struggling to keep up with the pace of innovation, and in particular to support products based on converging technologies, said Guido Rasi, executive director of the EMA, in a foreword to the strategy.

The aim is to "build a more adaptive regulatory system that will encourage innovation," Rasi said. "Now we want to hear from our stakeholders whether they consider this strategy is ambitious enough."

Amongst other suggestions is a plan to set up an artificial intelligence (AI) test lab; a proposed new system for regulating combination drug/device products; and the implementation of a "learning regulatory system" that gives the EMA the capability to continuously analyze electronic health records and data routinely collected in clinical care.

The EMA should be constantly challenging its capacity to regulate, said Rasi. "How ready are we to engage with emerging science and technological innovations – for example, big data, personalized medicine, novel manufacturing, novel clinical trial designs, the revolution of synthetic biology?"

The agency needs to question whether it has the necessary skills and competencies to guide companies' clinical development programs and if it is giving the right guidance to foster innovation.

Following the consultation, the EMA said it will prioritize the list of actions and agree how to measure its success in implementing each one.

That will allow for the best allocation of limited resources, both in the EMA and the network of national regulators in Europe. The need to prioritize is "made even more acute as a result of the U.K. referendum on membership of the EU," the strategy document says.

In addition to losing the 25 percent of its 890-strong workforce who have elected not to make the move to Amsterdam, the EMA will no longer have access to experts in the U.K. Medicines and Healthcare products Agency.

In preparing to relocate, the EMA wound down its activities during 2018 and there will be further reductions in the first half of 2019. The agency expects to start gradually reinstating work streams in July but faces further upheaval in the second half of the year, when it moves from temporary accommodation to a purpose-built office in November.

Meat on the bone

Regulatory science is defined by the EMA as the range of scientific disciplines that are required to assess the quality, safety and efficacy of drugs across their lifecycle. The strategy document identifies five goals that body of knowledge should address.

Goal 1, "Catalyzing the integration of science and technology in medicines development," and the headlines of the four other goals, may have a ring of motherhood about them. But unveiling the five goals for initial discussion with stakeholders at a meeting in October, Hans-Georg Eichler, the EMA's senior medical officer, promised, "There is substance behind these titles; there is meat on the bone."

In the case of goal 1, building the latest research into drug development will require the EMA to collaborate more closely with academics, research centers and the EU's research infrastructures, and to have ongoing dialogue with companies at all stages of clinical development.

There should be early engagement to qualify biomarkers; work to address the use of 'omics technologies across the development lifecycle; and collaboration with health technology assessment bodies, payers and patients on the use of biomarkers as measures of clinical outcomes.

To help the industry in implementing novel manufacturing technologies, such as replacing batch with continuous manufacturing, the EMA said it will recruit new experts, for example to evaluate computer software used to control new manufacturing processes.

New manufacturing technologies "fit poorly" into traditional regulations, and the agency said it will identify bottlenecks, modernize its regulations, and devise a way of overseeing point-of-care manufacturing.

Similarly, the EMA recognizes the requirement to have greater understanding of, and shape its regulatory response to nanotechnology and to new materials used in pharmaceuticals, which interact with external stimuli to change their properties.

Another feature of goal 1 will be to create an integrated evaluation pathway for assessment of drug/device combinations. That is needed because the legal definitions of pharmaceuticals and medical devices do not fit complex products coming through to market, for which it can be hard to say if they act primarily as drugs, or by physicochemical mechanisms.

Products based on smart materials are particularly likely to be borderline and new expertise is required to evaluate the risk/benefit without increasing the regulatory burden.

Another goal – driving collaborative evidence generation – aims to provide regulators and payers with better information for decision-making. Digital tools should be applied to enable data to be collected more widely and efficiently, from preclinical development to real-world use, the strategy document says.

In preclinical development, the EMA foresees greater use of in vitro human cell and organoid testing and in silico modeling, to replace animal models. In clinical development, the agency wants to enable the use of mobile and wearable technologies to improve patient access to clinical trials, develop novel endpoints and make it easier to incorporate patient-reported outcomes.

As the EMA acknowledges, there is limited experience of those technologies in the drug regulatory system. There is the risk of collecting data that are not relevant and the problem of ensuring data privacy and security to be overcome.

By the same token, the EMA is not in a position to capitalize on advances in AI to analyze and process large unstructured datasets that will be generated by digital technologies. To address that, it proposes to set up an AI laboratory to explore how machine learning and cognitive computing can be used to support regulatory decision-making.

Evidence generation is at the heart of another goal, of improving patient access to medicines. Amongst other measures, that will mean ensuring the evidence needed by health technology assessment bodies and payers is factored in from the beginning of development; setting up a system to allow for access to real-world data across a product's lifecycle; and implementing a "learning regulatory system" through the routine analysis of electronic health records and data collected routinely in clinical care.

Those, and two further goals, of building partnerships with academics to undertake research in regulatory science, and addressing emerging health threats, will be open for consultation until the end of June. The EMA will hold a further workshop to agree on the final strategy in the third quarter of 2019 and will publish the final document at the start of 2020.