It’s a red-hot question: how to bridge the gap between available and desired levels of evidence for a drug’s safety and efficacy? FDA’s approval of Biogen’s Alzheimer’s treatment Aduhelm (aducanumab) was based on data from a surrogate marker – levels of amyloid beta in the brain – which may not correlate with clinical benefit. Evidence of that actual benefit can be gathered post-approval, in a Phase IV trial, the agency said.

Collecting data from the real world is an increasingly popular way to bridge the evidence gap, as regulators prioritize faster access in areas of unmet need. But RWE has its challenges. How to ensure consistent conditions and measurements? If some people are found to benefit, but even more don’t, can you reasonably pull a drug from the market?

In silico supports evidence bridge

In silico data provides a valuable second pillar to support and stabilize the evidence bridge. By using computer simulations to model the impact of a drug or drug-combination on particular physiological functions or pathways, in silico adds knowledge and insight on efficacy and safety, without putting actual patients at risk. It can be used to simulate both patient and population level effects, over time, unconstrained by real-world practicalities (and costs). This has proven particularly valuable  during the pandemic. Disease models are built using an expansive range of existing data and evidence from the literature, reflecting current reality as well as helping predict future reality. Further, they can incorporate both clinical and surrogate markers, providing another route to validating surrogate endpoints.


Aduhelm’s approval was not the first time FDA used surrogate markers. But the decision was controversial because of some unconventional post-hoc efficacy data analyses and concerns over side-effects. (Aduhelm was found to cause brain swelling or bleeding in some instances.)

The independent expert committee unanimously opposed the approval, aware too that post-marketing data collection promises often aren’t kept. Aduhelm’s $56,000/year price tag further inflamed the debate.

One of the three committee members who subsequently resigned called in a New York Times opinion piece for an independent body in the US to make evidence-based assessments of a drug’s impact on patients and on budgets.

There already is such a body. What the non-profit Institute for Clinical and Economic Review (ICER) lacks in legal mandate, it makes up for with an increasingly loud voice on pricing and the wider injustices of the US healthcare system. ICER ruled on Aduhelm a month before the FDA did, and later went on to condemn FDA’s decision .

Swelling healthcare spending means that deliberations by health technology assessors (HTA) increasingly overlap and overshadow those of regulators.

In silico reduces uncertainty in HTA

HTA groups – well-established in Europe and other developed markets – are grappling with a similar evidence gap to regulators. They are starting to systematically incorporate RWE in their reviews as the quality and volume of such evidence grows. ICER is using claims databases and clinical record data as a reality check or some of its drug benefit assessments. NICE’s new strategy  calls on RWE to support guidance that evolves dynamically as evidence emerges.

In silico is next. It offers the ideal complement to RCTs and RWE, helping reduce uncertainty and strengthen the bridge from available to ideal evidence levels. Building this new pillar won’t happen overnight. But since regulators and HTA agencies already accept evidence beyond RCTs, and as the value of computer-powered evidence shines through across all of R&D, it’s just a matter of time.