Look behind the lecture: Real-world evidence uses in clinical development to commercialization

In this feature, David Thompson (Syneos Health, NC, USA) discusses his presentation from ISPOR (18—22 May 2019, New Orleans, LA, USA) on the use of real-world evidence (RWE) in the pharmaceutical industry.

Sep 10, 2019

Could you briefly introduce yourself and describe your current role at Syneos Health?

I am a Senior Vice President in the Real World and Late Phase Research (RWLP) Group at Syneos Health (NC, USA). Syneos Health comprises a contract research organization (CRO) as well as what we refer to as a 'contract commercial organization' (CCO). We provide a balance of clinical research and commercialization services, in between which sits RWLP. I like to think that RWLP sits as the very fulcrum of the company; we help our clients navigate the difficult transition from clinical evidence generation needed to obtain regulatory approval, to real-world evidence (RWE) generation needed to support product commercialization.

My specific role in this is research design – ensuring that our clients’ investments in RWE generation are on target to meet the evidentiary needs of the health system stakeholders who will be making decisions regarding the pricing, reimbursement and use of their products. I am a health economist by training and some of the first studies that I cut my teeth on involved analyses of healthcare claims databases. The term 'real-world data' (RWD) did not exist back then but that is exactly what it was. RWE has become a high growth area; this has been my focus at Syneos for over 3 years.

In your presentation at the Pharmaceutical Pricing & Market Access (PPMA) Summit (19—20 March 2019, Amsterdam, Netherlands) earlier this year, you discussed how, historically, RWE design and execution has mistakenly followed a linear process: how do you think this needs to be improved?

In that talk, I contrasted how manufacturers pursue clinical evidence development compared with RWE development. The former involves a continuous process beginning with upfront stakeholder engagement. This stakeholder input is utilized to inform study design and execution, after which stakeholder engagement resumes to share findings and map out plans for additional studies as the product progresses through development. Surprisingly, we have observed that companies have, historically, abandoned this continuous process when it comes to RWE generation. Instead, they favor a linear process; study design and execution are undertaken in a ‘hidebound’ manner, with stakeholders only involved at the backend when results are shared.

 “You are going to have to develop a program of evidence generation that meets the differential needs of different stakeholders.”

Not surprisingly, this approach has been met with mixed results. Stakeholders frequently reject the RWE as interesting but unfortunately irrelevant to their decision-making process. These mishaps could be avoided if stakeholder engagement occurs to ascertain evidentiary needs upfront, so that measures and analyses can be built into the studies that specifically target what the stakeholders indicate they are interested in seeing.

We are trying to get companies to retain that same well-tested process for clinical research in the real-world realm. Unfortunately, this is more complicated than you might expect because instead of having simply one stakeholder to engage with, there is a multiplicity. You have regulatory, payers, patients, physicians and health technology assessment bodies. On top of that, and this is something I particularly highlighted in the PPMA presentation, each of these stakeholders has a different perspective and evidentiary need. Another layer of complexity comes in when it is time to design the studies; it is important to recognize that different real-world study designs or approaches are more or less compatible with different kinds of study measures. The whole thing starts to get exponential in complexity. This is okay, but it is important to recognize that whereas you might have done two big pivotal trials in Phase III to achieve approval, once you reach Phase IV there is not going to be a one-size-fits-all approach. You are going to have to develop a program of evidence generation that meets the differential needs of different stakeholders – and you must prioritize along the way because you know there is no such thing as an infinite budget.

Do you think trust in RWE is increasing?

Absolutely – it cannot be underestimated just how impactful the new-found interest of the FDA and other regulatory bodies in RWD is. This is having a tremendous influence even though they are approaching it from a different angle, looking at uses of RWE specifically for regulatory decision making. Of course, in some areas, the regulatory bodies are already excellent at using RWE, such as ongoing safety surveillance. But now they are pushing the envelope.

“…it cannot be underestimated just how impactful the new-found interest of the FDA and other regulatory bodies in RWD is.”

For example, in oncology or rare disease studies, it can be difficult to randomize to a placebo due to ethical or feasibility considerations. As a result, the regulatory bodies are now permitting single-arm trials with controls recruited or identified from RWD sources. They are also getting to what could be the holy grail for manufacturers – enabling label extensions to be based on RWE studies. The first one that I am aware of, in the post-21st Century Cures Act era, is Pfizer obtaining a label extension for Ibrance® – their breast cancer product. All of their original clinical trials were conducted in women and the product has been on the market with some off-label usage in men, in whom breast cancer is a rare disease. Pfizer tapped into a few RWD sources to assess the product in men and submitted the data to the FDA. This resulted in the label expansion. This is a huge development in terms of the credibility of RWD.

Furthermore, we are seeing methodological advances that will help address issues of trust. Biostatisticians and methodologists are helping address some of the major concerns that have always existed in RWD, particularly bias and confounding. Randomization levels that playing field, but you cannot randomize a care decision made 6 months ago: so, what can you do? Methods in this area are getting more sophisticated and I also feel that machine learning and artificial intelligence are going to have a mind-blowing impact, in terms of gaining insight from RWD sources.

 “…the regulatory bodies are now permitting single-arm trials with controls recruited or identified from RWD sources.”

Finally, in your Issue Panel at ISPOR 2019, you discussed the potential problems surrounding attempting to reproduce randomized control trial outcomes using RWD. Did you have any further comments on this?

The Randomized Controlled Trial Duplicate initiative is an ongoing attempt to replicate approximately 30 traditional clinical trials utilizing RWD sources. I think these efforts are interesting and useful, but problematic on a fundamental level – and potentially counterproductive if the outcomes are used to judge the quality of RWE and the underlying RWD sources. That seems to be FDA’s intent and that has me concerned.

For as long as we have been conducting analyses of RWD sources, we have recognized that products perform differently in clinical trials versus clinical practice; this is a phenomenon that has come to be known as the “efficacy–effectiveness gap”. Why does the gap exist? That is an interesting thing to investigate and the factors that often arise are differences in patients, providers and the care patterns that they generate. A big issue is patient compliance with therapy; this tends to be strictly enforced in trials but not in practice. Obviously, if patients do not take their medicines according to doctor’s orders, they will not get full treatment benefit. Putting all this together, we must recognize that it is highly unlikely that analyses of RWD will match well with those of traditional clinical trials; therefore, ultimately, where does that leave us? My hope is that discrepant findings will not be used to impugn the quality of RWE, but I feel that that is the risk.

“I think these efforts [to replicate traditional clinical trial data with RWD sources] are interesting and useful, but problematic on a fundamental level – and potentially counterproductive if the outcomes are used to judge the quality of RWE and the underlying RWD sources.”

How do you see the management of RWE generation evolving in pharma over the next few years?

My observation is that the RWE function within pharma is almost a paradoxical combination of a turf battle and hot potato. It is unfortunately the case that there are a lot of different departments within the typical pharma company that have some jurisdiction on RWE generation: the health economics and outcomes research department, the epidemiology department, safety surveillance, clinical, medical affairs and more. To some extent, they all administer the development of information that falls under the RWE umbrella and so there are inherent turf battles that must be navigated.

There seem to be two parallel trends afoot in terms of companies trying to reconcile this. The first is that we are seeing many companies setting up RWE hubs. These are the companies that tend to be in-licensing a lot of the RWD sources and have found that if they do not centralize it, they will be double paying, with different internal constituents licensing the same data. Therefore, they have addressed the issue by centralizing – with a department that is responsible for helping all the different therapeutic areas within their portfolio address their RWE needs.

However, this approach tends to be limited to retrospective data sources, such as claims data and electronic health records. When it comes to prospective RWD collection, including, for example, data derived from wearables, smartphone apps and so forth, the second trend we are seeing is an increasingly fortified medical affairs function that is taking on the responsibility of RWE generation and becoming the internal authority on how companies go about doing it. Those are the two interesting trends afoot, but every company is doing it differently. Sometimes, no matter how you try to structure or restructure things, there will be tensions that cannot be eliminated entirely. I see this as requiring further development over the near term.


The Evidence Base

Community, Future Medicine

The Evidence Base is a community site covering the latest news, opinion and insight into the collection and application of real-world data to real-world problems.
40 Contributions
0 Following

No comments yet.