Sanofi’s Ameet Nathwani Brings Life Sciences R&D Expertise To The Digital Space
Executive Summary
Nathwani talks to Scrip about his unique dual role as chief medical officer and chief digital officer, Sanofi’s approach to digital therapeutics, and major hurdles to turning the promise of such technologies into a reality for patients.
Sanofi's Ameet Nathwani readily acknowledges he’s not your typical chief digital officer.
Nevertheless, the self-described “technology geek” is excited about being able to meld his life sciences R&D expertise – with its focus on scientific rigor and patient protections – into Sanofi’s growing endeavors in the digital space.
A cardiologist by training with more than 20 years’ experience in the pharmaceutical industry, Nathwani joined Sanofi in May 2016 as chief medical officer. He added a second title in February, chief digital officer, tasked with scaling up Sanofi's portfolio of digital initiatives by developing broad external partnerships, building out internal infrastructures and exploring new business opportunities.
“I’m not a typical phenotype for a CDO. Even in our industry right now the CDOs generally come from a stronger technology background, much more of the data and engineering kind of arena.” – Sanofi’s Ameet Nathwani
During the Biotechnology Innovation Organization’s recent annual meeting in Philadelphia, Nathwani sat down with Scrip to talk about his unique job title, Sanofi’s approach to the digital therapeutic space, and what he sees as the major hurdles to turning the promise of digital technologies into a reality for patients. The conversation took place shortly before Sanofi unveiled its broad partnership with Google and offers insights into the company's digital philosophy (see sidebar).
[This interview has been edited for length and clarity.]
We felt that it was such a wonderful opportunity to sort of blend what actually is our business, which is life sciences, with the digital health ecosystem that’s really accelerating. And so our focus has been very much, and my focus has been for the last now two months, how do we set up a strategy that doesn’t just make us more productive [and] efficient about digitizing what we do – every company is doing that – but looking to see whether or not we can actually capitalize on the technology in the digital health revolution, to say look can we create new revenue streams? Can we look at patient outcomes very differently? Can we create a portfolio of new opportunities, such as technology-enabled drugs or digital therapeutics that can also drive the top line and that can almost help us to transform some of our business opportunities as well?
That’s been the focus of the strategy and that’s also [the reason behind] the kind of combined role of having the chief medical officer, who will bring the life sciences R&D background and the rigor of scientific principles, into digital health, where we still want to make sure we kind of secure patient integrity, safety, take the regulated pathway, look at patient access, understand physicians’ needs and actually how that’s evolving, and bring it together.
But you know one of the key things that we have to keep in mind and what we had an internal discussion on is what is it that we want to achieve? What’s our focus going to be? And if you see the amount of technology companies entering health and also how technology is changing what we do, which is to discover new innovative medicines, we felt two things.
Number one, can we enable technology to do what we are good at doing, which is innovation and making drugs, and do that more efficiently and in more novel ways? … So it’s the use of artificial intelligence in R&D, the way we digitize clinical trials, the way we leverage real-world evidence and how we kind of work with the regulations and regulators to actually optimize what we do. … Also customer engagement, so the way physicians work.
[However, we also] felt that we needed rigor as you would do for biological science in digital science. And as a consequence having someone who comes from a life sciences background, it just makes sense. I take an R&D approach to what we do in digital. We do a proof of concept, we make sure that results are reproducible. Even things like a trial design on a digital therapeutic requires a different way of thinking. It requires us to understand well what is a digital placebo? How as data evolves would the regulators look at software? … Even things like AI in diagnostics or digital biomarkers requires us to understand the regulated pathways and understand the access hurdles ... so actually having someone who is a bit of a technology geek but has also got a life sciences background just makes sense to us.
The first one is can we take a technology and adjunct it with our drugs? We call that Drugs Plus. It’s basically a combination digital therapeutic … and for us this is either a software, a hardware or a service, or any one of the three combined, which can improve adherence, efficacy, safety, the patient experience along the way.
We have a number of connected devices that we’re looking at for insulin delivery. We have something called My Dose Coach that we use to actually give specific algorithms and recommendations which are individualized in the way they would use insulin. We’ve just created and are launching right now a patient assistance app … for atopic dermatitis which does several things. It looks at your itch, your sleep, you can take photographs of your lesions [and] send that off your physician. [It also will look at triggers] for your atopic dermatitis to see whether or not there is either a seasonal element or a contact element to it as well, and all of this can then personalize your approach in how you manage it.
The other piece is digital therapeutics which stand alone. These are basically nonpharmacological therapeutics.
The reason that we like those is because there are many disease areas … where the pharmacology or the biology is not as clearly understood, where you can through some of the technologies, especially behavioral diseases or neurological diseases, address them with a digital approach because you’re basically stimulating the neurons to do things. I think those approaches are very attractive for us because the cycle times of development are very short, they can be up to … one and a half years, and the costs are completely different to normal development costs, probably in the range of $15m-$20m. So you can get a digital therapeutic reasonably rapidly and now there are routes through the FDA to actually get that and do that.
[An example of Sanofi’s work in the standalone digital therapeutics space is a collaboration on depression with Happyify Health, which has developed a digital therapeutics platform aimed at building healthier habits and improving mental health.]
Imagine depression as a meta-condition. So many people with chronic illnesses have depression and actually those with depression tend to have a worse outcome on their underlying diseases anyway than those without depression. And the question is if you could treat depression, in a nonpharmacological way, but with a digital therapeutic, number one is that better, can you get a better outcome? And secondly can you also modulate some of the underlying diseases? Will your blood sugar be better? Will you lose more weight? … If you have MS and depression, will your fatigue be improved as an underlying condition? So we’re actually looking at that as a pilot experiment right now but with a view that actually a prescribable digital therapeutic for depression would be quite attractive, as an example.
[Sanofi’s third area of focus involves taking] population health and look at chronic disease management by merging technologies, telehealth, wearables, to actually manage chronic diseases.
[An example of Sanofi’s efforts in this space is its Onduo diabetes-focused joint venture with Google’s health care spinout Verily Life Sciences LLC. [See Deal]]
Through a combination of algorithms, nanoengineering which has small continuous glucose monitors linked into a loop with coaching, with things that can actually help you to diet, with patient nudges, with AI-driven coaching influence back to the patients which are customized … can we manage their blood glucose?
Now the early data coming out of this actually looks extremely compelling. We’re seeing that the very chronic, very poorly managed patients are getting extremely good glucose control, and the CGM component really helps to drive the control, the AI and the coaching is contributing a significant amount to it. And you know if we can do that at scale we can make that modular. You can imagine an approach where the very complicated multi-morbid diabetic or cardiovascular patient can have many of these things managed at a distance, and at a very different cost level, and it’s kind of agnostic of our drugs.
The major hurdles to be fair are how you design the right trials. What are the right control groups, just as an example. Can you have a digital control group or do you have to use standard of care, and then how do you do it? What’s the level of evidence? Effectively, what is the right outcome? Because clearly your drugs when you go into a clinic, you get sporadic data on which clinicians make decisions, and you’ve got a very different density of data now coming through from technologies. You can get data everyday all the time downloaded.
I think the regulatory route is very similar to the normal drug route, just the level of evidence, its robustness, the rigor of it. Is the scientific methodology correct and is it rational? Is there a logic that can be applied? So I don’t think the regulatory hurdles are as onerous as we would believe.
But then after the registration, the long-term surveillance, the monitoring of continual effectiveness especially starts to evolve or change, what’s that going to look like? What’s the IP landscape, because IP around algorithms is not clear today. We know that you can probably copyright it, but you can’t really IP it. … What is a generic copy of a digital therapeutic becomes more complicated as a question. … Those questions still need to be worked through.
[Then there is] the access hurdle. So what is the payer going to look for? … You don’t have a big factory to develop it, so what’s the right pricing model? What’s an appropriate outcome that they’d be willing to pay for and how does that compare to standard of care, and how differentiated is that going to be?
And then the acceptability by both physicians and patients. I think this issue of trust is a big thing. The areas of privacy will continue to be challenged because in the new world you can imagine a digital therapeutic company by definition will have to have much more intimacy on the patient data than we as a pharma company would have to go through a physician.
A digital therapeutic probably has direct access to the data and the nudges are very specific back to the patient. So how does that work from the data privacy perspective long term? What’s the trust issue that we need to ensure and what if something goes wrong? If there’s a mistake on a disease and an algorithm misdirects a treatment and harm occurs, who’s at fault? How does it get allocated? It’s kind of getting into the world of autonomous cars, which is having similar issues today. It’s how good do you have to be, and how safe does it have to be? So I think that there are a lot of still unanswered questions around that whole space.