0:00:01 – Hi, everyone.
0:00:03 Welcome to the A6NZ podcast.
0:00:04 I’m Sonal.
0:00:06 Today we have one of our reruns,
0:00:09 which was recorded during the JPMorgan Healthcare Conference
0:00:12 last year, where the A6NZ bioteam had a lot
0:00:14 and has a lot going on this year as well.
0:00:18 And it’s with Vas Naurasimhan, the CEO of Novartis,
0:00:19 one of the largest healthcare
0:00:22 and pharmaceutical companies in the world.
0:00:23 In terms of volume,
0:00:25 they’re the largest producer of medicines
0:00:27 with 70 billion doses a year
0:00:29 across a wide range of therapeutic areas
0:00:32 from cancer to cardiovascular disease and more.
0:00:34 Joining me to interview him
0:00:37 are A6NZ general partners Jorge Conde and Vijay Pandey
0:00:39 from the A6NZ bioteam.
0:00:42 And we cover the latest trends and therapeutics,
0:00:45 including the journey in chemistry and medicine
0:00:48 from large molecules and antibodies and proteins
0:00:51 to small molecules and other new modalities with RNA
0:00:55 and now moving more into the cell and gene engineered world.
0:00:58 We also cover when science becomes engineering
0:01:00 and what does that mean at an industry
0:01:02 and a big company innovation level?
0:01:05 And then we touch on topics such as clinical trials,
0:01:09 healthcare go to market, shifts in talent in the landscape
0:01:11 and startups working with Big Pharma.
0:01:14 But we begin with the business of science
0:01:17 with R&D and innovation both inside and out.
0:01:21 – You build up R&D expertise in our industry
0:01:23 over long periods of time.
0:01:25 If you think about cardiovascular disease,
0:01:26 we’ve been in it 40, 50 years.
0:01:28 When you think about transplant and immunology,
0:01:32 again, 40, 50 years, oncology, 25 years.
0:01:34 So you build up an accumulated expertise.
0:01:37 And really the art of it is to make sure you have a depth
0:01:40 of new medicines to keep filling your pipeline
0:01:42 in each one of those therapeutic areas.
0:01:44 Now there are instances where we find new breakthroughs
0:01:46 in areas we’re not in.
0:01:47 Those you have to really think about,
0:01:50 are you gonna really stay in that area for the long term?
0:01:52 The other element of the story
0:01:54 is when you really have exhausted your pipeline,
0:01:56 we’re not so good as an industry at this,
0:01:58 but you have to also be prepared to exit, I think,
0:02:00 areas where you’re gonna be subscale.
0:02:01 And that’s something we’re working on.
0:02:03 We’ve made a number of exits actually this year
0:02:05 where we just said this is areas
0:02:07 we just can’t sustain longer term.
0:02:08 – Can you give us a little bit more color
0:02:09 on how you make those decisions,
0:02:12 especially as a CEO steering this?
0:02:13 That’s a pretty big, I mean, frankly,
0:02:14 it’s a decision that every big company,
0:02:17 regardless of industry has to think about,
0:02:20 which is essentially what to proactively invest in
0:02:22 and what to proactively opt out of,
0:02:24 which killing things as we call it in the media business
0:02:26 is a pretty hard thing to do.
0:02:27 How do you think about that?
0:02:30 And how do you tease apart the signal from the noise
0:02:31 when you get a lot of inputs,
0:02:32 both internally and externally?
0:02:34 – So we do it currently at two levels.
0:02:37 One is from an overall portfolio standpoint.
0:02:39 We’ve made the decision to really focus
0:02:41 as a medicines company powered
0:02:44 by advanced therapy platforms and data science.
0:02:46 So in order to really make that happen,
0:02:51 we transacted in 2018 around $50 billion of deals
0:02:53 to really change the shape of the company.
0:02:56 We took principle decisions to leave consumer healthcare
0:02:58 ’cause we just didn’t believe we would be a long-term
0:03:00 leader in consumer healthcare.
0:03:02 A decision to spin our Alcon business,
0:03:05 which is to get out of medical devices and contact lenses.
0:03:08 And alongside that, as we moved out of those other areas,
0:03:11 we made significant investments in acquisitions
0:03:14 in this next wave of therapy, cell therapy,
0:03:18 gene therapy in an area called radio drug conjugates,
0:03:21 which is a nuclear medicine kind of area.
0:03:24 So that was at one level, at the portfolio level,
0:03:27 changing a 20 plus year trajectory
0:03:29 to actually become a very diversified company.
0:03:31 It really came out of a conviction in my mind
0:03:33 that science is moving so fast.
0:03:35 You have to focus your capital
0:03:37 and really focus your energies.
0:03:38 That’s at the macro level.
0:03:40 Now, when you zoom in to innovative medicines
0:03:42 and we have to decide, okay, which therapeutic area
0:03:44 do we stay in cardiovascular disease
0:03:46 or do we stay in ophthalmology?
0:03:48 I mean, those are pretty tough decisions
0:03:52 because if you take down an R&D effort,
0:03:55 so one example for us was infectious diseases,
0:03:57 where we had a longstanding effort.
0:03:58 It was not an easy decision.
0:03:59 I mean, it was a lot of going around.
0:04:00 Are we really sure?
0:04:03 Because you can’t change your mind now in three, four years
0:04:05 and say, I wish I had it back.
0:04:07 It’ll take you another 10 years to build it back up again.
0:04:12 The cycles of innovation and science are accelerating.
0:04:13 The science is moving much more quickly
0:04:14 than it has in the past.
0:04:16 So assuming that that’s the case,
0:04:17 will it continue to be true
0:04:21 that it will take decades to build up expertise
0:04:23 in any given therapeutic area?
0:04:26 In other words, will there be future emerging players
0:04:29 that come much more quickly than they have historically?
0:04:32 I think there can be very fast players
0:04:34 who are really working on a couple of medicines
0:04:36 or a couple of assets.
0:04:38 But when I talk about building up a capability,
0:04:41 I’m really talking about a scaled capability
0:04:45 that could generate new medicines consistently over time.
0:04:47 And while I do believe that pace of science
0:04:49 is improving dramatically,
0:04:51 we also have to keep reminding ourselves
0:04:52 and being humble with the fact
0:04:54 that we understand a fraction of human biology.
0:04:57 And actually, when you look at attrition rates
0:05:00 in our industry, really the chances of success that we have,
0:05:03 they haven’t moved in the last 15 years.
0:05:05 Still, when we bring a medicine into human beings,
0:05:08 on average, only one out of 20 works.
0:05:09 Finally. – Really?
0:05:11 – And that has stayed constant,
0:05:13 despite the fact we’ve had this explosion in new science.
0:05:14 – I just wanna quickly pause on that for a moment,
0:05:16 because that’s a pretty important point.
0:05:19 So only over the last, say, 10, 20 years,
0:05:23 only one out of 20 medicines actually work in the human body.
0:05:24 – Once we get it into human beings,
0:05:28 we have about a 5% success rate, 5% to 10% success rate.
0:05:30 And it varies by therapeutic area.
0:05:32 We’ve actually been fortunate at our company,
0:05:35 we average in that same metric about 8% to 10%.
0:05:39 But if you look industry-wide, it is about 5%.
0:05:40 – The attrition rates are pretty constant,
0:05:42 but the costs still keep going up too.
0:05:43 – They do.
0:05:44 – How does that work out?
0:05:47 ‘Cause in a sense, one analogy that people use
0:05:50 is almost like trying to get oil out of the ground.
0:05:52 And the low-lying fruit, I’m mixing analogies here,
0:05:54 but the low-lying fruit has been taken,
0:05:57 and it’s just harder and harder to find new therapeutics.
0:05:59 Or do you feel like the science is moving fast enough
0:06:00 that that’s not an issue?
0:06:02 – You know, I think we go through waves.
0:06:03 I think there was a period of time
0:06:06 where probably in the 1990s and early 2000s,
0:06:08 we had a pretty big wave of innovation,
0:06:09 and we could bring a lot of medicines forward.
0:06:12 We went through a lull for seven, eight years.
0:06:14 Now, I think with, again, explosion and ability
0:06:17 to really understand the mechanisms of disease,
0:06:18 we’re seeing a renaissance,
0:06:21 a record number of FDA approvals.
0:06:23 We’re investing heavily in new therapy areas.
0:06:25 I mean, 15 years ago, people would have said,
0:06:27 you’re crazy if you think we’re gonna do gene therapy
0:06:29 and cell therapy and all the things now
0:06:31 that we’re doing at scale.
0:06:32 You know, the costs really come
0:06:34 from our ability to manage complexity.
0:06:36 When you look at it, over time,
0:06:38 the trials get more complex.
0:06:42 The requirements from regulators get more complex.
0:06:43 Because the science gets more complex,
0:06:45 we can actually measure more things.
0:06:48 So we add and add and add and add.
0:06:52 And that’s led to an interesting, pretty linear increase
0:06:55 in cost per patient in our clinical trials.
0:06:56 I don’t think it has to be that way.
0:06:59 I think what really our industry’s not been great at
0:07:01 is really deploying technology
0:07:02 to make this much more efficient.
0:07:04 So I think there’s a lot of opportunity.
0:07:05 – Yeah, well, why do you think that’s the case
0:07:07 that’s been so hard to deploy?
0:07:10 – I think it’s, you know, we’re a high margin industry.
0:07:13 You know, unless it’s easy enough
0:07:14 just to keep arguing to yourself,
0:07:15 it doesn’t really matter.
0:07:18 As long as we get another big medicine out, it’s okay.
0:07:19 Let’s just keep going.
0:07:21 – Well, and if you screw things up,
0:07:22 there’s a huge cost.
0:07:22 – There’s a big downside.
0:07:24 But I think now we’re reaching the point
0:07:25 where we have no choice,
0:07:28 but to really now engage technology.
0:07:31 I mean, there are estimates now from various sources
0:07:33 that believe you could take out 20%
0:07:34 of clinical trials costs
0:07:38 if you were actually to really deploy technology at scale.
0:07:40 – If the attrition rates have been flat
0:07:41 for as long as they’ve been,
0:07:43 and there have been all of these proliferations
0:07:46 of new platforms, cell therapies, gene therapies,
0:07:48 is there a measure that you qualitatively
0:07:50 or even quantitatively can look at that says
0:07:52 that the medicines that are getting through
0:07:55 are meaningfully better medicines or different medicines?
0:07:58 In other words, the failure rate may be the same,
0:08:01 but the impact of success is at greater now
0:08:02 in any measurable way.
0:08:04 – So it’s a very important point.
0:08:07 And there’s no objective measure.
0:08:10 I mean, various institutes have different measures,
0:08:13 so it’s nothing I think that is used externally.
0:08:16 We internally have just set a very clear bar now
0:08:18 for ourselves, primarily because we live in a world
0:08:21 where nobody wants a Me Too medicine
0:08:24 or a medicine that’s just incrementally better.
0:08:25 We say to ourselves,
0:08:27 it has to replace the standard of care.
0:08:31 And that usually means it gives such a big clinical benefit
0:08:34 to patients that it just becomes the de facto medicine
0:08:36 of choice in that therapeutic area.
0:08:37 That’s a shift.
0:08:39 It means a lot of projects no longer make the cut
0:08:41 because you’re really asking yourself,
0:08:43 if I don’t have something really transformative,
0:08:45 I’m not gonna take it forward anymore.
0:08:47 And so all of our research teams and development teams
0:08:49 are having to now come to grips with that,
0:08:52 that we will stop projects unless we really believe
0:08:54 it can redefine the standard of care.
0:08:55 – I have a process question behind this
0:08:57 ’cause it’s a parallel to this idea
0:08:59 of basically going for a slugging average
0:09:02 versus batting average and like outsize hits
0:09:04 with great outsize impact.
0:09:07 So behind the scenes, what are some of the mindsets
0:09:10 that you and the R&D teams bring to bear
0:09:13 to make these investments for slugging versus batting average?
0:09:15 How do you set things up to make that happen?
0:09:18 – We have all of the various review committees
0:09:19 and portfolio meetings, et cetera,
0:09:23 but really what it takes is a lot of discipline
0:09:24 about the criteria that you’re using.
0:09:26 I mean, so we have very clear criteria.
0:09:28 We’ve tried to apply that rigor.
0:09:31 I think people, there’s a lot of romanticization
0:09:33 and R&D about big ideas.
0:09:36 So much of it is just about discipline and discipline.
0:09:37 – Nitty gritty.
0:09:38 – Nitty gritty, disciplined execution
0:09:39 of how you look at projects.
0:09:41 I think that’s one element.
0:09:43 I think second, you have to build patience
0:09:46 because part of the reason mediocre projects go forward
0:09:49 is you start to worry you don’t have enough in the pipeline
0:09:52 and you start to lose faith that something’s gonna come.
0:09:54 And you have to believe in your own scientists
0:09:57 and your own R&D engine to say I’m gonna say no five times
0:10:00 ’cause I believe the sixth one could be the big one
0:10:03 rather than get worried and just start letting things through
0:10:06 because actually what you do is you crowd out the money.
0:10:07 – It’s an opportunity cost.
0:10:09 – It’s a huge opportunity cost when you take those.
0:10:13 And that’s been a real ongoing challenge for us.
0:10:16 I think that the third element is to bring a real lens
0:10:19 of what does it take to be successful in the market.
0:10:21 I think historically we just had a belief
0:10:22 that if we had a great product,
0:10:23 it’ll all work itself out.
0:10:26 Now we actually ask the market access teams
0:10:28 that have to negotiate with payers
0:10:30 to show up at every meeting and say actually,
0:10:33 even in phase two, so really early for us,
0:10:35 what is it really gonna take to let’s say,
0:10:37 bring a new medicine forward in Asmar,
0:10:39 a new medicine forward in multiple sclerosis.
0:10:41 And if we don’t make the cut,
0:10:43 we just have to be brutally honest with ourselves.
0:10:46 – Reimbursement’s more important or more on your mind
0:10:48 than sort of just getting past FDA.
0:10:50 – It used to be we think about reimbursement
0:10:51 as we got to launch.
0:10:54 Now we’re thinking about it really early in development.
0:10:56 – For people that are new to the space
0:10:57 or just a lot of entrepreneurs,
0:11:00 they think that the FDA is the real challenge.
0:11:02 And just getting something to clinical trials
0:11:04 is expensive and hard and that’s true.
0:11:06 But the reimbursement being first in class,
0:11:09 being having this huge jump in care,
0:11:10 that is the real challenge.
0:11:12 And so what I would love to see,
0:11:15 especially in our founders is for them to work backwards.
0:11:17 And but work backwards, not from getting through trials,
0:11:19 but work backwards from reimbursement.
0:11:21 – Yeah, and the way that Voss describes it,
0:11:22 I think is absolutely true,
0:11:26 is a lot of people view reimbursement as a process
0:11:28 to get to market access.
0:11:31 But reimbursement is really just a proxy
0:11:32 for value proposition.
0:11:34 So what are the actual user stories?
0:11:35 Who’s gonna actually value this?
0:11:36 Who’s willing to pay?
0:11:38 It’s almost like a pricing study.
0:11:40 It’s almost like a price discovery in the consumer world.
0:11:43 In this case, it’s obviously the payer’s
0:11:45 not the direct beneficiary of the therapeutic,
0:11:48 but they do bear the burden of the cost.
0:11:49 And so they’re the great arbiter of saying,
0:11:51 is there true value proposition?
0:11:53 And actually that’s why when you talk about moving away,
0:11:55 industry moving away from me too drugs,
0:11:59 it was because a me too drug arguably could not show
0:12:03 a very significant marginal increase in value proposition,
0:12:05 and therefore you could be very difficult to justify
0:12:06 and increase premium price.
0:12:09 And so that historically has been the big challenge.
0:12:10 – On that note, I do find it ironic,
0:12:12 a big part of your business is still generics.
0:12:14 So I mean, what is that but a me too drug?
0:12:17 Like how does that fit into this big picture?
0:12:20 – Yes, if you look overall at Novartis generics,
0:12:23 from a sales standpoint and value standpoint,
0:12:26 is a small portion of the company,
0:12:28 but you look at a volume standpoint,
0:12:29 it’s the biggest part of access.
0:12:32 And so really what a generics, our generics business does
0:12:35 is take when medicines go off patent,
0:12:36 we then produce them at scale.
0:12:38 And we were the largest producer, for example,
0:12:40 of penicillins in the world.
0:12:42 I mean, so we have a huge role to play
0:12:45 in providing access to medicines around the world.
0:12:47 I mean, right now Novartis reaches
0:12:50 about a billion patients a year through our work.
0:12:52 And a lot of that is through our Sandoz generics unit.
0:12:54 – So if you break it down, so if you,
0:12:57 there’s 70 billion doses that are Novartis drugs
0:13:00 every year, how many of those 70 billion are generics?
0:13:03 – I would say roughly 80%.
0:13:06 – Is it standard that big pharmaceutical companies
0:13:08 have their own manufacturing facilities?
0:13:10 And do you see that changing anytime in the near future?
0:13:14 – Most pharmaceutical companies have their own manufacturing.
0:13:15 I mean, there’s different trends right now.
0:13:19 There’s a pretty significant increase of use of Chinese
0:13:23 and other producers from many elements of the manufacturing,
0:13:26 but still historically we’ve had our manufacturing facilities.
0:13:27 The biggest trend we have right now
0:13:30 is a shift to these advanced therapy platforms.
0:13:33 So what we’re having to do is as our volumes go down
0:13:36 and kind of the older medicines that were produced
0:13:39 and huge volumes in the innovative medicines,
0:13:41 we’re now building up cell and gene therapy
0:13:43 production facilities around the world.
0:13:45 So that’s a shift we’re seeing.
0:13:49 – You talk about Novartis becoming a medicines company
0:13:52 using data science and novel platforms.
0:13:54 You’re very specific about saying medicines.
0:13:57 Are medicines and therapeutics synonyms
0:13:59 in the Novartis mindset?
0:14:00 – I would say yes.
0:14:01 There’s of course a gray zone here.
0:14:03 So what is a therapeutic?
0:14:07 We would say medicines is our proxy for therapeutics.
0:14:10 I mean, one example we launched in the US,
0:14:12 a digital medicine.
0:14:14 I mean, with paratherapeutics,
0:14:17 this is the first digital app with an FDA label
0:14:19 that’s being used for opioid addiction
0:14:22 and other psychiatric illnesses.
0:14:26 And it is literally an app that has run clinical trials
0:14:28 and has gotten an FDA approved label.
0:14:31 So that’s truly an example of a therapeutic.
0:14:34 But I would put that within our world of medicines.
0:14:35 – So software is a drug.
0:14:36 – Yeah, software is a drug.
0:14:39 – Most surprising indication that you would expect
0:14:41 to see for a digital therapeutic.
0:14:45 ‘Cause I think most people assume that it’s gonna be around
0:14:48 behavioral health issues or addiction
0:14:50 like with the work you’ve done with Pear.
0:14:52 Can you imagine moving beyond that
0:14:55 from an indication standpoint for digital therapeutic?
0:14:57 – I mean, my hope would be we could develop one
0:14:58 for obesity, right?
0:15:00 That somehow that a digital therapeutic,
0:15:02 they could actually just move the needle
0:15:04 a little bit more on obesity.
0:15:06 It’s such a massive issue for society.
0:15:10 And it should be one where a behavioral intervention
0:15:13 on top of other interventions
0:15:14 could actually move the needle.
0:15:16 Because so much of it is behavioral.
0:15:19 – I mean, there’s not an example
0:15:21 that’s non-behavioral in your future.
0:15:24 – You’re not curing sickle cell with an app.
0:15:26 – I mean, I would put a guess around fertility,
0:15:28 but one could argue that’s also psychosomatic.
0:15:29 – Well, I mean, the things that actually,
0:15:33 so you think about like the modern medical sort of marvels,
0:15:34 I think about like an antibiotic.
0:15:35 Like I was sick when I was in college
0:15:37 and I had a super high fever.
0:15:40 I got an antibiotic and like next few days I’m fine.
0:15:42 Maybe without that, I’d been dead.
0:15:43 And so that’s kind of magical.
0:15:44 And it’s not like I have to take antibiotics
0:15:46 for the rest of my life or whatever like that.
0:15:47 I’m just cured.
0:15:50 But the amazing thing about behavioral
0:15:52 is that that’s where you don’t have this.
0:15:54 I can’t imagine that you have a molecule
0:15:55 that cures depression.
0:15:57 You take that and then you’re just done.
0:15:59 Or you take one, a couple of doses
0:16:01 and then you’re no longer up type two diabetes.
0:16:02 And behavioral is really broad.
0:16:05 It’s depression, it’s smoking sensation.
0:16:06 It’s type two diabetes.
0:16:08 It’s even quite possibly Alzheimer’s.
0:16:09 I don’t know if you’ve seen like all these.
0:16:11 – I’ve seen a lot of recent papers on this.
0:16:12 It’s fascinating.
0:16:13 – And so these are actually the areas
0:16:16 where if you look at the biology of Alzheimer’s disease,
0:16:18 that’s just a mess.
0:16:19 So it could be that for these things
0:16:21 where you have a very clear target,
0:16:23 I just have to hit the ribosome of the bacteria
0:16:23 and then we’re done.
0:16:24 That’s easy.
0:16:26 But there may be actually the future of things
0:16:28 where just it’s hard to hit with a molecule.
0:16:30 And all that is primarily behavioral.
0:16:31 – Interesting.
0:16:32 So basically you’re almost arguing
0:16:33 the question might be moot
0:16:35 because all of disease is behavioral
0:16:35 and some capacity.
0:16:37 – Well, no, all the stuff that’s hard.
0:16:39 – The complex is complex.
0:16:44 – The low-lying fruit molecularly is not behavioral.
0:16:46 – There’s this infrastructure layer
0:16:48 that’s being created now around gene therapies.
0:16:50 So as folks figure out manufacturing
0:16:52 as people think about delivery,
0:16:55 as people think about all of the various components
0:16:56 of modular aspects,
0:16:59 do you think those are things that necessarily
0:17:02 would be owned by one company
0:17:05 or these horizontal infrastructure layers
0:17:08 that a third party should develop
0:17:09 and sort of deploy across the industry?
0:17:10 How do you think this plays out?
0:17:13 In other words, is there a startup
0:17:14 that figures out AAVs?
0:17:17 Do they sort of supply AAV to the industry
0:17:20 or do they go and develop their own gene therapy?
0:17:22 – It’s a very timely question.
0:17:23 We don’t know the answer yet.
0:17:28 I think right now in this nascent phase that we’re in,
0:17:29 we believe we need to just own it
0:17:31 because the launches are so important
0:17:34 that we can’t afford there to be a lot of experimentation
0:17:37 and not really owning the supply chain.
0:17:39 We’ve done $15 billion of acquisitions
0:17:40 just last year in the space,
0:17:43 not including all of our internal work
0:17:44 in each of these areas.
0:17:48 So we’ve chosen to build out the infrastructure ourselves.
0:17:50 I think as the technology matures,
0:17:51 we’ll get more comfortable
0:17:53 about which areas we could send out.
0:17:56 I also think the entrepreneurial world
0:17:58 will also figure out where they can play a role.
0:18:01 I think that’s still all being figured out right now.
0:18:03 And I actually don’t have a view yet.
0:18:07 I don’t know what’s gonna be the elements we must own
0:18:09 and what are the elements that we could afford to give
0:18:10 to other parties?
0:18:11 – You know, on that note,
0:18:13 I’d love to hear from you more about
0:18:16 how you figured out the build versus buy piece then
0:18:19 because a big part of your work is, you know,
0:18:20 focus on innovative medicines.
0:18:21 And you made this argument that it takes 10 years
0:18:24 to build up a base even longer, 20, 30 years.
0:18:27 And yet you’re also acquiring the expertise
0:18:29 for the very new cutting edge things,
0:18:31 which almost makes it seem like you don’t,
0:18:32 seem like you don’t have to even bother
0:18:34 building up that base, why not just acquire it?
0:18:35 So how do you sort of navigate
0:18:37 the build versus buy part of this?
0:18:40 – I think when you wanna enter very new areas,
0:18:42 sometimes it’s prudent to ask yourself,
0:18:45 does somebody have this much more figured out than you do?
0:18:47 So if you take the example of gene therapy,
0:18:49 we acquired a company called AVEXIS,
0:18:51 though it really is, I think the front leading edge
0:18:53 gene therapy company.
0:18:56 Now the scientists at AVEXIS, you know,
0:18:58 they’ve been working at this actually
0:19:00 in their academic labs for 25 years.
0:19:02 I mean, they’ve been working on trying to hone
0:19:06 how to use AAV vectors to get to the neuromuscular system
0:19:09 of children to address these issues.
0:19:11 They’d actually figured out the manufacturing.
0:19:13 They built the manufacturing site.
0:19:15 We were working on gene therapies ourselves in-house,
0:19:17 but when we looked at that, we said,
0:19:20 this is an opportunity to really accelerate what we’re doing.
0:19:24 And so it made sense, I think, to go external.
0:19:26 There’s always that balance.
0:19:28 You know, we are a company that’s very focused
0:19:29 internally on research.
0:19:32 We consistently invest at the high end on internal R&Ds
0:19:35 simply because we believe that’s the heart of the company.
0:19:38 But what I’m trying to keep asking our people is,
0:19:40 if there’s somebody out there who’s got it better than us,
0:19:43 let’s just go get that and build off of it.
0:19:46 I love that, but there is a classic NIH
0:19:47 non-invented here syndrome.
0:19:50 And when you have a strong internal R&D culture,
0:19:53 it does compete with NIH a lot.
0:19:55 So the question that really begs is how you then,
0:19:57 with all these amazing acquisitions,
0:20:00 integrate them into the company and actually make sure
0:20:03 the classic Chesbro study of all these acquisitions
0:20:04 not being killed by the big company,
0:20:06 like how do you balance that piece?
0:20:08 So I think there’s two things I’d say.
0:20:10 One is, as a R&D person,
0:20:13 I have a sort of the ability to really get in there
0:20:16 and have the discussions directly with the scientists
0:20:19 and argue why we need to actually go external
0:20:23 and really evaluate the case with hopefully objective eyes.
0:20:24 The other thing we’ve decided to do,
0:20:26 at least with these very new tech,
0:20:27 three new technology platforms,
0:20:29 is leave them as independent units
0:20:31 and really let them grow up independent
0:20:35 from the big R&D and manufacturing machine.
0:20:37 Because I think exactly for that concern,
0:20:40 it makes sense to let them build up
0:20:43 and really incubate these new technologies,
0:20:44 get them all sorted out,
0:20:45 and then we can ask the question,
0:20:47 what’s the right setup down the line?
0:20:48 – Right, right, right.
0:20:49 That is what the classic studies show,
0:20:51 that that sort of is the way to do the success.
0:20:52 I did, by the way, find it very fascinating
0:20:56 ’cause I wasn’t aware that you have a scientific background.
0:21:00 It reminds me of this idea that we have around CTO-led,
0:21:03 really having technical people at the helm.
0:21:04 So I am curious about your view,
0:21:06 I mean, besides being able to talk to the internal scientists,
0:21:09 like how has that affected your own career
0:21:11 and trajectory at Novartis so far?
0:21:13 – Given the company’s heart is innovative medicine
0:21:16 and most of my background has been in drug development
0:21:17 and really developing vaccines
0:21:20 and then developing various medicines.
0:21:22 I think it gives me a really good insight
0:21:25 into the heart of the company, our key technology.
0:21:27 If you think about our pipeline today,
0:21:29 I know every asset, every clinical trial,
0:21:31 I know all the clinical trial endpoints.
0:21:34 So that, I think, gives you a certain insight
0:21:36 into where the company is heading.
0:21:39 And also, I think enables you to hopefully guide the company
0:21:43 into the right areas in the future.
0:21:44 I think it’d be self-serving to say that it’s better
0:21:48 to have an MD, R&D person running companies.
0:21:51 But I think it does give you a different perspective
0:21:53 on an R&D industry like ours.
0:21:55 – Right, it might even be able to help,
0:21:57 to be able to empathize when you are killing a project
0:21:59 that you actually know what it’s like to feel that.
0:22:00 – Well, that’s for sure.
0:22:02 – Okay, so on that note,
0:22:03 what are some of the most interesting
0:22:06 and most innovative medicines categories?
0:22:08 – You know, when you look broadly right now,
0:22:13 I think you’re seeing a few big areas of high innovation.
0:22:16 I mean, I think in the whole world of CAR-T,
0:22:18 so cell-based therapies,
0:22:21 really what this is is the harnessing the power
0:22:23 to take cells out of the human body,
0:22:26 reprogram those cells and put them back in the human body.
0:22:29 CAR-T is the way we do that in cancer,
0:22:30 but there’s certainly the opportunity to do that
0:22:32 in many other diseases.
0:22:33 There’s companies working on trying
0:22:35 to cure sickle cell disease,
0:22:38 others working on other inherited disorders.
0:22:41 So really reprogramming cells.
0:22:43 – So if you go back 10 years ago,
0:22:46 something like a CAR-T therapy would have seemed
0:22:49 science fiction-y, or at least maybe 20 years ago.
0:22:51 If we look forward 10 to 20 years,
0:22:55 what are the modalities of the future, do you think?
0:22:57 – I think a couple of things will likely come.
0:22:59 I think xenotransplantation, I mean,
0:23:01 which has been in and out and worked on,
0:23:03 and what’s interesting is every one of these
0:23:04 comes up and down.
0:23:08 So gene therapies, cell therapies popped up in the ’90s,
0:23:10 kind of went away, popped up in the 2000s,
0:23:11 kind of went away.
0:23:14 And then the key linchpin issues were solved,
0:23:16 and then it was, you know,
0:23:18 I mean, xenotransplantation where you were able
0:23:21 to make organs for transplantation in animals
0:23:24 that enable then to have a sufficient number
0:23:26 of transplantable organs for human beings.
0:23:28 I think we’re gonna probably get there
0:23:30 in the next 10 to 20 years. – Interesting.
0:23:33 – So regenerative medicine makes a real comeback.
0:23:36 – I mean, I think, well, I think another area, yes,
0:23:38 I think on xenotransplantation being one,
0:23:39 I think the other is gonna be,
0:23:43 we are gonna start to solve problems of regenerating tissue.
0:23:46 We already see examples where we, in our own labs,
0:23:48 where we can start to crack,
0:23:50 how can you regenerate cartilage,
0:23:53 or how can you regenerate other tissues in the body?
0:23:55 Which would, again, seem like science fiction,
0:23:58 but I think actually harnessing the pathways
0:24:00 to really get regeneration to happen,
0:24:03 which would help healthy aging is another thing
0:24:05 I think will likely come.
0:24:10 So there’s a lot of things that are still on the way.
0:24:12 – Can you imagine a moment in time
0:24:16 where aging becomes a therapeutic area for pharma companies?
0:24:19 – Yeah, we had actually an aging program,
0:24:21 a small aging program for some time
0:24:24 where we were trying to work on things like sarcopenia,
0:24:28 which is muscle wasting and similar kinds of conditions.
0:24:30 It turns out to be very, very difficult
0:24:32 because, again, multifactorial,
0:24:36 and you probably need a medicine with behavior,
0:24:39 with diet, with exercise,
0:24:41 with all kinds of things to actually help
0:24:43 healthy aging happen.
0:24:44 But like I said, I mean, we continue to focus
0:24:47 on more the pure regenerative parts.
0:24:50 I mean, you think about the whole world of joints
0:24:52 and movement has not really been addressed and cracked.
0:24:55 And so this is an area where we have exploratory programs
0:24:57 to see maybe we could find something.
0:24:59 I mean, if you could regenerate cartilage or tendons
0:25:02 or enable muscle strength incrementally,
0:25:05 you might be able to improve a healthy aging quite a bit.
0:25:06 – Fabulous, why don’t we actually shift
0:25:09 into the innovative medicines set of therapies?
0:25:13 – Another big area, hot area, is in the world of RNAs.
0:25:16 So these are really ways to deliver,
0:25:20 let’s call it genetic instructions into specific cells.
0:25:22 This has been an area that’s been worked on for many years.
0:25:23 It’s always been difficult,
0:25:26 but I think companies are now starting to crack the problem
0:25:29 of delivering RNAs into specific cells
0:25:31 in a highly effective way.
0:25:32 – Can you give me just a concrete example
0:25:34 of how that plays out with like a real disease?
0:25:36 – So there’s a couple of really nice examples now
0:25:38 with RNA interference.
0:25:42 So one that our company is working on is RNA interference
0:25:46 to impact a factor that’s really a big part of heart disease.
0:25:47 It’s called LP little A.
0:25:50 LP little A is actually thought to be one
0:25:53 of the remaining risk factors for heart disease
0:25:54 that have not been addressed.
0:25:55 You know cholesterol,
0:25:58 everybody’s of course addressed cholesterol extremely well,
0:26:01 triglycerides, LP little A is another factor,
0:26:04 but there’s never been a medicine against it.
0:26:08 And it turns out it’s really hard to drug LP little A.
0:26:10 And so the only way to really target it turns out
0:26:12 to be using RNA-based therapies.
0:26:17 These RNA-based therapies are able to block the production
0:26:22 of the gene, a translation of the gene into the protein,
0:26:25 and then actually reduce the LP little A in the blood.
0:26:26 And so this is one example
0:26:28 of how we’re trying to take this into an area
0:26:31 where otherwise you wouldn’t necessarily have a therapeutic
0:26:33 against something that could have a big impact
0:26:35 for patients with prior heart conditions.
0:26:38 – So RNA interference is essentially a mute button
0:26:39 for a gene of interest.
0:26:40 – That’s right, yeah.
0:26:42 – I love that, and that’s a great example.
0:26:44 And by the way, LP little A sounds like a name of a rapper.
0:26:46 (all laughing)
0:26:47 And just remind me really quickly,
0:26:50 obviously I know what I learned about RNA
0:26:53 from like biology class in the sense of proteins,
0:26:55 but can you give us a little bit more distinction
0:27:00 about what’s unique about RNA-based therapeutic modalities?
0:27:01 – Absolutely.
0:27:02 So when you think about the history of our industry,
0:27:06 maybe another way to describe the trend I see that’s happening
0:27:08 is we used to be about chemicals, the small molecules.
0:27:10 So for probably a hundred years,
0:27:14 most of the pharmaceutical companies had their basis
0:27:16 in the chemicals industry.
0:27:18 And so we made these small molecules
0:27:20 that happened to have various effects on the body.
0:27:21 And over a hundred years,
0:27:22 we figured out we could really target
0:27:24 what those chemicals do.
0:27:26 Around the late 1980s,
0:27:28 we realized you could actually make large molecules,
0:27:31 large proteins, and make them be therapeutic.
0:27:34 So this is antibodies and recombinant proteins.
0:27:37 And that led to a whole new renaissance in our industry.
0:27:40 And so over the next 20 years and up to today,
0:27:43 probably still the largest category
0:27:44 is so-called biologic medicines.
0:27:47 These are antibodies and proteins.
0:27:51 What I see happening now is a shift to a next set of modalities
0:27:54 that move beyond small molecules and proteins.
0:27:56 And that is now really touching other elements
0:27:57 of what happens in a cell.
0:28:02 So one is RNAs, which is really the way DNA gets translated
0:28:04 into a protein that goes through an RNA.
0:28:07 So that’s one new modality.
0:28:10 Another modality, both of them really are about
0:28:12 editing DNA in different ways.
0:28:14 One is to take the cells out of the body
0:28:15 and edit the DNA of the cell
0:28:18 or enable the cell to produce something different.
0:28:21 The other is to do it inside the body,
0:28:22 what we call gene therapy.
0:28:25 So we make that distinction as cell therapy and gene therapy.
0:28:28 So cell therapy is ex vivo, gene therapy is ex vivo.
0:28:29 Inside, outside.
0:28:29 Inside, outside.
0:28:34 So these are new ways of actually delivering medicines
0:28:36 or creating medicines in the human body.
0:28:38 And now you see early stage companies
0:28:40 doing even more radical things,
0:28:42 trying to turn red blood cells into therapeutics
0:28:43 amongst other things.
0:28:46 So it’s really an expansion.
0:28:48 Let’s think about, if you think about it of the game board
0:28:50 of how you can address human diseases.
0:28:52 I love sort of the sweeping history you have here
0:28:54 in terms of starting with chemistry
0:28:56 and then moving to large molecules
0:28:59 and then now moving more into the cell engineered world.
0:29:03 Historically, every single sort of drug program
0:29:05 has been a very bespoke thing.
0:29:08 A very sort of, you know, it’s on ground war, right?
0:29:10 You have your target discovery
0:29:11 and then you have your validation
0:29:12 and then you have your lead
0:29:15 and then you optimize that molecule and then so on and so on.
0:29:18 And at least my sense has always been
0:29:20 that because it’s so bespoke
0:29:23 that there are some learnings that are generalizable
0:29:24 in any given disease area
0:29:27 but every sort of program is a unique thing.
0:29:29 When you start to move to the RNA world,
0:29:32 to the cell world, to the gene world,
0:29:35 is it going to become much more of a modular world
0:29:38 where, you know, the first version of a CAR T
0:29:41 is going to be by definition less sophisticated
0:29:41 than the second version.
0:29:43 But the second version will be built off the first.
0:29:47 And you go from being in a bespoke world
0:29:50 to going much more into sort of an iterative world.
0:29:51 – Unfortunately, in our industry,
0:29:53 it’s always the answer is it depends.
0:29:56 I think in the specific example of CAR T,
0:29:57 I do think that’s what’s going to happen
0:30:00 because you have such a complex manufacturing
0:30:02 that you’re going to have the first generation,
0:30:04 let’s say, of a CD19 card,
0:30:08 which is a card that targets B cell cancers.
0:30:10 And you’re going to try to then move into a next generation
0:30:12 that hopefully has more rapid manufacturing,
0:30:13 maybe higher efficacy,
0:30:15 and then even more rapid manufacturing.
0:30:17 So you’re going to get into that iteration.
0:30:19 Now, it’s not like medical device iteration.
0:30:21 I mean, this is still going to take years to do,
0:30:23 but you are going to get to that iteration.
0:30:25 I think another way, what I see happening though,
0:30:29 with these new technologies is real platforms in so far
0:30:33 is once you have the backbone of the production
0:30:36 and even the go-to-market model, depending,
0:30:39 you can put multiple products onto the platform.
0:30:41 What we’ve done at our company
0:30:44 is build a global network of manufacturing sites
0:30:46 that can take cells out of human beings
0:30:48 and reprogram the cells and put them back in the body.
0:30:51 And we’ve built the links into hospitals
0:30:53 to enable us to do that.
0:30:55 So you have that as a capability.
0:30:57 You also have the capability to understand
0:31:01 how to use what’s called a Lenti virus to reprogram a cell.
0:31:03 So we’ve got all of that.
0:31:05 Now we can apply that in very different ways,
0:31:08 in cancer and sickle cell disease and inherited disorders,
0:31:10 and use that same infrastructure
0:31:13 to actually then keep pushing the medicines through.
0:31:15 That’s very different than what we’ve had to do in the past,
0:31:19 where every single medicine had a bespoke production process,
0:31:22 had to have its own manufacturing facility.
0:31:24 Now we can actually build that platform
0:31:25 and then layer medicines on.
0:31:27 It’s no different in gene therapies.
0:31:30 When you think about AAV vectors,
0:31:33 these are ways to deliver these gene therapies into the body.
0:31:35 Once you solve it, the process, let’s say,
0:31:36 for one of these vectors,
0:31:38 you can apply it to multiple different diseases
0:31:41 and not have to recreate everything again.
0:31:45 That’s a shift I see in how our industry operates.
0:31:46 – You know, I find that fascinating
0:31:48 ’cause it actually sounds a lot like what we talk a lot about
0:31:50 around this theme around engineering biology
0:31:52 and when you bring engineering principles
0:31:53 and mindsets to biology.
0:31:56 – You know, you’ve just mentioned multiple places
0:31:59 where there’s sort of repeatability
0:32:02 and sort of different aspects of engineering
0:32:04 have already come in.
0:32:05 How is this trend gonna continue?
0:32:06 Where are there gonna be the new places
0:32:08 where engineering can play a role?
0:32:10 – I think the easiest place is gonna be
0:32:13 in continuing to innovate on the processes
0:32:16 by which we really manipulate cells and gene
0:32:18 and really get to the next wave of manufacturing.
0:32:23 ‘Cause I would say we’re really on the only learning to crawl
0:32:25 with respect to most of these technologies
0:32:28 and how we produce them, pretty rudimentary.
0:32:30 And so I think there’s gonna be an engineering problem
0:32:32 of how do you handle cells
0:32:34 and how do you handle the vectors
0:32:37 and make this a much, much more efficient process.
0:32:38 And there’s a lot of, I think,
0:32:41 very smart engineering firms now working on that space.
0:32:43 So I think that’s one place.
0:32:44 The area I’m quite interested in
0:32:46 is how we can get much smarter
0:32:49 at actually engineering the medicines themselves.
0:32:52 I mean, we spend a lot of work investing in AI
0:32:54 and 3D visualizations to say
0:32:56 in the so-called world of chemical biology
0:32:59 or if you even think about using quantum chemistry
0:33:01 that really understand how to define
0:33:02 your monoclonal antibody.
0:33:06 How can we do a lot more engineering of medicines up front?
0:33:08 Because we really come from a heritage
0:33:10 where everything was just trial and error.
0:33:12 We just tried many, many, many molecules
0:33:15 until we found one that worked and we just took it forward.
0:33:17 How can we become much smarter about that?
0:33:18 And so at our research labs,
0:33:21 we’re spending a lot of time thinking about
0:33:23 how do we engineer the medicine up front
0:33:25 to do what we want it to do.
0:33:27 And that’s a whole new world, I think.
0:33:29 – Yeah, also I think there’s gonna,
0:33:32 presumably have to be a culture that shifts along with this.
0:33:35 I read Alan Greenspan’s book, “The History of Capitalism.”
0:33:38 And he talked about how actually in Europe,
0:33:40 like furniture was bespoke
0:33:41 and you’d make this beautiful chair
0:33:42 and it’s this handicraft.
0:33:46 And they actually hated the idea of factories and engineering
0:33:48 because it takes the art out of it.
0:33:49 – It’s not artisanal anymore.
0:33:50 – Yeah, it’s not artisanal anymore.
0:33:53 But I think once you can have this ability to shift
0:33:56 towards that mindset where you have reproducibility
0:33:59 and almost like a factory process that can be built,
0:34:00 once you can have that shift,
0:34:03 as long as everyone is ready to make that shift,
0:34:05 then things can really start rolling.
0:34:07 But there has to be a major shift.
0:34:08 In terms of like in America,
0:34:11 people really care about that artisanal part as much.
0:34:14 And we got factories and that was a huge part of the early,
0:34:16 like late 1800s.
0:34:18 And I’m curious, you spoke so much
0:34:20 about how the virus is changing.
0:34:23 And so presumably there’s an internal cultural change as well.
0:34:25 – Yeah, we’re making,
0:34:27 trying to make a quantum change, I think, in our culture.
0:34:30 I mean, what we have is as context,
0:34:32 I believe we’ve moved to become truly
0:34:33 just a knowledge organization.
0:34:36 I mean, so much of the rudimentary tasks
0:34:39 have been either automated or sent to third parties.
0:34:42 So we have a whole organization of knowledge workers,
0:34:44 50% of them are millennials.
0:34:46 And they want to work in a very different environment
0:34:49 than let’s say an industrial company 20 years ago.
0:34:53 And so we call our new culture inspired, curious and unbossed.
0:34:56 And we want our people to feel inspired by the work,
0:34:58 really curious about the outside world
0:35:00 and not lived in a bossed company,
0:35:03 but really live in an unbossed, much more empowered company.
0:35:05 And when we talk about areas like digital
0:35:08 and data science, cell and gene therapies,
0:35:11 it’s so critical because these are so complex areas,
0:35:13 you need your people to figure out the answers.
0:35:15 And we can’t be in a world where everybody’s waiting
0:35:17 for management to tell everybody what to do
0:35:19 because none of us know what to do either.
0:35:22 ‘Cause these are whole new spaces for us.
0:35:23 So that’s a big shift.
0:35:25 The other element of that journey
0:35:27 is to get a lot more comfortable with rapid failure.
0:35:30 I mean, we have to be much more rapid cycle.
0:35:32 We can’t expect that we’re going to sort it all out
0:35:33 and it’s all going to work perfectly
0:35:35 because the first thing we’ve learned already
0:35:38 in cell and gene therapy is nothing works
0:35:40 the way you expected it to work, right?
0:35:43 And so you built a platform for rapid iteration.
0:35:44 That’s the idea.
0:35:45 What I love about that is it reminds me
0:35:47 of computing software companies
0:35:50 and the shift from waterfall to like more DevOps, Agile.
0:35:51 Yeah, the same principle.
0:35:52 Even microservices, architecture, enable.
0:35:54 We’re a little late to the party, but yeah, that’s the idea.
0:35:57 It’s the same kind of principle, that’s fascinating.
0:35:59 So we haven’t talked about the big elephants
0:36:01 in a good way in the room of AI and ML,
0:36:04 you know, artificial intelligence and machine learning.
0:36:06 Let’s talk about AI and ML and data.
0:36:09 I mean, it’s not a question of if, when, it’s how.
0:36:10 The question I have,
0:36:13 because quite frankly, it’s a very hype topic too.
0:36:15 And people sort of promise all kinds of things
0:36:19 when they talk about applying AI and ML to medicine.
0:36:21 I’m very curious from your take as a head of Novartis,
0:36:26 like where do you see the strongest applications of AI and ML?
0:36:27 Well, I have to first say, I completely agree
0:36:29 about the hype cycle here.
0:36:31 I mean, as we’ve gotten quite scaled
0:36:33 in working on digital health and data science,
0:36:37 we’ve learned that there’s a lot of talk
0:36:41 and very little in terms of actual delivery of impact.
0:36:42 But we’ve learned a lot.
0:36:44 I think the first thing we’ve learned
0:36:47 is the importance of having outstanding data
0:36:49 to actually base your ML on.
0:36:52 And in our own hands, in our own shop,
0:36:55 we’ve been working on a few big, big projects.
0:36:56 And we’ve had to spend most of the time
0:36:58 just cleaning the data sets
0:36:59 before you can even run the algorithm.
0:37:02 That’s just taken us years just to clean the data sets.
0:37:04 And I think people underestimate
0:37:06 how little clean data there is out there
0:37:08 and how hard it is to clean and link.
0:37:11 It was never intended to have this type of analysis done, right?
0:37:14 It was intended for a given project and that was it.
0:37:16 Yeah, that’s been so much of it.
0:37:20 And then the other thing is there are patterns
0:37:22 that can be really learned from the day.
0:37:25 I mean, do you have a good training data set
0:37:27 to actually train the algorithms?
0:37:28 So there’s a few places I think
0:37:30 we’ve seen a lot of traction.
0:37:32 One, I think the vision or image problem
0:37:34 has been very well solved.
0:37:37 So right now we’re in the process of digitizing
0:37:39 all of our pathology images
0:37:41 and having AI just be able to scan
0:37:44 all of the three pathology images at Novartis.
0:37:46 And we have millions of, of course,
0:37:48 records of biopsies and tissue.
0:37:52 And so that’s a huge project we have called Path AI.
0:37:54 I really work on that as a single example.
0:37:56 I mean, that’s like a gold mine.
0:37:57 It should be.
0:37:57 I mean, it should be.
0:38:00 And if you then apply that as well to the vast stores
0:38:02 of imaging data we have from our clinical trials,
0:38:05 we have two million patients in clinical trials,
0:38:07 at least in the last 10 years.
0:38:11 And we have MRI, CT scans, retinal scans, heart scans,
0:38:13 and all of that as well.
0:38:16 I think ML can have a significant potential
0:38:19 to really find, hopefully, new insights.
0:38:20 So I think the vision image problem
0:38:23 has been one we’ve been able to really take on.
0:38:25 Another area is in our operation.
0:38:27 So we’ve built an operational command center.
0:38:30 Take us, as I said, two and a half years to build it.
0:38:31 We call it SENSE.
0:38:33 And what it enables us to do a team sitting
0:38:35 centrally in our headquarters
0:38:37 to look at all of our clinical trials in the world.
0:38:40 And AI is predicting which trials are gonna enroll on time
0:38:42 or not enroll on time,
0:38:44 predict which ones are gonna have quality issues
0:38:45 or not quality issues.
0:38:47 And the reason we could do that is we had 10 years
0:38:49 of history to train the algorithms.
0:38:53 And we run about 400 to 500 clinical trials a year.
0:38:56 So we have a lot of data that we could train,
0:38:57 train the algorithms.
0:38:59 – Does that mean you’ve had to dig all the way back
0:39:03 into automating sort of real-time information
0:39:05 on clinical trials?
0:39:07 So the data entry on a clinical trial
0:39:08 as a patient is enrolling,
0:39:10 has that all been automated as well?
0:39:11 ‘Cause that used to be done on pads.
0:39:12 – It’s a great question.
0:39:14 I mean, really what we focus on is the operational data.
0:39:16 So one level up from the patient.
0:39:17 So is the trial enrolling on time?
0:39:19 Are the sites open?
0:39:21 All of that, all of those elements?
0:39:25 On the operational side, it was really easier to do this
0:39:27 than trying to get all the way down
0:39:29 to patient level data.
0:39:30 On the other area, interestingly,
0:39:32 in the financial area as well,
0:39:34 we find that AI does a great job
0:39:36 predicting our free cash flow,
0:39:39 predicting a lot of our sales for per key products.
0:39:42 And it does better than our internal people
0:39:43 because it doesn’t have the biases
0:39:45 and the data is very clean
0:39:47 and we’ve got very long-term data.
0:39:48 So that’s been all positive.
0:39:50 But there’ve been other areas where I think
0:39:52 it’s just simply not met up.
0:39:54 I mean, I think that the holy grail
0:39:57 of kind of having unstructured machine learning
0:39:59 go into big clinical data lakes
0:40:01 and then suddenly find new insights,
0:40:04 we’ve not been able to crack mostly because the data,
0:40:05 to link it up.
0:40:09 And I mean, we are spending a lot of our energy
0:40:11 just trying to get all of our data harmonized
0:40:16 so that some algorithm could maybe find anything of use.
0:40:18 – There’s an area that’s desperately in need,
0:40:19 I think of innovation,
0:40:22 is how we think about clinical trials.
0:40:24 Recognizing we have to operate
0:40:26 within the system that we live in.
0:40:30 But if you could design testing safety
0:40:32 and efficacy in humans on a blank sheet of paper,
0:40:33 what would look different
0:40:36 from a clinical trial perspective versus where we are today
0:40:37 and the way we do it now?
0:40:40 – I mean, the ideal world, if we could ever,
0:40:42 if we could get there would be,
0:40:44 we would have integrated health records
0:40:47 where we couldn’t easily insert the fields
0:40:50 that we needed for clinical trials.
0:40:52 And then we could use something like a blockchain
0:40:55 or some other distributed architecture
0:40:56 that enabled patients to consent for us
0:41:01 then to access the data and then run the trials through that.
0:41:04 And that would eliminate so much of the effort
0:41:08 of creating a second database versus the EHR,
0:41:11 monitoring that database, QA’ing that database,
0:41:13 locking that database.
0:41:15 You could get the data on an ongoing basis.
0:41:18 I mean, we would radically simplify this.
0:41:22 I believe that’s a huge, huge opportunity.
0:41:25 I think we have a long way to go because EHRs
0:41:26 are not where they need to be.
0:41:28 We’re probably not where we need to be to get there.
0:41:31 But I see opportunities in baby steps
0:41:32 to actually get towards that.
0:41:35 And I think we’re experimenting with that.
0:41:38 I think other companies are as well.
0:41:39 The other thing people talk about,
0:41:41 but I mean, I’ll take a skeptical voice around it,
0:41:44 is the ability to use real world evidence
0:41:46 to try to get at these things.
0:41:48 But as somebody who’s worked in clinical trials
0:41:52 for most of their time in the industry,
0:41:57 I do believe that the power of randomization,
0:42:00 the power of blindedness is what enables us
0:42:04 to control for all of the things we don’t know
0:42:07 about the complexity of human life and human biology
0:42:09 and to think that we’re gonna take that away
0:42:11 and then be able to really determine
0:42:13 the efficacy of the medicine,
0:42:16 puts a lot on the statistics that I don’t think we have.
0:42:20 And so I’m more of a real world evidence,
0:42:21 I don’t know if it’s a skeptic,
0:42:25 but realist who sort of says after we have
0:42:28 randomized placebo control data that really tells us
0:42:30 that something has the effect we think it is,
0:42:33 then to explore more effects or explore more uses
0:42:35 through real world evidence makes a lot of sense.
0:42:38 But I don’t see this as a panacea
0:42:41 that suddenly will make the world much easier.
0:42:42 I mean, that’s my expectations as well,
0:42:45 is that you’ll see it first come out like as a phase four,
0:42:48 something where you’re using real world evidence,
0:42:50 which was right now used for reimbursement anyways and so on.
0:42:53 And, but then maybe see how far it can go back,
0:42:55 but it’s not gonna replace it.
0:42:57 You guys don’t think a secular, I mean, not to sound naive,
0:42:59 but you don’t think a secular shift,
0:43:01 like censorification of everything
0:43:04 and everyone really truly has continuous wearables,
0:43:07 like everyone’s wearing a CGM by default.
0:43:09 I hear you on the statistical side,
0:43:10 and there’s a lot of other various variables
0:43:13 and things introduced into that equation,
0:43:17 but it is a huge, it’s a very deep, nuanced,
0:43:19 patient level set of data
0:43:20 that seems like we can’t ignore the power of that.
0:43:23 Like where do you fall on that?
0:43:24 – When I think about, first of all,
0:43:26 I would say just in general in sensors is another place
0:43:29 where there’s been a lot of hype above expectations.
0:43:31 I mean, we’ve been really trying to explore the use
0:43:33 of sensors in clinical trials now for,
0:43:35 in my own experience, at least six years.
0:43:37 And it’s been tough to get sensors
0:43:40 that really meet clinical trial grade outcomes.
0:43:44 I mean, to really show that they can be validated
0:43:47 versus our current clinical endpoints.
0:43:50 Now, if it’s consumer products, fine.
0:43:52 I mean, they’re the perfect, perfect people can feel.
0:43:53 – But you’re talking medical grade.
0:43:55 – But here we need to really be able to replace
0:43:57 what are pretty rigorous tests.
0:43:59 And we haven’t seen that, seen that yet.
0:44:02 Now, we’re exploring, I think, use of many different sensors.
0:44:05 The real power of it is a continuous variable
0:44:07 to actually see how a patient’s doing
0:44:09 in between the study visits.
0:44:11 And so I think that will help a lot,
0:44:12 but I still think in the end,
0:44:15 you’re gonna need to randomize and blind.
0:44:18 I mean, I think if you don’t randomize,
0:44:20 I think it’s really hard to figure out
0:44:23 what is going on in a complex system.
0:44:24 – I agree with the short term.
0:44:28 I think longer term, my good feeling is that statistics,
0:44:30 this is a solvable problem statistically,
0:44:33 because there is even issues with clinical trial design
0:44:36 that one has to overcome today,
0:44:39 because randomization isn’t just picking people
0:44:41 literally randomly, you know, necessarily.
0:44:41 – True.
0:44:42 – It’s a sample, not a population.
0:44:43 – And there’s been a lot of work
0:44:46 on causality theory and statistics of around.
0:44:49 So there are advances, but I think it’s not there now.
0:44:50 – Yes, I agree.
0:44:51 Small n, not capital n.
0:44:53 More to say there, that was really interesting.
0:44:55 – What’s the role of bringing innovation in
0:44:59 from the outside through partnerships and M&A and ML, yeah.
0:45:03 – Yeah, I think one of the things we’re working through
0:45:05 is how do we get the talent, you know?
0:45:08 As we really start to organize the data,
0:45:09 and we’ve brought in some great talent
0:45:11 to really help us work on data architecture
0:45:14 and come up with a whole data landscape for the company.
0:45:16 So that we’re always now thinking about
0:45:18 how do we treat data as an asset?
0:45:20 That’s one of the things we keep harping on,
0:45:22 is data as an asset, whatever data we collect
0:45:25 from the external world has to be organized
0:45:27 in a clear data architecture.
0:45:30 But then to take the next step to get the data scientists
0:45:32 to really find the insights,
0:45:35 we’re not the traditional place where a data scientist
0:45:38 coming out of Stanford is looking for
0:45:39 where they want to come to.
0:45:41 So we’re working through partnerships with universities,
0:45:44 potential partnerships with startups.
0:45:45 Actually here in the Bay Area,
0:45:46 we have a center called the Biome,
0:45:48 where we’re working with different startups.
0:45:51 And so these are the things we’re trying to do to engage
0:45:54 and hopefully create an ecosystem that helps us do this
0:45:56 and not just do it ourselves.
0:45:58 I don’t think we’ll be able to track the scale
0:45:59 that you would need.
0:46:00 – Yeah, there’s a Reese’s Peanut Butter Cup issue
0:46:03 because startups sometimes have some innovation
0:46:06 on the data science, but not the data.
0:46:07 And so bringing the two together,
0:46:09 I think seems like a very natural combination.
0:46:11 – Where is Reese’s Peanut Butter Cup?
0:46:12 – Peanut Butter and Chocolate,
0:46:14 like she’s got the peanut butter and the chocolate.
0:46:16 – Oh my God, I’m like, I don’t remember those commercials.
0:46:18 – I don’t remember them.
0:46:19 I watched a lot of TV shows growing up,
0:46:20 but I don’t remember that.
0:46:23 I find it fascinating ’cause a lot of our bio entrepreneurs,
0:46:25 the number one thing that they tell me
0:46:27 that drawing data scientists to bio companies
0:46:30 is one of the hardest challenges they have to face.
0:46:32 And so you’re saying with the biome and other things
0:46:33 that you’re doing, that you’re essentially saying
0:46:35 you have to kind of create the pipeline,
0:46:36 not just source it.
0:46:38 – That’s right, that’s right.
0:46:40 And it really, to what you told your earlier point,
0:46:42 I mean, the opportunity is to say,
0:46:43 look, come and work with us
0:46:45 and we’ll let you work with our data
0:46:47 and you can learn and we’ll learn.
0:46:49 And maybe then there’s a partnership that’s created
0:46:50 or maybe you want to come work for us,
0:46:52 which would also be great.
0:46:53 But that’s how we’re approaching it.
0:46:55 – Well, and there’s actually an interesting shift
0:46:57 that can happen in academia
0:46:58 with my group at Stanford.
0:47:00 Many people actually, during their PhD,
0:47:03 have gone to work in pharma and it’s hard to,
0:47:05 it’s possible to pull the data out of pharma,
0:47:07 but it’s actually easier to put the grad student
0:47:08 into pharma. – Oh, nice.
0:47:11 – And so the grad student comes with the code,
0:47:14 runs it, you know, internal through the firewall of pharma
0:47:14 and we see how it does.
0:47:16 And then you can still publish papers
0:47:19 where maybe you have to obscure what the target is
0:47:19 or something like that,
0:47:21 but you can at least see how things are going.
0:47:24 And there’s nothing like sort of trying in the real world.
0:47:26 – Yeah, yeah, makes total sense.
0:47:28 So on this question of bringing in talent,
0:47:29 as you guys operate globally,
0:47:32 obviously you’re in 150 countries,
0:47:34 some of your headquarter in Switzerland,
0:47:37 Nibers in the Boston, Cambridge area,
0:47:38 you have a presence out here in Silicon Valley.
0:47:41 So how do you guys think about innovation hubs?
0:47:45 Very simplistically is all of the machine learning,
0:47:48 artificial intelligence talent going to be based out here.
0:47:51 What, you know, how do you sort of distribute teams
0:47:52 across the world?
0:47:54 – So it’s interesting.
0:47:55 You know, when you look at research,
0:47:56 we have three main hubs,
0:47:59 our three main hubs are in Cambridge,
0:48:01 in Basel, Switzerland and in Shanghai in China.
0:48:04 Those are three main research hubs.
0:48:06 In terms of development centers for product development,
0:48:09 you would add on to that list Hyderabad, India
0:48:13 as kind of the main, the East Hanover, New Jersey.
0:48:15 But when it comes to data science and digital,
0:48:16 what we’ve actually decided to do
0:48:18 is take a much more distributed approach.
0:48:20 So we’re building up these biome centers
0:48:23 in San Francisco and in London,
0:48:26 other locations in the Middle East, perhaps in China,
0:48:27 just trying to say,
0:48:30 we’re not gonna constrain ourselves
0:48:31 with our current locations.
0:48:33 We’re gonna just try to source talent wherever it is,
0:48:35 particularly because talent in these areas
0:48:38 doesn’t necessarily have to be housed,
0:48:40 you know, next to the other functions.
0:48:42 We’re really asking these people to explore our data
0:48:45 and find big, big new insights.
0:48:48 So that’s the approach we’re taking right now.
0:48:49 It was really saying, you know,
0:48:50 let’s go where the talent is
0:48:54 as opposed to force everyone to come to us.
0:48:57 So we’ll see, that’s the experiment we’re undertaking.
0:48:59 – How do you see the future of that sort of working out?
0:49:00 Like, do you see that, you know,
0:49:02 Boston, Silicon Valley, Basel,
0:49:04 like these places will specialize?
0:49:06 Will they distribute?
0:49:08 – Yeah, we have lots of debates as if we were to build
0:49:11 a scaled hub in digital or in data science health,
0:49:12 where would we go?
0:49:14 I think one of the challenges in the Bay Area
0:49:17 is again, just the competition for talent is so intense,
0:49:19 especially in the tech sector.
0:49:22 So we’re in the business of funding early stage companies,
0:49:24 supporting entrepreneurs.
0:49:26 If I’m an entrepreneur,
0:49:28 I obviously see a ton of benefit
0:49:30 in partnering with Novartis.
0:49:33 Access to data that doesn’t exist elsewhere,
0:49:35 obviously validation in my approach
0:49:37 and my technology, et cetera.
0:49:38 But if I’m an entrepreneur,
0:49:43 I’m also scared to approach a large company like Novartis,
0:49:45 ’cause I’d worry about, you know,
0:49:47 basically you’re an elephant and I’m a mouse
0:49:48 and if I want to dance,
0:49:50 I have to hope you’re a very graceful elephant.
0:49:52 (laughing)
0:49:53 Otherwise, you’re gonna crush me.
0:49:56 What advice would you give to entrepreneurs
0:49:58 about approaching biofarm,
0:50:01 a large biofarm in the spirit of collaboration?
0:50:03 – Yeah, I think in data and digital,
0:50:06 what we’ve tried to do is make us feel a lot smaller.
0:50:09 ‘Cause I think we recognize that we are a huge beast.
0:50:11 And so with things like the biome,
0:50:14 we work with many other entities to try to say,
0:50:16 how can we make ourselves feel smaller,
0:50:18 work in smaller units?
0:50:21 We created our own digital data organization
0:50:25 so that entrepreneurs would have an input into Novartis
0:50:26 where it’s people like them.
0:50:27 I mean, the people in that team
0:50:30 are all come from the tech sector.
0:50:32 They’re working in a much smaller, agile way.
0:50:34 They do sprints and scrums
0:50:36 and they work in all the ways
0:50:39 that the people are being used to working.
0:50:42 And so I would say really engaging through some place
0:50:46 in a large company that I think has a natural affiliation
0:50:48 for the entrepreneur makes a lot of sense.
0:50:49 I think it is harder
0:50:52 on the kind of traditional biomedical side, right?
0:50:55 I mean, we have, I mean, if you just think of,
0:50:57 we have 17,000 R&D people
0:51:01 and spent $9 billion plus a year in R&D.
0:51:02 So if you’re a small entrepreneur
0:51:04 who wants to start working with us,
0:51:07 it’s easy to get lost in the fray.
0:51:08 We’re trying to work on that.
0:51:10 I think most of the companies in our industry
0:51:14 try to have external offices that try to engage.
0:51:16 I mean, we have external scholars program
0:51:19 where we really try to enable scientists
0:51:23 to use our facilities, interact with our scientists.
0:51:24 So we’re trying to experiment,
0:51:26 but I can’t say that we’ve completely figured that out
0:51:27 on the biomedical side.
0:51:30 I’m much more optimistic on the data and digital science side
0:51:33 mostly because we just brought people in from that world
0:51:35 and they just think differently.
0:51:36 – There was something I wanted to ask you earlier
0:51:38 which was about measurement.
0:51:41 ‘Cause when you talked about the portfolio approach,
0:51:44 I wanted to know how you think about actually measuring
0:51:47 the way you make those investments in a portfolio.
0:51:49 And the reason I asked is because there’s all these mindsets
0:51:52 like pastures quadrant, like here’s a place
0:51:54 where we’re gonna put more emphasis on basic research
0:51:55 and we’re gonna put more emphasis
0:51:57 on something more practical.
0:52:00 Or there’s another approach in Xerox PARC.
0:52:03 They used a modified real options analysis
0:52:05 as a way to figure out how to do like short-term,
0:52:07 long-term, mid-term type investments.
0:52:09 Do you have a way of sort of closing the feedback loop
0:52:11 for how you measure the success
0:52:15 of how you’re allocating and deploying investments in R&D?
0:52:16 – Yeah, I mean, we have financial measures.
0:52:20 So we look at return on capital employed, NPV, NPV, peak sale.
0:52:23 So all the traditional financial measures,
0:52:27 we look at really the scientific innovativeness
0:52:28 for lack of a better word.
0:52:30 Is this really something that’s changing the game
0:52:31 from a scientific standpoint?
0:52:33 That’s a little bit more of a subjective measure
0:52:35 but we try to ask teams,
0:52:37 is this really moving the needle
0:52:39 from a standard of care science?
0:52:42 And we actually score that based on six different parameters.
0:52:43 – Oh, interesting.
0:52:44 Are you allowed to share those parameters?
0:52:46 – I don’t know them off the top of my head.
0:52:50 But we really try to score the medicines to say,
0:52:51 is this really transformative?
0:52:53 So you have a financial score,
0:52:55 you have a transformational score.
0:52:58 And then another kind of subjective element
0:52:59 is does this strategically fit?
0:53:02 So is it in one of our core therapeutic areas?
0:53:04 So if somebody comes with a great breakthrough,
0:53:08 which happens not quite often,
0:53:09 in an era that we’re not in,
0:53:11 that’s the toughest one
0:53:12 because it can break through,
0:53:14 but we’re not in this space.
0:53:15 And what do we do now, right?
0:53:17 And do we really want to build this up
0:53:19 or do we want to just send it to an out license
0:53:22 to a fund or do something else?
0:53:24 Those are tough discussions.
0:53:25 But we try to be disciplined
0:53:27 because it’s again, the patience
0:53:30 and being really sure you build depth in your key areas.
0:53:31 Because if you take another program on,
0:53:33 that means that there’s another program you have to stop.
0:53:34 I mean, it’s a zero sum game for us.
0:53:36 – It’s an opportunity cost.
0:53:36 – One thing that’s funny,
0:53:39 just listening to you talk about what Sonal brought up,
0:53:43 this question of not invented here syndrome.
0:53:46 And when you contrast that with managing,
0:53:49 having an organization that is naturally curious
0:53:51 and unbossed, as you said.
0:53:52 – Inspired.
0:53:53 – Inspired.
0:53:56 But managing that not invented here syndrome
0:53:59 versus maintaining sort of the skepticism
0:54:01 that things might be in a hype cycle
0:54:03 and not sort of chasing hype.
0:54:05 It’s a very fine balance, right?
0:54:08 It’s kind of like the not invented here.
0:54:10 The other side of that coin is not invented yet.
0:54:12 And you got to figure out like where you are in that.
0:54:14 And I think that is one of the most difficult things
0:54:16 that I would imagine that an innovative company
0:54:18 at this scale at which Novartis operates
0:54:21 has to always find that balance between.
0:54:22 – Absolutely.
0:54:24 I mean, there is a balancing act
0:54:26 between the different forces.
0:54:28 And I find a lot of it comes down
0:54:32 to just encouraging people just to have open, frank debate.
0:54:32 – Yes.
0:54:35 – And be comfortable with task conflict
0:54:36 without personal conflict.
0:54:38 That’s what I keep telling our team.
0:54:41 We have to be incredibly curious about one another,
0:54:42 what one another thinks.
0:54:45 Think that’s just all about trying to get the best ideas
0:54:46 and we’re just trying to debate.
0:54:48 But it’s never personal.
0:54:49 And it’s never, ’cause I think when,
0:54:50 particularly in the world of science,
0:54:52 it often becomes personal.
0:54:55 It becomes, this is about me and my science
0:54:57 versus you not believing in my science.
0:55:00 As opposed to saying, we need to just find a great medicine
0:55:01 or we need to just solve this problem.
0:55:04 That’s a journey I think we’re taking the organization on.
0:55:07 But I think that’s going to be what’s really critical
0:55:10 is having that radical transparency in the open debate.
0:55:13 – I find it fascinating because it alludes to the concepts
0:55:14 around skin in the game
0:55:16 because you want people to have skin in the game.
0:55:18 But at the same time, they need to have just enough out
0:55:19 that they can see things a little clearly
0:55:23 where you’re not like only attacking their sacred cows.
0:55:25 – Skin in the game, but not vital organs.
0:55:25 – Yes, exactly.
0:55:27 That’s a great way of putting it.
0:55:27 I love that.
0:55:29 – How long have you been in the CEO chair now?
0:55:30 – One year.
0:55:33 – What’s the, having come up through the R and D side
0:55:34 of the organization,
0:55:37 given the most surprising thing to you now as the CEO,
0:55:40 given that R and D is such a big part of what the company does.
0:55:44 – I’m just amazed by how vast our company is.
0:55:46 I mean, I think even though I’ve been at the company
0:55:50 since 2005, now actually overseeing a company
0:55:53 that’s 120,000 people in 150 countries
0:55:57 and you go anywhere, we are just a vast, vast company.
0:56:00 So that’s one thing that’s really, I think surprised me
0:56:02 just to have to know, when you think about
0:56:04 making a transformation happen
0:56:07 and you try to make that happen in such a large enterprise,
0:56:11 that certainly really, I mean, that really hits you.
0:56:15 I think the other thing about this job is crisis management,
0:56:17 which you just not exposed to.
0:56:19 I mean, this job is a lot about managing crises
0:56:22 and that’s been a big learning curve for me
0:56:25 because in the world of R and D, we had clinical trials
0:56:26 the last two or three years
0:56:28 and everything’s sort of predictable.
0:56:32 I mean, we sort of know what the decisions we need to make.
0:56:35 A lot of documentation that you can lean on.
0:56:38 Now you’re in the world of the ambiguous, the uncertain
0:56:41 and then things hit you completely from the blind side
0:56:43 and then you gotta keep moving ahead.
0:56:45 – If you were to write a letter to grad students
0:56:48 or just people kind of entering the space,
0:56:50 like what kind of skills would you encourage them to have?
0:56:54 Like if you could have added things 20 years ago,
0:56:55 what would you tell them to do?
0:56:58 – I’d say focus a lot on how you lead people.
0:57:01 I think there’s so much of a focus on technical expertise
0:57:02 and thinking that that’s gonna get you there.
0:57:05 It matters, of course, competence matters tremendously
0:57:07 but what really makes the difference
0:57:09 is how you lead people, how you lead yourself.
0:57:13 And I think investing more in that would pay off a lot.
0:57:15 I think the other thing I’d say is don’t underestimate
0:57:18 the importance of getting multidisciplinary exposure.
0:57:21 I mean, I think most people get worried
0:57:22 when they have to make those jumps.
0:57:24 I’ve had a career at Novartis
0:57:27 where I’ve worked in commercial areas and marketing areas
0:57:30 so most of my time in R and D worked across
0:57:32 four different areas of the business.
0:57:35 And so with that diversity of experiences,
0:57:38 it enables you, I think, to take the right decisions.
0:57:40 – There was one other point I wanted to raise.
0:57:42 I think what’s often lost some people,
0:57:44 ’cause you mentioned the miracles, right?
0:57:46 And how incredible it is
0:57:48 that we find any human medicines at all
0:57:49 because if you think about it,
0:57:52 every human being is probably 40 trillion cells
0:57:54 that are working together.
0:57:55 – It’s amazing anything even works.
0:57:56 – It’s amazing.
0:57:58 We understand a fraction of the proteins,
0:58:02 what they do, 1,200 drugable proteins,
0:58:03 and there’s only a fraction of those
0:58:05 that we can actually drug.
0:58:08 We don’t know what most of RNA does, non-coating RNA.
0:58:12 We don’t know most of what the genome’s even talking about.
0:58:15 And if you look at it, since the creation of the FDA,
0:58:20 there’s only been about 1,500 new molecular entities ever found.
0:58:21 – Wow.
0:58:23 – And most of those are actually overlapping
0:58:25 in similar therapeutic areas.
0:58:26 So actually, if you were to count for,
0:58:27 I haven’t done the analysis,
0:58:28 but if you count for double counts,
0:58:32 my guess is it’s in the hundreds of medicines
0:58:33 that we’ve actually found.
0:58:36 – And by the way, what’s the predominant therapeutic area?
0:58:37 – Probably, I would guess,
0:58:39 hypertension, cardiovascular disease,
0:58:41 but I’ve not looked carefully.
0:58:45 But it’s worth reflecting on how hard it is to do what we do.
0:58:47 And when we find, I tell our people,
0:58:49 you have to think every medicine we find
0:58:51 is a miracle that fits in the palm of your hand.
0:58:56 We’ve unlocked, in a sense, a billion years of evolution
0:58:58 of the eukaryotic cell and human biology.
0:59:02 And somehow we found something that was able to move the needle
0:59:04 in this incredibly complex system.
0:59:05 I think that’s easy to forget
0:59:10 when we just kind of overly simplify what we do.
0:59:12 – That’s a great note to end on.
0:59:14 Voss, thank you for joining the A6 and Zee podcast.
0:59:15 – Thank you.
0:59:16 – Thanks so much. – Thanks so much.
0:59:26 [BLANK_AUDIO]