Artificial intelligence tools are revolutionizing nearly every stage of the drug discovery process, offering substantial potential to reshape the speed and economics of the industry. As the drug discovery and preclinical stages speed up and potentially produce more drugs to test in the clinical trial phase, how do clinical researchers prepare and respond to these challenging opportunities?
In this podcast episode, Toban Zolman, Chief Executive Officer at Kivo, shares his thoughts with Katherine Vanderbelt, host of The Latest Dose, on how AI-enabled successes in drug discovery will affect clinical operations and regulatory operations. We discuss how advancements in technology and data analysis are reshaping the way we conduct clinical research.
Read the full transcript below, listen on Spotify, or wherever you get your podcasts.
Full Episode Transcript
Katherine Vanderbelt: Hi, everyone, and welcome to The Latest Dose, the podcast that explores the depth of innovation in human compassion in clinical research. I'm your host, Katherine Vanderbelt, Global Vice President of Clinical Innovation at Oracle Health Sciences.
Drug discovery is a time consuming and expensive process. A host of artificial intelligent tools, AI, are said to be revolutionizing nearly every stage of the drug discovery process, offering substantial potential to reshape the speed and economics of the industry. According to Boston Consulting Group, as of March 2022, biotech companies using an AI-first approach had more than 150 small molecule drugs in discovery and more than 15 already in clinical trials. Does the drug discovery and preclinical stages speed up and potentially produce more drugs to test in the clinical trial phase? How do we prepare and respond to this exciting, new and challenging opportunity?
Today, our guest will share his thoughts on how AI-enabled successes in drug discovery will affect Clinical Operations and Regulatory Operations. We will discuss how advancements in technology and data analysis are reshaping the way we conduct clinical research. Joining me today is Toban Zolman, Chief Executive Officer of Kivo.
Toban has 20 years of experience in regulatory and clinical operations, drafting some of the first guidelines for electronic submission at Image Solutions. Toban has consulted with 47 of the top 50 pharma companies in the world. After working in regulatory, Toban ran product teams for several tech companies. Toban has been at the forefront of multiple tech revolutions, such as cloud computing and the Internet of Things. And Toban thinks the time has come for clinical trial management to level up.
Toban, it is great to speak with you today. Welcome to the latest dose.
Toban Zolman: Thank you. Great to speak with you as well.
Katherine Vanderbelt: In the intro, I mentioned that you believe the time has come for clinical trial management to level up. What do you mean by that?
Toban Zolman: Let me give you some context on where that comment is coming from.
I spent a chunk of my career helping Tier 1 pharma transition to electronic submissions.The promise of electronic submissions was improved process, improved visibility, faster review times by regulatory agencies. The way that we went about that as an industry 15 to 20 years ago was to take this new challenge of managing a 10x increase in the amount of documents going back and forth to a regulatory agency and controlling that incredibly tightly. Literally, I spent years in windowless conference rooms with committees, trying to figure out how to manage every aspect of an increasingly complex process. And honestly, it was soul crushing. So I left the industry and spent over a decade working in other industries that were on the edge of major transformations - eCommerce, social, cloud, IoT, and eventually circled back to Life Sciences. As I came back into Life Sciences and started to talk to clinical and regulatory leaders who were dealing with all of these advancements in how clinical trials operate, [I noticed] this was the same song, new verse.
The pace of clinical trials was accelerating, the complexity of tools was increasing, and the number of assets that they were having to manage that resulted from those advancements was also increasing. And the approach to managing all of that was to just have leaders in Life Sciences at these pharma companies just tighten their grip on the process even more. That's just not a model that works and it's not a model that any other industry has embraced. So really, what we've really focused on at Kivo is helping companies loosen their control a little bit. Not control of process, but really not trying to manage everything in a monolithic, top-down approach and instead move to more nimble, more decentralized, more collaborative processes to manage this massive increase in the amount of activity that's happening in the clinical pipeline.
Katherine Vanderbelt: Well, welcome back to the Life Sciences!
You mentioned how these individuals are sort of holding on to the existing process. In preparation for this episode, I read a number of articles and they continue to talk about how pharmaceutical industry resists adopting digital tools, the need for them to change their strategic priorities and also evolving the workplace culture, perhaps in some of the ways you just mentioned. What are your thoughts about these statements now that you're back? Is this true, or are you seeing something else?
Toban Zolman: Great question. I think you're correct in terms of meta-level trends. Life Sciences, and especially folks that work in Operations, whether that's ClinOps, RegOps, et cetera, are a very risk adverse group of people - and for good reason. The nature of those jobs and their remit within the drug development process is fundamentally to be risk adverse and that's what helps create safety in drugs.
With that said, Kivo is focused fairly exclusively on working with emerging life science companies. The majority of our customers do not have a drug and market yet. They have active clinical pipelines, but they are new companies, new in Life Science terms. Many are 15 years old. They are hitting growth inflection points in a post-pandemic world. These smaller companies that are growing rapidly and hitting inflection points post-pandemic are leaning into decentralized teams. Maybe not even by choice, it's just the nature of how you scale a company now. But they're leaning into that workplace culture of small, decentralized teams relying heavily on partners, whether that's CROs, contractors, medical writers, Reg Affairs shops, whatever it is. They are figuring out how to scale organizationally, scale technologically, and scale their clinical trial process in that landscape.
So the conversations we have with leaders in those companies who are really building the organization from the ground up differ significantly from the conversations we have with companies that reached a scale point a decade ago or pre-pandemic, where the workplace culture was centered around in-person, everyone working in the same office sort of a culture.
The industry is risk adverse, Ops folks are risk adverse. The customers we work with that are most successful are the ones that are baking into their corporate culture from the ground up a more nimble, decentralized approach to managing this influx of data. So that makes sense to me.
Katherine Vanderbelt: I still think there's a disparity that I'd love to get your thoughts on. We talk about AI, the promise, the culture...but we also see that we've had cloud around for more than 20 years. Yet there are articles that say that 50% of clinical trials are still utilizing paper processes somewhere in it. So how do we deal with this disparity? How do these large companies deal with this? What are your thoughts on what they need to do?
Toban Zolman: I would frame the conversation about AI and Cloud this way. Cloud in Life Sciences is very different than cloud in other industries. The majority of the incumbent software vendors, especially that are offering Part-11 Compliant solutions, software that's used deep in the regulated process...They may be cloud-based, but this is technology that was created before the iPhone was invented. So the paradigm in which a lot of these platforms were built has not fundamentally changed from software and processes that were developed in the 90s and early 2000s. AI as a layer on top of that creates so much acceleration - increased data, process challenges - that those two are never going to play well together.
What you are starting to see in the industry, it's almost like looking at geology - where you've got three strata of incompatible technology. You've got paper on top of that, you have SaaS-based, cloud-centric solutions that fundamentally aren't very cloud-like, and then you have AI swirling on top of all of that.Those three things are very difficult to stitch together, especially if a company is attempting to take a monolithic view of how to control that process.
If you look at organizations outside of Life Sciences that have adapted and grown quickly, there are some commonalities in how they approach new technology and apply that technology in the organization. Amazon is a great company that comes to mind in terms of how they approach this.
Amazon is obviously massive in scale, but the way that they run that company follows the "two pizza" rule, where there's no team that can't be fed with two pizzas at lunch. That team manages all aspects of a project or product and has effectively total autonomy to drive features, process, et cetera. Within life sciences it's possible to take a "two pizza" mentality, especially as AI helps accelerate the pace at which net new assets are spun out that may be completely discrete from other products in the company's pipeline.
What we've seen that's been successful with companies that have grown rapidly at Kivo is not to try and scale the organization in proportion to the pipeline or in proportion to the amount of assets being created, but rather to create tightly constrained teams that have high degrees of autonomy and authority. Those roll up into ultimate decision-makers on clinical and regulatory but have the ability, operationally, to adapt and dictate their own workflows. That may sound scary to some folks, especially coming from a paper world where it was possible to have a pharma company with 50,000 people and everyone does the same process. That's just not super practical these days. Picking tools and defining process that enables teams to be autonomous and nimble is really the only way to proportionately scale an organization to keep up with the tsunami of advancements being driven by cloud and AI.
Katherine Vanderbelt: In our last episode prior to this one, we actually had a conversation about creating drugs at the speed of AI. You talked about the increased input that's coming into these organizations. You talked about the two pizza teams, you talked about utilizing technology...This is a big change for pharma. I've been in pharma for many years. This is a big change. Really doing the work differently.
What are the implications of this and what advice do you give folks to scale?
Toban Zolman: First off, it's been fascinating to be on both sides of this. Helping companies really codify a process in the early 2000s around how to manage, how to transform the entire organization from paper to electronic. Obviously there's still some holdouts in the process there, but really that was a transformational change. And now seeing another transformation of industry (cloud and AI) really driving changes in how assets are developed and more rapidly finding promising new drugs.
Customers that we are working with that are managing this transition effectively are really doing two things. The first thing is they are running guardrail management, which means for their organization, from the top down, they are defining guardrails in terms of process and technology that they want individual groups to follow. They are not dictating every step, every workflow that has to happen with every team, but rather creating a North Star that everyone is working towards defining policies and operating procedures that define the parameters in which individuals and departments have authority and autonomy to work within. Generally giving those teams the discretion to identify the most effective way to work.
Because let's be real about that. The speed at which things happen in clinical pipelines today is faster than [the time in which] a typical company could author, approve and train an SOP. So by the time you get your massive global process defined and implemented, enough tech has changed and enough insights have been drawn out of the data that it no longer makes sense. So defining guardrails for groups and then letting them operate within those while still staying compliant and still meeting the goals of the company is the key. A Chief Medical Officer or a VP doesn't necessarily have the operational insights to be that prescriptive anymore. So I think that guardrail-based management is super effective.
We have a customer that in the past ~ 16 months has gone from something like 2 to 15 assets that they're managing with a very small number of employees. They have not scaled their organization. They have not doubled and then doubled and then doubled again. In terms of headcount, they've maybe grown 30%. That growth has been really centered and focused around asset classes where individual groups have the ability to figure that out on their own within budget and some general guardrails.They're one of the fastest moving life sciences organizations I've worked with as a result of that. A key lesson that we are seeing is changing that top down mentality.
The second trend that I would point to is taking a similar view of technology. At Kivo, we see this from a document management or a process management perspective because that's the software we build. But this really is true throughout the entire stack. Especially on the tools that are used on the AI side, the machine learning side, and workbench tools to try and find insights into pharmaceuticals, or tools to speed up and better analyze data on the clinical side. Throughout that stack, I think teams that have the ability to select, implement and iterate on those tools in a rapid fashion are the ones that seem to be adapting and increasing their pipelines the fastest.
With modern cloud tools, with APIs less reliance on centralized IT, it's possible for a very small company to go very quickly and do all of it in a really pretty controlled way. But it takes really thinking through the tools and thinking through the process in a way that is not nearly as top-down and prescriptive as it may have been a decade ago.
Katherine Vanderbelt: Thinking through those two suggestions you have around the Guardrails and also how to handle technology advancement. Are Clinical Operations and Regulatory Operations prepared for this big change? I get the examples you've given with companies that are starting and growing and so forth, but what sort of investments or what improvements... was what they experienced going from paper to digital in the early 2000s enough to prepare them? What else have you seen prepare them for the change?
Toban Zolman: I would say is it is a mixed bag. That's not a way to dodge the question; the amount of deviation that we see across organizations is significant.
I think there are individuals in the industry who get it and really are embracing these trends as a way to accelerate development, and see that there is a path to do that while preserving safety of the drug development process. And I think that there are others, some of whom have legitimate perspectives, that are very much under prepared for the sea change that is happening with these tools. [Those people] are continuing to frame everything all the way back to paper.
Much of what I think happened in the transition from paper to electronic is that electronic processes were still fundamentally rooted in how you operated in paper. They were more efficient. Literally the constructs in software UI, the steps in the process, all of that still came back to underlying philosophies around where documents sat in what file cabinet and what that file cabinet represented, whether that was draft documents or approved documents or things of that sort.The entire paradigm of managing electronic data is still fundamentally anchored in a paper view of the world.Organizations and software that I think have gone beyond that have been able to create much more nimble processes and are probably better prepared for the AI tsunami. Organizations and individuals that are still managing electronic data in a paper paradigm are in for a world of hurt.
When we're talking to a Life Science company for the first time and we're asking questions about how they solve specific problems...if it's anchored in references to file cabinets, we have one set of responses. If it's anchored in in terms of decentralized teams and collaboration and process management, it's a different set of responses.
I think what you're really seeing in the industry right now is - and I hate to use the term paradigm - but a paradigm shift, really a transformation in how work gets done and how companies think about what that is. We have a close partner at Kivo who talks about how the product of pharma companies is documents, not pharmaceuticals. And I think for the majority of individuals within a pharma company, that's totally true. That is what they do day in and day out is prepare documents that represent some portion of the narrative of their clinical trial and ensure that those get teed up to a regulatory body for approval.
While that narrative element and communicating everything to regulatory agencies through documents is true, organizations that can shift that thinking and really understand that they are developing pharmaceuticals, biologics, whatever they are working on, in the context of an ongoing, evolving process are the ones that are ultimately going to be able to adapt to this and be prepared for the transition most effectively.
Katherine Vanderbelt: That's great. So I agree with you that I still come across individuals that are working with technology based on a paper-fundamental process. So my question to those individuals that you're familiar with or collaborate with or discuss with through the COVID-19 pandemic...we had to change the way we're working. In many cases it pushed us to do things very differently than we had to do in the past and to be very creative yet safe and produce quality. Do you feel that these trends actually helped people to move towards where they should be going or do you think it's more of the same?
Toban Zolman: I actually do think that COVID was a game changer. I think that's true in a lot of industries, but certainly in Life Sciences.
I think it affected two things in a fundamental way. One was business process. You had to move to a decentralized approach. And the second is so many clinical trials were affected by COVID. If you had a clinical trial in progress when COVID hit, well, doctors can't be in the same room as their patients for every single visit or there has to be social distancing. Data collection methods changed and that forced many most clinical trials to switch up protocols, update documentation and frankly, created a highly dynamic environment on the clinical side, where changes to protocols for clinical trials changed on a weekly or daily basis and changed from one country to another in a way that was way more aggressive and dynamic than I think most individuals had ever dealt with before. And that, frankly, was a forcing function to be more adaptive, to leverage a different class of tools, a different class of partner and I think forever changed the operational aspects of how clinical trials are run.
I think the adoption of AI and the willingness to transition to more decentralized models in terms of process and technology was accelerated by a decade or more due to COVID because it changed the baseline of what is acceptable in terms of how dynamic a process can or should be to move drug development forward while still maintaining safety. I think more than AI, more than cloud, COVID was probably the biggest accelerant in the drug development process in the last two decades at least.
Katherine Vanderbelt: So now that we have your advice around guardrails decentralized teams, leveraging the cloud, leveraging AI, take the learnings from COVID-19...Embrace them, take the learnings and keep going. Is there any other improvement or change that the Regulatory Operations and the Clinical Operations work team need to take in account that we haven't touched on yet? T
Toban Zolman: There's three aspects from a meta-level to any sort of transition, really, that you can anchor against. So 1) people, 2) process, and 3) technology.
On the people side, redefining who makes up a team is a critical part of this. This is a pendulum in life sciences where business process outsourcing becomes super critical. Organizations move to that to restructure cost, et cetera. Then the pendulum swings back the other way and they bring those roles in house.
When I talk about people, there is a different trend happening now. And it's not just business process outsourcing and small companies work with CROs. It is really a fundamental rethinking of how organizations scale and how teams are built. It is not where you have employees of a sponsor and employees of a partner and there is a wall where documents are lobbied over from one organization to another. What we are seeing is a much more fluid arrangement between sponsors and partners, where partners - and they could be sizable organizations, they could be independent contractors - are bringing specific domain expertise into the team and are a core part of that team and process for whatever duration makes sense. Could be six months, could be six years. Companies that seem to be doing exceptionally well are ones that really are looking for the most efficient way to bring the right domain expertise into a team. And it doesn't matter necessarily, at least in the short run, where that person sits; if they're a sponsor, employee or a partner.
At Kivo, we can see companies that are moving very quickly. Our internal stories about them are "man, did you see Acme pharma? They are cruising." If you look at their users in our system, you see a high intermix of email addresses that belong to the sponsor and email addresses that belong to partners because they are shoulder-to-shoulder, elbow-to-elbow, working through the challenges of bringing a drug to market as a team. Everyone bringing their best expertise to the table. Rethinking where that domain expertise comes from and how you scale a team internally versus through partners is a key piece of that.
On the process side, I won't repeat what I've said previously, but I think a guardrail-based approach to process development versus a prescriptive approach seems to be what is really helping successful companies drive innovation faster.
Finally, on the technology side, drug development is no longer a monolithic process even within a single company. Individual teams are going to be using individual tools. You'll use tools for short periods of the process and then wind those down, especially with AI. So I think the tools that you do pick to be anchor points in your process need to be nimble amd highly configurable so that you can adapt them over time.
Frankly, you need to have a change management process that can go incredibly fast so that you can adapt those systems to your people and process as they change, because they're going to be highly dynamic. All three of those areas have to be able to scale and change quickly and all adapt around each other. The days of mapping out a process in a conference room, paying a vendor to implement that process in software and then revisiting that in five years is no longer a model that makes sense. So people, process and technology, that's the key. They all got to work together in a dynamic fashion.
Katherine Vanderbelt: Thank you for that great advice.
As we come to the close of this episode - you've shared some excellent, great content for our listeners to think about - I really want to ask you this question. What do I, what do the listeners do? What do we need to start doing today in order to embrace this new way of working, this new reality, to really capitalize on this opportunity of faster drug development coming at us? Share your thoughts with us. What we could do tomorrow after we've listened to you?
Toban Zolman: There are a couple levels to think about this from. At the top level, taking one step back from the day-to-day and not necessarily focusing on what do I need to do next, but instead thinking about what is the ultimate objective. In my role in the trial, what actions actually move the needle the most?
As someone who talks to customers every single day and gets asked for features, being told a process is changing, whatever it is, what I have found to be really effective is asking, "Why do you want that feature? Why are you making that process change?" The Five Whys (asking why five times) to get to the root reason why something is turns out to be really effective because a lot of decisions, a lot of process... it turns out the original reason why you are doing that has been so mutated by process, by committee, by interpersonal challenges, political challenges within the company, that if you really strip down, what is the end goal and why are we making changes to get to that? It turns out you can often short circuit a whole lot of added complexity and instead strip away complexity to really get to the core challenge. That's kind of a big idea. But in terms of what people can do today or tomorrow, I really think it is less about focusing on explicit deliverables, even though that's how everybody gets measured for their job, and instead really thinking about overall outcomes.
I don't mean to sound like a process consultant when I say that, but fundamentally, that's what drug development is about. It's about producing positive outcomes for patients that are safe and effective. Reframing process based on the outcomes of a process, not based on individual deliverables, not based on a description and composition document that was produced. Instead, the ultimate goal is not the document, but it's getting engagement with the regulatory authority faster or moving a clinical trial forward faster.
Reframing day-to-day tasks in a way that is measurable and visible across an organization turn out to really be key.Then the conversations seemed to be less about horse trading on deadlines and trying to figure out who's responsible for a certain deliverable and really more about how you align what you're working on with other contributors, employees, partners, et cetera, to really drive the ultimate outcome. For clinical, it could be getting into a clinical trial or completing a clinical trial. On the regulatory side, that could be getting a submission to a regulatory agency or approved by a regulatory agency. Understanding what the ultimate outcome is instead of the deliverable...then you just figure out how to get the deliverables marshaled through...turns out to be a positive way to reframe things and think about things at an organizational level.
Katherine Vanderbelt: Well, Toban, as we prepare and respond to this exciting, new and challenging opportunity, I am very grateful, and I'm sure my audience is grateful for you sharing your experience and advice on how to embrace this change and prepare for the bottleneck. I really appreciate your time with us today. Thank you so much.
Toban Zolman: Thank you. I enjoyed it.
Katherine Vanderbelt: Thank you for listening to The Latest Dose, the podcast that explores the depths of innovation and human compassion in clinical research. Before you go, show us some love by subscribing, and make sure to look for us next month. Goodbye.
