Optimizing Life Sciences Commercialization Through Digital-First Strategies with Tarun Mathur

In today’s rapidly evolving life sciences landscape, the fusion of digital technologies with commercialization strategies has become indispensable. From research and development to marketing and distribution, the adoption of digital-first approaches is reshaping how life sciences companies operate. This shift promises to accelerate innovation, improve patient outcomes, and optimize the entire commercialization process. By leveraging data analytics, artificial intelligence, and other digital tools, organizations can unlock insights, streamline operations, and personalize engagement with healthcare providers and patients. However, this transformation also presents challenges, including the need for significant investments in technology and organizational adaptation.

This week, we welcome Tarun Mathur from Indegene, a seasoned enterprise technology leader and Chief Technology Officer with over 20 years of experience driving digital transformation solutions in the life sciences industry, focusing on the innovative application of AI across the pharmaceutical lifecycle to positively impact clients’ businesses and patients’ well-being, driven by a passion for uniting technology with deep domain expertise and a commitment to solving complex problems and advancing technology within the industry.

In this interview, we explore Tarun Mathur’s 25-year journey in digital technologies within the life sciences industry. Tarun discusses key differences observed in operations and functions within this sector, alongside the impact of digital advancements on his role. We delve into life sciences commercialization, focusing on digital-first strategies and Indegene’s evolution in providing solutions amidst technological transitions. Tarun also addresses factors enabling and limiting the potential of data-driven technologies in life sciences, emphasizing the importance of overcoming challenges to serve clients better. Lastly, the discussion underscores the critical need for responsibility and ethics in deploying these technologies to ensure societal well-being in the life sciences domain.

AIM:  Could you share insights into your journey within the life sciences industry? With your extensive experience, what are some notable differences you’ve observed in operations, sales, or other functions within this sector compared to others? 

Tarun Mathur: I’ve been in the Enterprise Life Sciences space for a little over 25 years, and I started with a background focused on e-learning. That was my main area of interest: how can we do a more effective job of teaching over the Internet? This was back in the 90s, and that led to some interesting innovations. The insights into how we can, not just teach, but also communicate digitally. The timing was good, as at that period, the Life Sciences industry was looking at training salespeople, and also considering how to digitally engage better. My original company was founded by physicians and entrepreneurs, so we had a diverse background. My own background is in physics, with a concentration on the medical side, and I am deeply plugged into that field. It became clear to us that conveying content and information through a new medium was the way to go, and so my journey began by thinking about e-learning and then upgrading to general digital communication – how to implement things, measure it, and make it more effective. In 2005, the company that started in the 90s with a couple of colleagues got acquired by Indegene. There, we saw the opportunity to leverage technology in a differentiated way, positioning ourselves at the intersection of healthcare domain expertise and technologies from day one. We had lots of physicians and other medical and business experts coupled with technologists, and we wanted to keep investing in that intersection. My journey started with small teams focused on web applications and early mobile CD-ROMs, and as the engagement increased, so did the team size. Eventually, we went from just engineering teams working with domain experts to more cross-functional teams. I started with a team size of 10, growing to hundreds of people in the engineering group. Now, Indegene has roughly 5,000-plus employees overall across the world. It’s a very different scale, but the journey was always with the lens of Life Sciences, exclusively focused on that. We looked at trend lines and how to exploit the intersection of domain and tech better to achieve better outcomes for patients or caregivers, and for the business stakeholders as well. It’s been a very interesting journey, and while my title has always been the same – Chief Technology Officer – the role within the organization has evolved to keep pace with these tech inflection points and the changing industry needs.

AIM: Can you outline your 25-year journey with digital technologies in life sciences? How have these advancements impacted different functions and your role?

Tarun Mathur: Most of us familiar with Enterprise Life Sciences acknowledge that the industry tends to be a little laggard on the new technology adoption front. While something new comes out, the pace of adoption and its impact on the industry tends to trail some other industries. And that’s rightfully so, given the nature of healthcare, the risks involved, and just how all those functions work. It’s good to be careful and measured, but we have seen disruptions.

I remember the inflection to cloud computing, which took a little time for Pharma to embrace, but then they went all in. That impacted the value chain, the entire drug life cycle across the board. Then we moved up to big data analytics, and all of a sudden, you see a huge impact, which part of it was on the R&D side, but there was a significant impact on the commercial end of things because now they’re starting to rethink how we digitally engage with physicians and patients, dealing with the broader ecosystem of channels and data sources, and questions about personalization. So, you needed to start looking at embracing this big data, and that infrastructure came in.

Then, we saw the acceleration of machine learning fall into this space. So, here again, all elements of the drug life cycle are impacted by it. We as an organization certainly saw a major impact on the commercial side because the pressures on there were ones that machine learning had a clear and very visible tangible ROI for adopting it. But we also see its impact on the search and retrieval of unstructured documents, using natural language processing, and, in the early stage, R&D through clinical trials and prediction algorithms, trying to predict if patients will fall off protocol.

So, you saw these major kinds of technology inflection points adopted. The commercial teams were the quickest to adopt, but it would cause a ripple effect throughout the entire life cycle. Today, we’re seeing it happen again. Pharma is also impacted by the whole nature of GenAI that’s sweeping the world. And we’re seeing a similar disruption to the value chains associated with each step of the drug life cycle. They’re being heavily reevaluated, and some areas will be transformed in some meaningful way. Some areas will be completely disrupted, but we already see that happen. So, I don’t think this has been isolated to any specific function. Still, these are experiences where we generally start to see quick adoption and quick results from the commercial side of the business. Still, then it ripples and ultimately affects the entire stream.

AIM: Can you clarify the concept of life sciences commercialization for our audience and elaborate on the digital-first strategies that have emerged around it?

Tarun Mathur: When we refer to commercialization, we’re talking about after a drug has gone through discovery, clinical research, development, and manufacturing and when you’re finally bringing it to market. How do you get it into the hands of physicians? How do you educate physicians about the benefits and the appropriate use of your new product? 

In today’s world, we’re dealing with precision medicine and personalized healthcare. That means we also have a responsibility towards the patients, ensuring they’re educated, have the correct information, and are timely and up to date. I also broadened that field because it includes medical information. So, physicians may have certain medical inquiries or questions, and so what are the services from a company that must have in place to support those? And as a pharmacovigilance side of things, which is finally, once a drug is out there and being used, the pharma companies have a responsibility to monitor the drug for safety events and risks and report those to the health authorities and take appropriate action in a very timely and responsible way. So, all of those things combined create the umbrella of the commercialization story of the pharma companies, and it’s a major initiative. You can spend billions of dollars developing the product, but unless you have the right way to commercialize it, you won’t see any return on that. 

I certainly look at it from the business lens but also from the patient and caregiver side of things. They need to know the latest and best treatments available; maybe there’s some new product that’s come out that’s more cost-effective or has fewer tolerability issues for a patient. If your physician doesn’t know about it, then it doesn’t help. So, the ability for a pharma company to engage particularly with physicians and prescribers, and what we measure are, broadly speaking, two things: one is the script and how many prescriptions are written for your new product, and then you also look at adherence. That means our patients are prescribed a certain treatment plan, and whether they continue to refill their medications. So, from those two dimensions, we look at business outcomes, and each report has many corollaries attached. And so, from our side, Indegene has historically been focused on this commercialization part. We do some work on the early stage of clinical research and clinical trials. But the heavy lifting has certainly been on commercialization of products because if you’re able to get successful commercial outcomes at a lower cost and burden for the pharma company, that frees up resources that can go into fueling more drug research and create better programs, ultimately, all of these things are linked to better patient outcomes, better HCP information, and information management. So, we see the business benefit of having a strong commercialization partner. It goes hand in hand with the overall industry or landscape of improving patient outcomes. That’s really where it comes on. For example, I work with pretty much all the major and most of the mature-sized pharmaceutical companies on various aspects of commercialization, and most physicians would have seen some sort of content produced by Indegene at some point.

AIM: How has your work at Indegene evolved in terms of building the right applications and providing solutions for customers, considering the transition from data analytics to Gen AI in the life sciences commercialization field? And what implications does this evolution have for the industry as a whole?

Tarun Mathur: In the earlier days, when we discussed engaging with physicians, it primarily involved what they call detailing. Field representatives would visit the doctor’s office, conducting a slide presentation that underwent rigorous medical legal review. During this visit, the sales rep would engage in a brief pitch with the doctor, leaving behind samples to foster a personal relationship. Typically, reps developed familiarity with their assigned doctors, fostering a personalized engagement model.

Over time, however, the amount of time reps could spend with doctors diminished. Physicians began expressing a preference for a blend of human interaction with their rep and access to digital content. They started requesting materials such as PDFs or research documents.

This shift reflects the evolving nature of physician engagement, with a transition towards a hybrid model combining traditional face-to-face interactions with digital resources. Physicians are increasingly requesting digital assets to aid in patient care, such as articles explaining product issues or tolerability concerns specific to their practice, and data comparing the cost of medications to competitors. This shift reflects a growing preference for digital resources among healthcare professionals, especially considering recent events like the COVID-19 pandemic.

The rise of digital health trends, including telemedicine and remote patient monitoring, further emphasizes the importance of digital interactions. Physicians are now more open to developing digital relationships with pharmaceutical companies, marking a shift away from the traditional field rep-centric model toward a hybrid approach that integrates both personal and digital interactions.

As companies working in the life sciences industry, we recognize the need to adapt to this changing landscape. Our role has evolved from creating detailed content for reps to rapidly producing precise digital content tailored to specific physician inquiries. This includes self-service digital content options for physicians, allowing them to submit medication inquiries and receive timely responses.

Moreover, we understand the importance of collecting data to tailor future experiences. Our approach to digital content creation encompasses strategy, production, and measurement, with a focus on modular content development to enhance efficiency and effectiveness. This evolution in our industry underscores the importance of embracing digital solutions to meet the evolving needs of healthcare professionals and patients alike.

To fully implement this approach, it’s crucial to orchestrate the entire process in the right way. While pushing content is essential, obtaining data feedback is equally important for tailoring future experiences and making informed decisions. Over time, we’ve refined our methodology for digital content creation, encompassing strategy, production, and measurement, along with key performance indicators (KPIs). This evolution has been significant, particularly in the past few years, with our industry adopting a modular content approach.

Initially, significant investments and extensive work are required to build primary digital assets for delivery to physicians. These assets undergo a rigorous process involving medical and legal reviews to ensure accuracy and compliance. Additionally, demographic and target considerations are crucial in content distribution. Agencies are typically responsible for creating these assets.

The next step involves repurposing these assets by breaking them down into reusable modules and reassembling them while adhering to compliance rules. This process aims to maximize efficiency and effectiveness. The technology and business processes required to facilitate this transformation span the entire value chain, from engagement strategy to delivery.

Our role in this transformation involves designing a framework that aligns with these changes, particularly in our interactions with pharmacists. Through our efforts, we aim to facilitate the adoption of these innovative approaches and drive meaningful outcomes in the healthcare industry.

AIM: What factors converged to enable the development of technologies like large language models and natural language processing in the life sciences field? Conversely, what factors are currently limiting the potential of life sciences, particularly in leveraging digitized data for data-driven technologies?

Tarun Mathur: Gen AI is certainly a major inflection point in our space, and there are several reasons why this is the case. The algorithms are known, but computing power and access to them have become democratized. Also, while the algorithms and models were known, the pre-trained models weren’t in place. So, we’re talking about large language models specifically and then the upcoming GPT-5. If you look at these, there’s a slew of these big models now. The access to that computing power, the large pre-trained models, and an understanding of what it means for the industry is much clearer today.

In life sciences in particular, if you look at any business functions like commercial content production and campaign management, you have a value chain that involves a lot of decision-making. There’s complexity due to the nature of regulated content. There’s a validation of claims – are those validated from the database and the literature? All of those things are manually intensive processes. Meanwhile, you have pressures to drive down costs. So, where do you make the trade-off?

Gen AI is exciting because it allows AI technology to be created by the right people, by domain experts, using domain-specific language. You’re now coding in English instead of Python or R or C. And that kind of mindset is very transformative and disruptive to how we normally think about these things.

So, when I look at what’s happening in life sciences, there have been some legacy business processes in place that we’ve been running for decades. And we do that because even the best AI, the best automation technologies, rule-based systems, and all those things in place do not outperform the legacy process. So, we’re willing to spend the extra time and money on these processes because that reduces the risks. Now, we can reexamine that, and we’re going back and saying that some of those challenges we had in the past are potentially breakable. We can change those through generative AI. So, we’re very excited about it, and Indegene adopted a Gen AI-centric point of view very early on. We jumped into this a while ago because it became clear that as a company that runs these services, we manage services along with the technology; we are consumers of these platforms. We need them for our operations, and it became clear that there is something new.

For life sciences, for example, modular content, when you have details or even a pharma-created emailer, it has images or charts, text, and safety prescribing info – there are elements. So, these are well-known from the structure and highly controlled content. Repurposing is the ability to use AI to dissect and break that down, tag it, and bring it in for repurposing. Gen AI far exceeds anything that was there before.

So, that’s one very specific example that makes a difference in the value chain, but in our case, the data ecosystem was there. Pharma had already started investing in big analytics and data warehousing. Those elements were there. Gen AI is opening new opportunities in this which is causing us to reexamine the entire cycle.

AIM: What are some of the challenges today that act as barriers to reaching the next phase of building data-driven technologies, despite the convergence of various factors that seem to offer endless potential? How can addressing these challenges better serve your clients?

Tarun Mathur: The challenges arising from the Gen AI transformation are widespread across industries. Issues like biases and misconceptions are particularly notable in pharmaceuticals and life sciences, where inaccuracies can be highly consequential. How do we address these risks? It’s a pressing question, with existing strategies, but the primary concern is whether they yield a positive ROI.

Moreover, Gen AI, being a relatively new technology, is clouded by misinformation about its true potential. Solely approaching it from a technological standpoint overlooks a vital aspect – the intersection of technology and domain expertise. Without an organizational structure that effectively leverages this intersection, realizing its full potential becomes difficult. Differentiating between achievable outcomes and mere hype in life sciences requires careful qualification of use cases, a challenge many companies encounter.

Some people have a lot of assumptions about what AI can do, which may not be true today. Who knows what tomorrow’s model will be, so being able to qualify use cases for the domain, is a big hurdle for some companies to adopt. I will add that we’re seeing the companies recognize this. So, you look at many major organizations. They’ve done restructuring.

However, there is growing industry recognition. Major organizations are restructuring. They’ve brought in Chief Digital Officers and, in some cases, Chief Technology Officers. They’ve put in place SWAT teams to examine these cases. So, they’re aware of this problem, but there’s still a long pathway to go before that becomes scalable into real-world production. One of our responsibilities in the space is to go through that education and work with the industry to bridge that gap between separating the hype and the fiction from what real-life sciences possibilities are.

AIM: How is Indegene ensuring responsibility and ethics in the development and implementation of data-driven technologies, particularly in the context of life sciences commercialization? From an industry perspective, what measures are being taken to address the ethical implications of these technologies for pharmaceutical companies and other stakeholders?

Tarun Mathur: When working with Gen AI and making substantial investments, we approach and recommend its adoption, which is centered around responsible and ethical usage. And so, there are a couple of things that we’ve put in place. One is, that we work with our domain experts to carefully craft the systems that use Gen AI in a controlled way. These are not end-to-end systems. Nobody can put raw queries into it. They’re very much focused on tasks, not on jobs. And that’s another key point about Gen AI, our point of view today. The state-of-the-art technology is suitable for some complex tasks. Still, it can’t replace a job yet because it encompasses parameters that encompass not only what you describe but many other elements to get the outcome. So, we’ve put together a structure and a testing methodology for how Gen AI is used and when it’s deployed. It’s task-centric.

There’s process engineering attached to Gen AI adoption. For any of the functions, if you look at the processes, we re-engineer those where we bring in Gen AI just at those points where it makes sense, and we can qualify it, measure it, control it, and see some output from it. Almost always, these systems involve human-in-the-loop type mechanisms. So, whether it’s reinforcement learning through human feedback or just the flow that goes through human curation and review, that certainly is there.

The other end of it is the data infrastructure itself. The quality of results we get from Gen AI is highly dependent on the quality of available data. Certainly, there are mechanisms for retrieval augmentation and other approaches toward improving the reliability and accuracy of your Gen AI. But all that hinges on your sources of truth being well-curated, validated, controlled, and trustworthy. So having the right mechanisms and infrastructure to manage your sources of truth that feed your Gen AI systems significantly contributes to getting to what you’re asking about, ensuring I get valid and trustworthy outcomes from this. So, indeed, those elements are there, but we’re also cautious today. We are using Gen AI in specific areas where we can qualify, attach numbers, weave in observability, and all those elements to say I’m using Gen AI to solve this task. I see a 40% efficiency gain, but I can also monitor it for compliance, bias, and ethical issues. For example, I would not be comfortable having a pure chatbot as a primary point of care. We wouldn’t want that.

AIM: Given the high stakes involved in life sciences, particularly regarding providing medication and consultancy to society, how crucial is it to ensure responsibility and ethics in the development and deployment of data-driven technologies?

Tarun Mathur: Consider this scenario: on the safety side, you’re analyzing a safety adverse event form or a MedWatch form. You have structured data and a narrative, whether it’s a transcript from a patient call into a call center or a physician’s notes. Your task is to ensure the structured data aligns with the narrative in the unstructured text. While people have attempted to solve this using NLP and similar techniques, they’re only partially effective. However, with Gen AI today, could you configure it through prompt engineering and retrieval augmented generation (RAG) to accelerate this process? Could the system automatically flag inconsistencies for a case reviewer? This could be a game-changer in observing an inconsistency in a patient’s social history or medication. And this can be transformative in the business process. It significantly influences what we refer to as case processing, with parallels in other business domains, facilitated by Gen AI. Implementing this approach can yield substantial benefits because it focuses on a specific, measurable task that can be controlled. With a validated human-in-the-loop process, significant time savings of up to 30 minutes per case are achievable, demonstrating the value of leveraging this tool to enhance productivity.