AI Throughout the Healthcare Sector: Interview with Nuray Yurt, Novartis

Nuray Yurt is the Worldwide Head of Artificial Technology Solutions at Novartis, leading the charge towards combining data analytics with innovative AI technologies to produce new insights and patient-oriented solutions. Combined with her prior experience in business insights and economics, she is in a privileged position to speculate on the future of AI in healthcare and the transformational impacts it promises to bring

PF: It’s a delight to have the opportunity to pick your brain in this discussion, Dr. Yurt. Thank you for being here! We could not possibly begin with a question other than why Artificial Intelligence (AI) in pharma? Your long career has taken many diverse paths, from Mathematics, to Economics and Business Insights – but what drove you where you are now?

NY: What actually drove me here is the potential impact of AI and its potential to transform healthcare. Obviously, the pharmaceutical industry is just but a part of our overarching health system – mathematics, economics, business insights and their associated data analytics also have a very large role to play in that system. All of these areas have been, and will continue to be, inextricably linked with the progress we see in the field of AI. My move towards a more direct, pharma-linked role seemed natural – my passion for information and data runs much hotter when it is applied to a field where it can have a positive effect on the wellbeing of people around the world. Technology can assist us not only in improving patient outcomes, but it can also assist the patients themselves for the betterment of their every-day decisions.

PF: Artificial Intelligence (AI) is still a new phenomenon in pharma. What areas do you foresee seeing a significant impact? We have already seen large moves to assimilate it in drug discovery, improve the integration of heterogeneous multi-omics datasets, protein structure prediction, automating and standardizing pharmacovigilance as well as medical imaging and diagnostics. What kind of progress do you foresee here? Are there other specific fields where you foresee a large effect? 

NY: You are absolutely right – it is a very new phenomenon in pharma. For better or worse, this industry is reluctant to change – not merely because of tradition, but also because of the regulatory dance that must follow in lockstep with innovation. Legal challenges and governmental procedures may have slowed down the uptake of progressive AI technology, but we are already seeing big moves in drug discovery – as you said. We also see great strides in other departments – such as how we communicate with patients, more systemic account-keeping tasks, and assisting decision-making. I certainly think there is a lot of space to be explored in data analysis and interpretation with the assistance of AI, in order to improve such low-level decisions that interface directly with patients and customers. These include treatment reminders, diagnostics – we have seen gradual steps to implement these, and at some point I expect the phenomenon to reach a critical mass where we will really see it flourish. 

PF: There are many challenges to overcome as we improve the AI tech available for the life sciences. One of these is simply that the technical expertise just isn’t there. We had seen this before during the rise of computerized bioinformatics – where the lack of an intersection between computer science, software development and biology hampered the growth of the field. Do you see the same phenomenon in AI – and what do you think should be done to ameliorate it?

NY: This is truly spot on – and aside from the regulatory challenges inherent to the pharma industry, is one of the largest contributing factors slowing down the growth of AI in the life sciences. I would not say the technical expertise is not there – it is, but just in different industries. We have the tech experts, the health business experts, the life scientists – but we need to improve in interfacing the former with the last two groups and enabling cross-disciplinary communication and collaboration. I have seen improvements first-hand, with pharmaceutical companies demonstrating a growing interest in hiring computer science and technology professionals; we see data scientists being in high demand across North America and Europe. As the talent shortage is handled, with data experts that also have an interest in the life sciences entering the field, we will see the field truly blossom. And I say data experts because a true data scientist is dedicated to understanding the data they work with – and the subject it is based on, rather than just computer scientists. I think this talent gap will be comparatively easier to bridge rather than what we saw in the 2000s with bioinformatics – back then, we truly were looking for two different fields of expertise to be embodied in a single person. With AI, a competence in data science, supplemented by an interest in the life sciences and healthcare, is sufficient qualification.  

The growth of AI in pharma will come with unique bottlenecks not seen in other fields. PF: For example, with increased involvement in clinical settings, AI will inevitably find itself in decision-making points that can have life-changing consequences. The problem of liability – who should be responsible for such decisions – as well as broader ethical problems will inexorably rear their heads. How do you think these problems should be handled on a societal and regulatory level?

NY: This is a truly loaded question, with widespread impact. I do not think anyone has a true, definitive answer that solves all the anxieties represented by the topics it raises – otherwise, we would already have a solution in our hands. This affects all industries that hope to use AI – not merely healthcare, although in this industry the impact becomes undeniably more personal. But even with entities such as self-driving cars assisted by AI, we see dilemmas regarding liability pop up. But the urgency with which these decisions must be answered is definitely increasing – and I do not think there are right or wrong answers: it is up to society to decide what kind of decision-making capacities it is comfortable to entrust machines with. In my own personal opinion, I do not think we will see AI entrusted with high-level, high-impact decision-making competencies in the near future – though I am confident it will find a niche where it will assist those decisions. I think this will be particularly true for healthcare, where we face significant other hurdles such as data fragmentation and heterogeneity and health privacy concerns.  

PF: Are there other technologies you are equally excited about to see make inroads in pharma, such as blockchains – particularly for supply chains – or quantum computing? How do you think such technologies will interface with AI, and what challenges do you anticipate?

NY: As far as I am concerned, I am extremely thrilled every time I see new developments in all of these technologies. At the heart of it, these techniques and methods promise to improve what we already do and transform our perspectives on data in revolutionary ways – although part of the hype is definitely down to catchy, attractive marketing! I expect all of these technologies to introduce small and incremental changes to the way we do things as the tech matures – rather than any large, singular watershed moment. Our understanding of quantum physics has been rooted in discoveries made since the early 20th century – yet we have only seen its practical application happen relatively recently. These are promising steps – and the computational improvements that quantum computing promises to bring will have drastic effects on AI, blockchains, and traditional bioinformatics. Obviously, we are still in the NISQ-era – true quantum computing remains conceptual – and will likely continue to do so for a while. What I do think we need to do is take a step back and, in a way, “deconvolute” the buzzwords we use to describe the technologies we refer to. Even within AI we see such a cornucopia of other terms that fall under that umbrella – natural language processing, machine learning, deep learning, neural networks. Even these differ vastly according to which field they are applied to. Using buzzwords is good for grabbing attention – but discourse can benefit more from clearer terminology. And while I think we are also seeing a lot of investigations into how AI can be applied in healthcare, we will not see it become the norm in all of these implementations – AI itself has a cost, and it will be interesting to see where it establishes itself as a mainstay.

PF: March was a month dedicated to the history of women. It would be remiss of us not to ask about your own experience in pharma as not only a woman, but also an ethnic minority woman. Did you face any challenges during your journey in the world of science – and how did you overcome them? Do you have any advice for the younger generation of women, and people in general, who may be looking for more diverse role models?

NY: I am extremely appreciative of this question – particularly because I make it one of my missions to empower women, particularly from more diverse backgrounds. Obviously, we have to start from the premise that discrimination against women exists – because it does! The reasons for it may be varied, but its existence is not variable. Such discrimination may not happen deliberately or consciously – one may not mean to offend. I think this is what kept me going when facing such discrimination – I understood that it may not necessarily be personal; rather, it is an institutional and societal phenomenon. This was particularly true in my experience – as a religious Muslim woman, I found people who often did not know how to approach me. As our discrimination against women, and other minorities, exists on a collective, societal level, efforts to combat it should be aimed at that level as well. And I think ignorance plays the largest part in causing it – not malice. Efforts aimed at increasing awareness and education can have largely beneficial effects and foster dialogue between groups. I also found that science was a good area to find myself in as a minority – there is definitely a meritocratic culture across many organizations in the industry, although at times I felt that I worked much harder or more than everybody else around me. Sadly, this is a common story among people who have to overcome biases and prejudice – and it should not have to be. Still, I tried not to throw up walls and often welcomed any questions about my role or background – hoping to build bridges; after all, they come from a different background than I do as well. My advice to other women, and people in general, would be simple – if you feel doubted by others, don’t doubt yourself. Instead, make your work indubitable; make it felt; make your contributions undeniable.  

Join Proventa International’s Bioinformatics Strategy Meetings to hear world-renowned industry stakeholders and experts discuss the latest trends in data science, AI and other hot topics in bioinformatics – in closed door roundtables facilitated by leading figures from the industry!