The Next Frontier: Reimagining Medicine with Data Science and Digital Technologies

Novartis is more than a year into our digital transformation, and people often ask me—do you still believe in the power of data science and digital technologies when it comes to finding new medicines, or is there too much hype? Can digital solutions actually move the needle in improving and extending lives?

As a trailblazer at the intersection of tech and health, Novartis has certainly learned a lot about what works and what challenges remain, and my answer is still a resounding yes. In the work of improving human health, there is no question in my mind that data science and digital technologies will accelerate our progress and bring us to new heights. 

I’ve written before about how technology is already transforming our work at Novartis, including how we’re leveraging digital solutions to make a real difference in the lives of patients and transform our clinical trials. Now I want to take a look at the promise data science and digital technologies hold to upend the way the scientific community approaches drug discovery and development, including how they can help us unlock the mysteries of human biology, find medicines faster, and get them to patients more efficiently. With an estimated 40 trillion cells in the body and 1 trillion molecules in each cell, the complexity we’re navigating is massive—and we will need the power of technology to help us discover the next wave of medicines.

We’ve found that the power of artificial intelligence (AI) and machine learning to analyze visual data is unparalleled. In the world of medicine, AI is already better at identifying skin and lung cancers than the human eye. Combine AI with unstructured machine learning, and technology can in some cases more accurately find markers of disease progression, which could help identify new patient populations and new ways to intervene in treating disease. 

Unstructured machine learning has already led Novartis to new insights related to early markers of disease in eye scans, and we’re in the process of digitizing every one of our millions of pathology slides to look for more insights. The possibilities are vast considering the range of information patients, doctors, and researchers rely on visual charts, such as CT scans and x-rays, to provide. 

The grand challenge we’re working to solve is how to bring all of our data sets together, across Novartis and even potentially across the industry, into data lakes we can interrogate. It’s a major endeavor, and our industry has been behind on this, but if we can bring together historical patient data that is clean and curated, I’m convinced we will make completely new discoveries. 

It’s already working in Novartis labs. We’ve made a few recent discoveries purely from virtual clinical trials we’ve run within our data lakes. In our boldest effort to explore our data, Novartis is creating a massive data lake with all of our clinical trial data, which spans a decade, 5,000 trials, and over 2 million patient years, to find new patient populations and drug targets. We know in many cases there are super-responding patient groups that benefit greatly from specific medicines, and we’re working to find them.

The final horizon for Novartis’s digital transformation is true drug discovery. There is incredible opportunity in that arena, but considering the complexity of human biology, it will take a lot of time and persistence to get it right. Given how little we actually understand about human biology, there is no solid training data set for the algorithm, which makes asking a machine to find the next set of drug targets tremendously challenging. I’ve come across many people and organizations around the world that are purely aiming to use algorithms and AI to find the next wave of drug discoveries, and I’m confident we won’t get there overnight. 

I do believe that we can generate insights by looking at vast numbers of chemical reactions and experiments and having machines continuously analyze them, and I also believe that the man-machine combination is a giant leap forward. There will be areas where we can leverage our knowledge of human biology and the power of computing to find better medicines.  

Novartis is already adding applications to our drug discovery toolbox, like 3D visualization, to help us find novel medicines. When we identify a molecule we want to impact with a potential medicine, it’s extremely difficult to create a drug that will fit into the keyhole and actually affect human health—so we’re using 3D imaging to enter into the molecule visually and identify the pockets in which to place a drug. This could improve the molecular structure of our medicines and further increase their efficacy and safety, and we’re rolling this approach out across our entire portfolio. 

In all, the potential data science and digital technologies have to help us reimagine medicine is profound. Novartis is committed to bringing the best minds together on this, from startups, large tech players, academic institutions, and anywhere else, because we believe technology will increasingly disrupt our industry, transform the art and science of drug discovery, and change the course of human health. 

To learn more about the scale and depth of Novartis’s digital transformation, watch my keynote speech and fireside chat at VivaTech.