AI improves patient care

Share this article

Curing and caring for patients is a comprehensive process that requires research to discover new drugs, clinical trials to test the safety and efficacy of those drugs, and coaching and mentoring to ensure patient engagement and the creation of healthier lifestyles. At each of these steps, there is an overwhelming amount of data—more than any human researcher could keep track of in a lifetime. Furthermore, most of it is unstructured, hidden in images or the text of publications. These are not searchable and are difficult to analyze. Two years ago, IBM founded a new division, “Watson for Health,” dedicated to applying its AI technology to help the healthcare sector cope with big data and improve patient care on all fronts.

The partner to any good detective

Watson is a cloud-based cognitive computing system that can understand, reason, learn, and interact. It can be applied in all domains where big data is available, from finance to life sciences. In each of these fields, Watson is initially trained on the ontologies. It can then dynamically learn from new data that is continuously added and discover new connections by cross-referencing different sources. It can extract and uncover meaning from structured and unstructured data on a large scale and help researchers discover new opportunities previously hidden in disconnected data. Ultimately, it helps researchers create hypotheses and make confident decisions.

While this sounds futuristic, many companies are already applying cognitive technology. IBM estimates that Watson will reach one billion people by the end of 2017.

Streamlining healthcare

“Today, on average it takes €2.6 billion to research and develop a new drug,” stated Antje Niemeyer, IBM Watson Health Business Lead, in her talk at Knowledge for Growth 2017. “These research projects have a less than 12% success rate, and it takes over 10 years to bring a new drug to the market. In the process, there are millions of studies out there that researchers have to scout through. This is a task that Watson for Drug Discovery can perform for you.” The ALS research performed at the Barrow Institute is an excellent case for Watson’s value in research projects. In a matter of months (instead of years), researchers discovered five new protein targets that could hold the key to the treatment of this devastating disease.

Watson could also help companies and research centers streamline their clinical trials. “Trials are responsible for 60% of the time spent in drug development,” Niemeyer stated. “Delays are encountered in 86% of the trials, and each additional day comes at a significant cost.” Emphasizing the importance of a better system, she said, “Watson Health can help companies to optimize clinical trials through better protocols, a better patient selection, patient recruitment, site selection, and overall a faster time to launch.”

Life science expertise from partnerships

While IBM provides the cognitive computing platform, a lot of real-world data is required to train the system. Here, the company teams up with the top pharmaceutical companies and research centers. “For Watson for Drug Discovery, we have announced a big partnership with Pfizer recently,” Niemeyer clarified. “Our platform Watson for Oncology has been trained by the Memorial Sloan Kettering Cancer Center, harboring around 1,000 oncologists.”

Of course, the partnerships are not limited to training the system. They also exist to create new applications. “We are also working together with companies like Medtronic, where we have developed an app that can predict adverse clinical events three hours in advance,” Niemeyer said. “A partnership with Under Armour focuses on bringing counseling and coaching,” she continued, “and with the American Society of Cancer, we are creating an interactive advisor to support patients when they don’t feel well.”

The democratization of cognitive computing

IBM is opening up the system for everyone who is interested. The company provides APIs in the cloud that can be used to develop new applications. The business model to generate income is built on a monthly subscription fee. This allows you to access the cloud-based system that is permanently updated with the latest data. Companies can choose to make their own private cloud to protect their own data and combine it with what is publicly available. There is also no need to worry about IP. “We are not the biotech expert, we focus on providing the platform,” Niemeyer said. “If you make any discoveries, those are yours. It is your data, your money, and your success.”

With its technology, IBM hopes not only to accelerate drug discovery and streamline clinical development but also to engage with patients and provide them with better and more personalized experiences.