IBM Watson-powered Patch AI innovation is just what the doctor ordered
The pharmaceutical sector is one of the most innovative in the world, as well as one of the most heavily regulated. Drug companies must go through extensive clinical trials with patients to get their products to market. The problem lies within keeping those patients on track. Drop-out rates exceed 30%, causing expensive delays and driving up costs.
Italian startup, PatchAi, has a solution that has already helped two of the biggest life sciences companies in the world to tackle this problem, earning it the IBM Outstanding Cloud and AI Embed Solution 2020 Beacon Award.
As a trained clinical nurse, PatchAi’s CEO, Alessandro Monterosso saw this problem up close when helping to conduct trials. “There was an outdated system for monitoring patients,” he says, adding that trials used cumbersome paper-based forms to collect patient data. They didn’t consider the patient experience.
In between visits, patients would often use messaging apps informally to check in with doctors. They could ask simple questions such as whether they should take a medicine with food, or whether they should visit the hospital due to pain. Monterosso saw this as an opportunity to combine messaging with AI and create a new kind of service: an empathetic chatbot that would be able to help patients and gather data simultaneously.
“The idea was having a conversational agent. A tool that personalized the conversation based on the patient’s behavior, preferences and demographic characteristics. It would understand their needs and clinical status,” he says.
Improving the patient experience
Monterosso enrolled for his MBA in healthcare, where he met many of his PatchAi co-founders. Like him, they came from clinical backgrounds and understood the practical side of patient engagement, but the technology was a challenge. “At the beginning, it was almost impossible to find the right technical skills,” he says.
Another challenge was to adhere to legal requirements. PatchAi had to stop short of generating entirely new conversations with patients because an open conversation could generate regulatory and liability problems. Instead, it needed a closed conversational path. Conversational outcomes had to be predictable enough to not break the rules governing medical conversations with patients, but complex and smart enough to support a wide range of patient issues.
Thanks to an IBM and Novartis sponsored initiative that included credits and consulting services, PatchAi enrolled in accelerator program BioUpper as they developed its initial prototype. This helped PatchAi form a relationship with IBM and develop its solution around the IBM cloud-based Watson AI service.
From there, the relationship between IBM and PatchAi flourished. “We got IBM’s constant support,” Monterosso says, explaining that PatchAi learned about the data structures and cloud technologies involved in Watson AI with the help of IBM.
Friction-free data gathering
Together, the companies navigated these hurdles to create a service that launched in January 2020. Delivered as a Watson-powered service through a patient’s smartphone, PatchAi accompanies the patient from the initial onboarding process and is available to the patient—whenever they need it.
For example, when a patient reports pain late at night, the Watson-powered service will ask questions about the level of pain and what events triggered it. It uses colloquial language to congratulate patients on their achievements and reminds them about their upcoming therapies.
It also reports the data from these conversations back to the cloud, filtering it for doctors with appropriate timestamps and quantitative data about pain levels and duration where available. Doctors can then interpret this data when producing their clinical trial reports.
The pharmaceutical sector that PatchAi targets has notoriously long sales cycles, yet the company has already signed agreements with Novartis and Roche, two of the top five pharmaceutical companies in the world. That’s due to the product’s success metrics. Monterosso says that PatchAi can generate savings of up to EURO 6.5 million per study in phases two through four trials, along with time savings of up to 17%. Patient adherence rates have rocketed from around 40% on average to 95%, Monteresso reports.
What’s next for PatchAi?
Monteresso hopes to expand to an omnichannel solution, complementing the text-based chatbot system with a voice assistant, possibly delivered through some of the existing consumer-focused assistants already on the market. “That would help patients with certain physical impairments along with elderly people,” he says. “We really hope that we can do that together with IBM.”
With a strong start in its first two years, we think PatchAi can do anything it puts its mind to.
Learn more about how other Beacon Award winners are changing the world with their solutions.
1. National Academy of Sciences, “The Prevention and Treatment of Missing Data in Clinical Trials”, 2010 https://www.cytel.com/hs-fs/hub/1670/file-2411099288 pdf/Pdf/MissingDataNationalAcademyof_Medicine.2010.pdf
2. ASPE, “Examination of Clinical Trial Costs and Barriers for Drug Development”, 2014
3. Applied Clinical Trials, “Non-Adherence: A Direct Influence on Clinical Trial Duration and Cost”, 2017