Alphabet's life science subsidiary Exactly announced a strategic realignment last year, shifting its focus to leveraging AI and data science to improve precision medicine options.
Andrew Trister, Verily's Chief Medical and Scientific Officer, said: Mobi Health News We talked about how the company uses AI technology, its work with tech giant Google, and what excites him about the future of the business.
Mobi Health News: Can you tell our audience about Verily and its features?
Andrew Trister: What really came out google x The company has a long history of looking at the entire healthcare ecosystem, starting with research. That means building devices, finding new ways to measure disease (kind of like a metrology problem), and being able to provide care. So this is primarily done through Onduo, our diabetes care management system. I also work in the field of nursing care finance. In other words, Granular is a stop-loss insurance product.
So if we think about each of the different issues across how people receive better care, we can think about the application of better measurements. That's the question about data generation, and using that data to create better insights and knowledge and take action.
These are all pieces that have been building since the initial investment by Google I think it's time to move forward. Equitable access to care and better outcomes.
MHN: Does the company often work together? Google?
Tristar: So there are a number of projects that were started within Google Health Research and brought to Verily to build the product. Some of this is more like a partnership with a memorandum of understanding between the two companies.
The best example is an algorithm published in Nature by Google Health Research that looks at images of the back of the eye, or the retina, to determine whether a person is likely to have diabetic retinopathy. They also did things that humans couldn't do to demonstrate the usefulness of AI applications, such as classifying whether the eyes were male or female.
But the question for such algorithms is where does it fit into today's workflows and how does it actually impact people? One of the big questions that came up was, this is where our partnership with Verily and their great hardware comes into play. It was actually about how do you get some of the images on the back? eye?
So we built a retinal camera that is almost completely automated. It's called Verily Retinal Camera. This device allows a person to take an image of the back of the eye, on which his AI will be mounted. We used Google's AI, but we've since worked with other companies who are also building other AI models that demonstrate this utility, starting first as a classification diagnostic and then more We are working from a broad perspective. We're excited to be able to apply what we can measure, similar to what Google is doing with things like gender classification.
Although it is not necessarily necessary to make a diagnosis based on retinal images, there may be other diseases that can be diagnosed based on such types of things. So what else can you do once you have the device in place? So it's an area of ongoing debate.
MHN: teeth Med-PaLM What is Verily looking to leverage within its services?
Tristar: We have discussed the utility of Med-PaLM, its model, and how new multimodal approaches can be exploited. We have Med-PaLM and obviously a lot of work is being done on Gemini. So we're certainly exploring what that might look like, but we're not locked into working exclusively with Google.
We are considering other applications. Once you know you have the right data infrastructure in place, you can be agnostic about the dominant (genAI) model. So that's a lot of what we've built that differentiates us at Verily and sets us apart from Azure and Google Cloud and AWS. And what really influences people?
I think a lot of the focus of generative AI applications is on the back-office part. We do some of that work primarily through insurance companies, but the biggest change we can make with these tools is to be visible to people even before they become patients. So how can we help people navigate this crazy thing that is the health care system?
MHN: What are you most excited about within Verily right now?
Tristar: There are a lot of very powerful engineering applications and very difficult problems in the medical field, and Verily decided to tackle them head-on. However, in the past, there were some siled projects.
For example, things like retinal cameras are great products, but they're not built to holistically solve all the different pain points that we see within the healthcare system. We are now in a position to start connecting things in novel ways.
If we can start to show that things are actually mutually reinforcing in multiple different ways, that's where the real value comes from, because people can live healthier, better lives without increasing costs. I think.
Technology has been a major driver of health care costs for decades now, so reducing technology enough in the future could bend the cost curve.