SIOUX FALLS, SD — The power of computer-driven artificial intelligence has held tantalizing promise as a tool to help doctors diagnose and treat disease, but so far it hasn’t met the high expectations.
Elliot Green and his co-founders of Dandelion Health Inc. decided to investigate why artificial intelligence in the clinic had so far yielded disappointing results.
“We saw it had the potential to do so much good and couldn’t understand why it never took off,” said Green, CEO of the company.
“We found out it was a data problem,” he said, “it wasn’t a technology problem.”
It turned out that the massive volume of information needed for computers to learn to perform tasks was not enough, a lack of data that was stifling innovation in clinical applications of artificial intelligence.
So Green and three partners formed Dandelion Health to provide a data platform that aggregates big data to enable researchers to develop new artificial intelligence applications for the clinic.
They quickly sought out health systems as partners to provide the data, and Sanford Health was one of the first to step in, along with San Diego-based Sharp HealthCare. Sanford is the nation’s largest rural health system, with a service area of 250,000 square miles.
“We believe this collaboration with Dandelion and Sharp will strengthen our means of providing care and improving lives,” said Kent Lehr, director of business development at Sanford.
“We see a great opportunity to leverage data to create innovative clinical tools, products and solutions,” he said.
Sanford’s participation ensures that rural populations will not be ignored in datasets used to develop artificial intelligence applications for clinical use.
To date, some of the leading examples of artificial intelligence are found in radiology. Computers are good at looking at pixelated information in medical images, Green said, and can sometimes spot things the human eye can’t.
With enough data, the potential to help doctors diagnose and treat patients with artificial intelligence technology is great, Green said. “In reality, we’re barely scratching the surface,” he said.
The idea behind the search for clinical applications of artificial intelligence is simple: “Where can a machine do what a human cannot?” said Green.
The clinic’s artificial tools should help, not replace, doctors, he said. “That’s what it should always be, a raise.”
Dr. Jeremy Cauwels, Sanford’s chief medical officer, agrees that artificial intelligence holds promise and will benefit patients.
“For me, the real benefit of this is identifying pathways that may not be obvious to a physician,” who lacks the time and isn’t as well equipped to sift through huge amounts of information, did he declare.
Computers, on the other hand, excel at analyzing masses of information. “Right now a lot of these projects are in what I call the infancy stage,” Cauwels said. “We don’t have a large database to train the computers on.”
When powered by data trucks, computer intelligence grows. A famous example from 2012 is the development of an artificial brain by Google. After scanning 10 million YouTubes, he was able to recognize a cat – a skill he achieved with 75% accuracy after reviewing more images over three days.
Sanford has doctors and researchers with ideas for AI applications, and they will have access to the combined data provided by Dandelion.
Sanford will also have priority in applications derived from the aggregated data.
“We will kind of have the first opportunity to deploy these tools,” Lehr said.
The partnership, he said, offers the opportunity “to try to bring the promise of AI” – short for artificial intelligence – “to reality”.
All patient identifying information is removed from the data provided to Dandelion, where it is stored securely in the cloud, where it can be analyzed by the developers, but cannot be downloaded or copied without the express permission of the health system, Green said.
“We are not interested in selling our patient data,” Cauwels said. “We keep the data. You have to get into our “sandbox” and do it our way. »
Protecting patient privacy in the use of data is a primary concern, Lehr said. “I can say without a doubt that the No. 1 priority we’ve had in the conversations we’ve had is patient privacy and its protection,” he said. “We can keep control of this data at all times.”
Individual patient information, which is not available to those using the Dandelion database, is of no use to developers, Green said.
Dandelion is in “active discussions” with other health systems, with the aim of recruiting five partners, he said.
AI developers will soon have access to the dataset, as soon as a month or two away, Green said.
“We have people waiting to use it,” he said.
In recent years, IBM announced Watson Health, an artificial intelligence initiative to effectively identify symptoms of heart disease and cancer.
Sanford University has announced that it is working on an artificial intelligence-assisted support system to detect behavioral changes in elderly patients who live alone or in intensive care units.