Health care is extremely complex, but artificial intelligence and machine learning are evolving in ways that can help consumers, caregivers and entire communities, said Michael Howell, Google’s chief clinical officer.
wednesday in Mayo Clinic Platform Conference, Howell discusses the ways AI and ML are serving healthcare and some of the lessons he’s learned since he started working at Google. Most of his experience was as a pulmonary and critical care physician and health system leader, but he joined Google about five years ago.
“Since working at Google, I’ve learned something about artificial intelligence and machine learning that I didn’t know before,” he said.
His first piece of advice to conference attendees is that AI and machine learning are “math, not magic.” That requires doubts about how it will affect health care, and it’s a long way off, he said. Like other forms of healthcare, it requires the publication of peer-reviewed research.
“Math is harder, we want to get it right, we think peer review will help… We want to show our work, we want to show our work over time because it’s a really important tool, ‘ Howell said. “This is part of our understanding of how technologies like AI and ML are coming into healthcare.”
Another important lesson is that ML learns from examples, unlike traditional programming that follows explicit rules such as “if-then” statements. An example given by Howell is spam. The code of the traditional program clearly states what types of emails need to be placed in the spam file. The ML program looks for examples of unopened emails and continuously adjusts to reduce errors.
“It changes what we can do,” he said.
How artificial intelligence can help today
Howell identifies three groups of people where AI and ML can help in healthcare today: consumers, caregivers, and communities.
When it comes to consumers, the goal is to help every day.An example he gave was Google Fitwhich allows people to use their phone’s camera to check their heart rate and breathing rate.
To measure their breathing rate using the app, consumers simply place their head and upper body in view of the front-facing camera and breathe normally. In order to measure their heart rate, consumers need to place their fingers on the rear camera.
“We want to help people in their daily lives,” he said. “Not only in the doctor’s office, not only in the hospital, but in their daily lives.”
Artificial intelligence and machine learning can also help ease the burden of healthcare when patients are in a healthcare setting, whether it’s doctors, nurses, payers or anyone else in the field, he said.
For example, Google is working with the Mayo Clinic to simplify the process Radiation TherapyWhen planning radiation therapy, clinicians perform a step called “contouring,” in which they identify areas of cancer and nearby healthy tissue that may be damaged by radiation. This procedure can take up to seven hours per patient.
Google’s collaboration with Mayo Clinic includes research into how AI can be used to shorten treatment times and assist in contouring.
Finally, Google is working hard to use AI and ML to improve community health, and has introduced some features to help treat infectious diseases during the Covid-19 pandemic.
For Covid-19, Google introduces features to improve research, including a dataset search trends about disease symptoms to help researchers understand the spread of the virus and how it affects communities.Google also launched tools to help vaccination. Its Covid-19 Vaccine Access Dataset helps find places where people have difficulty accessing vaccination sites and where pop-up clinics or new locations are needed.
“We do see a lot of progress in the research and public health community using these aggregated, anonymized datasets to advance research efforts around Covid-19,” he said.
Howell also provided insights into some of the areas Google is currently researching that may be discussed at the next Mayo Clinic Platforms conference.One of them is named Motherdesigned to identify people in crisis and connect them to the best resources when they seek help on the internet.
“I think a year from now … we’ll be talking a lot more about these upcoming new technologies,” he said.
Photo: ipopba, Getty Images
