On the opening day of Patient-Centered Oncology Care, Robert Groves, MD, of Banner | Aetna, discussed the use of psychographics to predict health behavior in local populations.
Why do some people always get their annual physical and faithfully take their medication, while others avoid the doctor until they end up in the emergency department (ED)? How do we predict which person—or at least, which groups of people—are likely to be the costly ED patients?
There are data sets that can tell us these things, according to Robert Groves, MD, executive vice president and chief medical officer at Banner | Aetna, based in Phoenix, Arizona. And while it might seem a little scary—these data can be used in harmful ways, to be sure, Groves said—health systems can leverage artificial intelligence (AI) to use such data—from the brands consumers select to their achievements and preferences—to predict their health behavior.
Groves spoke about using the power of AI to leverage these data, called psychographics, in a talk during the opening day of Patient-Centered Oncology Care® (PCOC), held in a hybrid format from the Omni Nashville Hotel. Groves appeared in person.
“When I say a term like ‘precision medicine,’ or ‘precision health care,’” what's your mind immediately go to? For most of us in this room, it probably goes to genetics and Cas9, CRISPR-based editing, genetic predictions, and that's fair,” Groves said. “But what if I said personalized health care or personalized medicine?”
The data that are collected on individuals, psychographics, "is a way of describing why people make the decisions they make,” Groves said.
Distinct from demographics—things like age, gender, or income level, which he said measure facts about a person in the moment—psychographics tend to measure characteristics that hold up over time. When psychographics are laid over geography the resulting “microcultures” are very local—but these are important to health systems, because that’s how health care markets operate.
Groves also discussed the collaborators Daniel Kahneman and Amos Tversky, who created the field of behavioral economics by upending the idea that human beings always acted rationally—they introduced the idea that in fact, humans make plenty of illogical choices and take short cuts in their decision making.
Separately, Kahneman and a one-time student, Richard Thaler, won Nobel Prizes for what is now called called nudge theory—the idea that behavior can be changed slowly by a series of small changes or “nudges.” Health systems are paying attention to how they can harness psychographics and deploy the data into nudges, giving them a chance to direct populations toward healthier choices, Groves said.
“Now, I'm not suggesting that you can take psychographic profiling down to the individual level and influence the behavior because, by definition, it is defining microcultures, not individuals,” he said. “But it can be extremely useful if you are a health system, and you're trying to plan strategically for the future.”
There are already several examples of nudge theory in practice. Groves is an unpaid advisory board member with the AI company CareCentra, which is developing strategies for health systems. In academia, Penn Medicine’s Nudge Unit has published work on using wearables to improve sleep and using text messages to boost flu shot rates. In the commercial sector, there’s the weight loss company Noom, cofounded by Artem Petakov, who studied under Kahneman at Princeton University.
An important feature of AI in these circumstances is that it is not judgmental, Groves said. He pointed to one example of a woman who was not taking her medication, but the AI in making decisions noted that she was having a Coke every day and recommended she put her medication with the Coke so she would not forget. “As a physician, I might say, ‘well, the first thing we have to do is get rid of that Coke,’ but that’s not really ideal, is it?” Putting the medications near the Coke helped the woman become compliant with taking them.
The research Groves presented showed people fall into 5 broad categories of health care decision making, with 26.7% labeled “willful endurers,” meaning they mostly go about life unfocused on their health, and not much will change their attitude. Groves said a key discovery of this work is that for most people, their predisposition for health care decision making is hard wired by age 18, “’and it takes a lot to move people—and the one thing that does more them is a diagnosis of cancer.”
There’s great power to these data, Groves said. Historically, pharmaceutical companies primarily used psychographics for marketing; understanding that the data could also be used to help people stay on medication or eat healthier food could allow health systems “to actually manage this beast and turn it towards good.”