“I have the distinct pleasure of talking about one of my favorite topics today, revealing my inner geek, which is data,” says Amy Compton-Phillips, Executive Vice President and Chief Clinical Officer for Providence St. Joseph Health and NEJM Catalyst Care Redesign Theme Leader.
“I don’t know about you, but when I look at my news feed, and I open the front page of a paper, or I turn on TV, I feel like the world is going to pot. I don’t know what’s happening, but it’s kind of all falling apart,” she says.
“The reason things make the news, and the reason we get urgent headlines and dramatic information fed to us, is because their perturbation’s off normal,” Compton-Phillips adds. If you want to understand what isn’t normal, you need a baseline for what is. “To understand that baseline, you need data to say what is normal. It is a conscious thought because it’s not going to come into your news feed, because normal isn’t news,” she says.
Compton-Phillips recommends Max Rosen’s Our World in Data for learning about what’s normal. For example, when comparing life expectancy at the time of Compton-Phillips’ birth to the time she began practice to today, one can see a remarkable difference.
“The world is getting better. The world is getting healthier,” Compton-Phillips says. “It’s not just getting healthier and better in rich western industrialized nations, but it’s getting better everywhere — low income, moderate income, high income, across the globe.” Life expectancy in places that in the past experienced deprivation now are seeing equivalent life expectancy to what the U.S. saw in the mid-20th century. Those improvements aren’t just from sanitation and immunization, but they’re also from reductions in mortality from things like HIV, pneumonia, heart disease, cancer, and accidents.
“Everything is getting better,” Compton-Phillips says. If you compare what the average human’s life is like today — what’s normal today — to what the average human’s life was like at any other time in history, today is better. “If you had to choose a time on the planet to be born, if you wanted to optimize your chance of having a life where you can live and thrive and be educated, and have a healthy, fulfilling time here on the planet, you would choose now.”
“That is not to say that we’re perfect, because I have to tell you, we’re not, and that’s why your news feed looks like your news feed looks like. But we’re better than we were,” she adds. “And a big part of why we’re better than we were is because we figured out how to gather and use data.”
Data to Illuminate, Data for Discovery
“We gather and use data because people are curious, and so we had to describe how does the world work, that’s data to illuminate, and because people are fairly altruistic, because we live in a group, we’re a social animal, and we want to make life better for not just ourselves, but for those around us, and so we also use data for discovery on how we might change things to make things better,” explains Compton-Phillips.
“And now, it’s a new dawn because we don’t just have data that we collected ourselves, but we have all this incidental data from the digital dust that comes along with life in the 21st century. Now we have opportunistic information, so we can start adding new sources of information into this quest for a better world.”
Stepping back to provide some historical context, Compton-Phillips talks about how Leonardo da Vinci, who lived during the Italian Renaissance, used observation, curiosity, and an ability to draw to collect data, make inferences, and discover things that today may seem obvious but at the time were new. For example, he found that in nature, when things like rivers, trees, or blood vessels branch, the sum of the area of the branches equals the sum of the area of the trunk. That rule always existed, but nobody knew it until da Vinci used data to illuminate and describe it.
In another historical data story, John Snow, a 19th-century English physician, believed that contaminated water caused a rampant cholera outbreak in London. He stuck pins in a map of where cholera deaths occurred, tracking them to their source: a water pump near Broad Street, where a woman had washed out her child’s diaper, contaminating the water in the well. Snow got the water pump handle removed, and the cholera epidemic dissipated.
Today, we constantly use two types of data: data to illuminate, and data for discovery and change. We can use illuminative data to help prove that a program is working, e.g., pointing to a food program that has reduced deaths from malnutrition in developing countries to say, “Keep doing what you’re doing.” This isn’t a change but a display of information that everyone should know.
Data for discovery is how we conduct clinical trials today, with an “if then” hypothesis that “if then we do this, perhaps we can change that.” The data sets are getting bigger, but the process is the same: gathering data to see if we can change things for the better.
The one thing that’s new in today’s data is the digital dust Compton-Phillips mentioned earlier.
“How do we take this amazing stat that 90% of the world’s data has been developed within the past 2 years [originally from 2013, via IBM and SINTEF, though sources repeat that statistic year after year]? That’s incredible,” says Compton-Phillips. “The volume of information that comes in today from everything from your Fitbit to your Twitter feeds to your refrigerator telling you you’re low on milk because it’s connected to the Internet of Things, all that data now is available for different information.”
“How do we take the human attributes of curiosity and altruism, put it together with this function of data to illuminate and data for discovery, mix it in with both intentionally collected data and opportunistic data, and create a future where we continue on this trajectory of making life better year after year after year?” asks Compton-Phillips.
She proposes three ways to get there:
- Asking the right questions
- Adding art to science
- Changing the equation
Asking the Right Questions
Compton-Phillips tells the story of a friend who woke up one night feeling ill. After calling 911, he made it into the ambulance and to the emergency room before collapsing from cardiac arrest. After 45 minutes of chest compressions, the doctor asked the man’s wife if they should call the code. She asked what his chances were of waking up and still being her husband, and the doctor said, “I don’t know.” But that is knowable information.
A year later, Compton-Phillips’ friend is healthy and lives a normal life. Understanding what the odds are, not of living to discharge from the hospital for an in-hospital code, which we measure today, but of knowing the chances of a person being fully functional, of having a normal life years later, is knowable if we ask and answer the question: Will health care make a life better?
“We can know that today and we need to know that,” says Compton-Phillips. “How can we give people probabilistic information, so in real-time the doctors can help make educated decisions, not make the decisions for the doctors, but have the information so that they can make educated decisions? How can patients put their preferences, values, and beliefs into that decision matrix so that together, they can decide on the appropriate course of action when and where they need that information? We have to ask the right questions to be able to present that information at the right time.”
Adding Art to Science
James Lind, a British naval surgeon in the mid-1700s — a time when scurvy killed more sailors than battle — ran the first randomized clinical trial. He divided 12 sailors on his ship who had scurvy into six groups and gave each group purported cures: lemons and oranges, seawater, vinegar, etc. The two he gave citrus revived and were back on the job within a week. The others didn’t fare so well.
Lind published his trial in a 400-page tome “full of words like putrescence and springiness of” — basically gobbledygook, as Compton-Phillips puts it, and nothing changed in 40 years because no one read that information.
“How do we take the vast reams of information we have today, use that information, and display it in intuitive ways so that people can make decisions?” asks Compton-Phillips. We must add art to science. For example, with a Napoleon’s March–style graphic, the Providence St. Joseph clinical analytics team asked: If we have a best practice alert, does it fire on the right people? If so, can people act on it? Does it get the right consult and the right change in behavior? Where that information falls off, do I need to change the alert, the workflow, or the workforce?
You can answer these questions from the way the data is displayed. “That takes thinking creatively about how you put huge reams of information into something actionable,” says Compton-Phillips. “Think about that for a second: that James Lind, 12 people, 400 pages. Or clinical analytics, 91,000 people, one graph, which gives you a lot more information. That’s adding art to science. That’s embedding creativity into the way we talk about data.”
Changing the Equation
During the Italian Renaissance and the European Enlightenment, people of means or with patrons of means conducted research and collected and used data. That went on for a while until people figured out that we should have academic programs where people can think about big ideas, funded through grants and government programs. And then in a capitalist society, we moved to targeted, for-profit problem-solutions from pharma, device manufacturers, and the Internet, with people asking how we are going to use all this information.
“People are using it for the profit motive, and we need a way, somehow, that we can reward as we move forward,” says Compton-Phillips. “What is the model that we can continue to help people get the intrinsic reward of solving big problems, and be able to put food on the table?”
We haven’t quite figured that out, and the calculus keeps changing, but we need to figure it out. “I don’t know about you guys, but I am hoping that my grandchildren’s grandchildren, the people who come after me, are going to be able to look around and say, ‘This is the best time. If I could choose anytime on earth to be here, to be on this planet, now would be the time,’” Compton-Phillips concludes. “If we want our grandchildren’s grandchildren to be able to say that, we need to keep on the path that our forebears started with the Enlightenment, using data to drive change so that they could deliver on the promise of health for a better world.”
From the NEJM Catalyst event Provider-Driven Data Analytics to Improve Outcomes, held at Cedars-Sinai Medical Center, January 31, 2019.