Speak Data, page 9
Again, I’m not saying that the current situation is bad. All these other entities need access to our data because we need data to come up with therapies. A patient with a rare disease is often willing to open-source their data, to share it, in order to help develop a lifesaving therapy. But most of us have so little control of how our health data gets used, who owns it, how long they can have it. I think we need to figure out ways for patients to safely have autonomy over their data and control who accesses it, how it gets shared, how it gets stored, especially as we’re able to collect more data from humans.
Eric Topol
Eric Topol is a physician-scientist at Scripps Research, where he leads the Scripps Research Translational Institute. He is also a prolific author, public speaker, and advocate for individualized medicine using genomic, digital, and AI tools. The author or coauthor of more than 1,200 peer-reviewed articles, Eric is one of the top ten most cited researchers in medicine, and his weekly newsletter is read by tens of thousands of subscribers. In this conversation, he talks about empowering patients with their data; the future of AI in medicine; and communicating scientific research to a larger public.
You were one of the most incisive voices in the medical community during the COVID-19 pandemic, advocating for science and treatments in a very patient-focused way. What’s exciting to you these days?
There’s a convergence of many things, from genome editing to vaccine drugs. Never in the history of modern medicine has there been a time where all these incredible, cutting-edge advances were occurring at once. That’s what gets me excited. We can improve human health more than ever before.
Why are you focused on the patient?
People talk about breakthrough drugs, but the patient is and can be a breakthrough driver of their own health. They are a missed opportunity. Right now, electronic health records exist in multiple provider and health systems. Nobody essentially has all their data. There’s genome data sequences or the microbiome. There are sensors that people commonly wear and use, and more to come. There’s the environment, you know, whether it’s pollution or carcinogens, micro- or nanoplastics. There are layers of data that are not brought together. Ultimately, the patient should be able to aggregate their own data.
But now, with AI, we have the capability to do all that analysis in real time and be a virtual coach to patients. This isn’t going to happen in the medical system; it has to be done at the individual level. It’s waiting to happen, and when it does, it will give patients more agency. Patients can interact with their results and the analysis of their results with their physicians. It will be a very powerful tool in the future, but it relies on multimodal AI.
Do you imagine a near future where one can go to the doctor with all this aggregated data, and the doctor will be able to make sense of it? It’s a lot about educating physicians too, isn’t it?
Yes. One of the layers I didn’t mention is the corpus of medical literature, which no doctor can fully digest. As much as I aspire to know the medical literature, no human being could keep up with the massive knowledge-base expansion that’s occurring every day. So, to get to your question: Progressive physicians will realize that we’re past the point of human beings as the experts. Doctors, who were historically the purveyors of information, will now provide guidance and wisdom and experience and oversight. We want patients to have all their data and have help from AI to validate it. Of course, this has to be done carefully. But as long as it’s accurate in helping patients and promoting health, it’s a big win, because it decompresses the doctors’ work.
What you’re bringing up, though, is this tension. Since Hippocrates, doctors have ruled the roost. They ran the show and had total control. Now they’re being undermined. They’re no longer the supreme being of sorts. Because the truth is that the patient knows best. Younger physicians and progressive physicians of all ages will realize that that’s where this is inevitably headed because data is eminently portable, and AI tools to analyze people’s data are building quickly and gathering tremendous momentum.
I think this is a vital step. When a doctor sees a patient, it might be once a year. It’s not in the real world. Whereas we now have tools to track, if need be, every five minutes, or even more continuously, certain metrics like glucose or heart rhythm or you name it. This is medicine that’s continuous. It’s far more insightful than the artificial white coat syndrome where you go in and you’re scared of what the doctor is gonna do to you. This is an upgrade, a profound upgrade, of medicine’s capability. But you’re right that it’s both data dependent and physician dependent.
This entire conversation is obviously inflected by what we learned from COVID-19. As a medical professional, as a researcher, as a physician, what did COVID-19 teach you? We’re always wondering, if we could zoom back to March 2020, what we would do differently.
There’s a lot in that question. I mean, a lot. We had one of the greatest triumphs of biomedicine history with the march from the sequencing of the virus to the vaccine. That was unparalleled in history. But so far, it’s just been a one-off. What if we had a similar success with long COVID? What if we had it for nasal or oral vaccines, and many other things that are desperately needed?
Then at the same time, there are these anti-science, anti-vax people. How can you question the need for a mask to protect against respiratory illness? Unfortunately, there hasn’t been any aggressive response to them, and they’ve just grown, gotten more organized, more financially backed. We’ve been really hurt by this. They’re the same people who are saying that there’s no such thing as long COVID. It’s just despicable.
You were a coauthor on one of the first and most formative long-COVID studies, which was published in January 2022.1 Notably, your coauthors weren’t fellow academic researchers, but members of a patient advocacy group. They’re researchers, but they are also dealing with long COVID themselves in a personal way. Historically, it’s an unusual thing to share credit and expertise in that manner. Did you get pushback?
Actually, I didn’t. There may have been some push-back, but I didn’t see it because I block those people online! But I actually can’t imagine how you could write a paper about long COVID without a coauthor who is experiencing long COVID. In fact, that’s been part of the problem from the get-go. We have tens of millions of people around the world affected by long COVID, and we have the pseudo-experts who have never experienced it themselves. I’ve been a proponent of citizen science for a long time—long before the term ever came alive. And this was a perfect example of why we need such collaborations. Even though my three coauthors on that paper have long COVID and I don’t, I could still be an external questioner, trying to get the right balance on the topic. We had, I think, an exceptionally important collaboration in putting it together.
I was actually the one the journal invited to write the piece. I told the editor, “I can’t write it myself. I don’t have long COVID, thank goodness. But I do know some people who have it who are exceptionally sharp.” That was the formula. We need more of that.
The paper was also incredibly accessible. Even a nonexpert could understand it.
That’s a great point. I mean, these days you can take a paper and put it through AI and make it understandable at any level of education. We didn’t have that at the time. But yes, it’s important to write in a language that everyone can understand. If you’re only trying to impress your peers, you’re in a microcosm, and the work is never going to have the impact it should.
How do you do that? How do you reach more people? It’s surely more than just dumbing down the language.
Stories are big. You need to tell a good story. But graphs, to me, are also so powerful. When you do that well, you’re also telling a story. It may not have the emotional charge, but it can quickly reveal that something big is happening. I rely a lot on simple graphs. I think they’re extraordinary. Qualitative data can be powerful. But hard data—now that’s compelling. It’s the tool I favor most.
How do you define hard data, or just data itself?
I don’t know if I can get more atomic-level than that. Data is data.
It just is.
Of course, there’s good data and bad data. But to me it’s the equivalent of an atom.
We like that.
NOTES
1 Hannah E. Davis, Lisa McCorkell, Julia Moore Vogel, and Eric J. Topol, “Long COVID: Major Findings, Mechanisms and Recommendations,” Nature Reviews Microbiology 21 (2023): 133–46, available at https://www.nature.com/articles/s41579-022-00846-2.
4
Machine Thinking: Data and Technology
Machine Thinking
We’re typing these words in late autumn of 2024. As we write, we’re thinking about data and technology—what they are, and what they do for us. Why they matter now, and how they will evolve in the future. Data and technology are so intertwined, and together so completely enmeshed in daily life, that it’s difficult to apply criticality to our thoughts. We reach for metaphors in hopes of making sense of it all: Data is like a power source, the computational jet fuel that animates technology’s potential. It’s the informational lubricant for digital processing. It’s the instructional code, telling the machine what to do. If technology is the frame, data is the art within it.
That sounds pretty good. We type it up.
On second thought, maybe we’ve gotten it all wrong. Data doesn’t tell the machine what to do—in fact, it’s the opposite. The machine tells data what to do. By providing the frame, technology shapes the data to fit inside it, sculpting binary code into a form that can be read, understood, and acted upon by machines (and the humans who operate them). Technology is a tool—increasingly, a dizzyingly omnipotent one—for manipulating data, for organizing it and slicing it open to see what’s really inside it. How we use this tool, and do so with an eye toward ethics and sustainability, seems to be the most pressing question.
That sounds good too.
But then we stop ourselves again. Because by the time these words make their way from our computer, to the printed page, and into your hands, they may well be obsolete. When it comes to data and technology, bankable descriptions—much less predictions—have always been a futile proposition. Over the past century the world has witnessed exponential growth in technological innovation, an annual doubling in computational power famously known as Moore’s law. Moore’s law was first theorized in 1965, when the full promise of personal computing and digital transformation was imminent, but still beyond the horizon. Today we stand on a new horizon line of a fourth Industrial Revolution, and many believe that Moore’s law no longer applies as originally conceived.1 Artificial intelligence represents an entirely new paradigm for computational reasoning. Big data, on humanly incomprehensible scales, operationalizes warehouses of knowledge with ease. Quantum computing bends the rules of physics to outdo the performance of classical computers one hundred million to one.
What does this all mean for us? As society hurtles headfirst into this new reality, we believe it will be essential to return to first principles. Let’s remind ourselves what technology really is, why it matters in the first place, and how it can improve our lives and the lives of others. This doesn’t mean being afraid of technological innovations, but instead embracing their possibilities with a necessary level of criticality and discernment.
Across our careers, we’ve used technology in many ways, engaging with new forms of hardware and software to create data-driven experiences for many types of users. Yet ironically, at the core of our practice has always been an emphasis on the least technologically advanced tool we have: the human hand. For Giorgia in particular, the creative process usually starts with drawing, sketching, and working ideas out on paper. One can draw with data in the mind, but with no data in the pen, to understand what is contained in numbers and in their structure, and how to define and organize those quantities for greater insight.
Today the word digital means any machine-readable code constructed in a binary language of ones and zeros, but its etymological origin actually refers to digits, as in fingers—our most human, individualized way of counting and communicating information. A foundation of analog knowledge and human agency helps us to remember that technology, no matter how powerful or awe-inspiring, is ultimately just a tool—a means to an end rather than an end unto itself. Just as we use data visualization to communicate meaning, we can use technology to bring the user into closer proximity to that meaning, to amplify its emotional resonance or underscore its real-world impact.
Back in 2018, this was Giorgia’s goal for a collaboration with Google News Lab, in which she and her team utilized augmented reality and an unusually poetic form of information design to connect the user with a shared sense of optimism for the future. In a world increasingly consumed by polarization and despair, the project was an attempt to tap into more positive ideas around progress and possibility, and use data from Google searches—the world’s great barometer of human interests—as a doorway into these ideals. This very nontraditional AR experience was named Building Hopes, and it allowed users to build “data sculptures” in virtual space based on what they believed mattered most to humanity’s future. Each sculpture’s size and color was specified by the user to indicate thematic focus (immigration, sustainable development, technology, et cetera) and intensity of interest.
These colorful, floating sculptures functioned as personal totems of belief, metaphorically embodying the desires and dreams of each user. Users could arrange their sculptures next to those of others to see how their values “stacked up” to the rest of the world. Do we share the same dreams? How much optimism is there for the things that matter to me? In its unique way, augmented reality allowed these questions to take physical form, to have real shape and presence, and manifested as human-scaled structures in our spatial environment. Technology was the tool, but hope was the ultimate reaction.
Sometimes the solution isn’t more technology, but less. This was the lesson when Giorgia and her team collaborated with Samantha Cristoforetti, the European Space Agency astronaut who made history in 2014 by becoming the first Italian woman to venture into space. In the lead-up to the mission, a chance conversation between Samantha and Giorgia on Twitter led to an interesting project: an “extraterrestrial” social network that would connect us earthlings to Samantha, and Samantha to us, as she orbited the globe in the International Space Station (ISS) for 199 days. Giorgia called it Friends in Space, and it served as an unlikely interface for very human conversations.
The project developed after a key realization about how technology, at least in this situation, might best facilitate connection between Samantha and those back on Earth. After all, a mission to the ISS is a feat of technological accomplishment. Each flight also generates terabytes of data—more quantitative material than we know what to do with. But the data and technology that would power the space shuttle’s ascent was in fact far less profound than the simple miracle of launching a living person into the cold depths of outer space. Samantha would be up there: a human being in a flying metal box. On a clear night, one could see her with the naked eye, and although she was just a tiny flash of light in the dark sky, that instantaneous connection would prove to be more moving than any statistical readout could ever hope to be.
Building Hopes, 2019
Friends in Space, 2014
Particle view, Plastic Air, 2020
Trash view, Plastic Air, 2020
This form of interstellar poetry became the inspiration for the design. With a simple digital gesture within the interface, a person could say “hello” to Samantha when she was in orbit overhead, and “hello” to all the other people around the world who were online at the same time. These greetings were rendered as yellow bursts of light, pinpointed to the user’s exact location on a map. Naturally, Sam could say “hello” back to the world via Twitter from the ISS, which also allowed users to see her own position along the ISS’s flight path. During its first three weeks online, Friends in Space hosted more than two million interactions, merging the physical, the digital, and the emotional in moments of joy and connection.
Friends in Space used technology to bring people together, but we can also use technology to shine a light out onto the world and reflect our place in it. In 2019, our team at Pentagram worked with Google Arts & Culture to create a digital data visualization experience intended to raise awareness about microplastics, the tiny particles of plastic waste that are expelled by everyday household items like clothing, furniture, packaging, and appliances. Microplastics are a pressing environmental issue, but they also can be a somewhat obscure concept. Most are smaller than the width of a human hair, and thus invisible to the human eye—literally out of sight and, thus, out of mind. As we learned more about microplastics, we became shocked and alarmed by their pervasiveness. While most concentrated in and around cities, microplastics have been found in remote locations around the world far away from human settlements, indicating their ability to travel long distances in air or water.
Using a simple web app and real data of microplastic incidence culled from scientific studies and conversations with researchers, our solution bridged the gap between the visible and the invisible, the tangible and the abstract. Called Plastic Air, the experience consisted of a speculative window onto a data-driven approximation of the plastic particles that float around us all the time. Users could drop identifiable objects like a lawn chair, a T-shirt, a cigarette butt, or a couch cushion—all containing plastic polymers—to “pollute” the sky, and then witness those items breaking down into millions of microplastics, like colorful confetti scattered to the wind. The result was, by design, both beautiful and horrifying. Here, technology gave us the power of sight. Plastic Air let us peer into a world that is normally opaque but nevertheless real, revealing the hidden impact that we as plastic consumers are having on the planet that we all share.
