Integrating diverse kinds of data key to precision health
June 1, 2017 – STANFORD. Environmental, behavioral and social data could help researchers and clinicians prevent disease in whole populations — not just diagnose and treat disease — according to speakers at Stanford’s fifth annual Big Data in Biomedicine Conference.
The conference, held in May 24-25 at the Li Ka Shing Center for Learning and Knowledge, drew about 500 attendees and 2,000 remote viewers. Titled “Big Data in Biomedicine: Transforming Lives Through Precision Health,” the conference focused on precision health in action.
A series of talks and panels highlighted, for example, the National Institutes of Health-funded Precision Medicine Initiative, which will use data from a million volunteers to build a future where prevention and treatment are tailored to individuals based on their genes, microbiome, health histories lifestyles and diet, and the Chan-Zuckerberg Initiative, which aims to rid the world of disease by 2100. Speakers also discussed applications of big data to cardiovascular disease and cancer; the use of artificial intelligence in imaging; and digital health and technology.
The key to transforming lives through precision health, experts at the conference said, is the integration of diverse kinds of data sets, including sequencing and imaging data, gene expression data and also behavioral data, such as that from fitness trackers. Finding ways to combine and explore such cross-disciplinary data sets will be key to the focus on prevention, they said.
“For years, health care really has been about sick care; it’s been about treating severe acute diseases or their chronic manifestations,” Lloyd Minor, MD, dean of the School of Medicine, said during remarks on the first day of the conference. “And there’s been comparatively little attention either in research or care delivery on prediction and prevention. But that’s all changing today, because of the work being done in this room, because of the work being done at Stanford.”
Advocating greater focus on prevention
Eric Topol, MD, a professor of genomics at the Scripps Research Institute, gave a passionate talk in which he criticized U.S. medical care and advocated for a greater focus on prevention.
The key, Topol said, is moving “from where we see people as all the same, at the 50,000-foot level — which is the way medicine is practiced and the way we give out drugs and diagnoses and do screenings — to ending this concept of an ‘average person,’ who doesn’t exist.”
“Where we are in 2017 is remarkably primitive,” Topol said, quickly reviewing a series of facts about U.S. health care. Mass screenings, for example, often do more harm than good. Of every 10,000 mammograms, Topol said, only five actually benefit the patient, while thousands are false positives that can result in actual harm. He also cited the high rate of medical errors and misdiagnoses, despite an annual health care budget of $3.4 trillion in 2016
“We have to be able to do better than that, you would think, if we can define each individual, which we haven’t had the tools to do — until now,” he said.
One way medicine can improve is by mapping human beings using approaches that resemble the geographic information system maps that geographers, mapping apps and urban and county planners use to understand different layers of information about the world, Topol said.
Today, biomedical researchers have the capacity to quantify and track a person’s environment, behavior, dynamic physiology and proteomics through time and across whole populations, Topol said. Sensors can track the data while algorithms can integrate and interpret multiple data streams.
Such data shouldn’t be understood like the black box approach of Netflix guessing which movies you’d like, though, said Hemant Taneja, PhD, managing director at the venture capital firm General Catalyst, who spoke during a panel discussion on digital health and technology. We need a higher bar for biological data such that not only can we predict the consequences of certain combinations of factors, but also know the biological mechanisms that cause those effects, he said
“Today you can accurately track your blood pressure through your watch,” said Topol, adding that it’s possible to see in real time how your blood pressure reacts to heavy traffic or a heated discussion with your spouse. And it’s now so easy to monitor blood sugar levels that healthy individuals who don’t have diabetes can monitor their blood glucose as a preventive measure.
Topol also described a portable ultrasound device that can create an image of the heart beating. He said it makes an old-fashioned stethoscope obsolete. “Why would you listen to ‘lub-dub’ when you can see everything?” Topol said.
Preventive health at population level
Apply such measurements to 10,000 people, as in the Baseline study by Verily, Duke and Stanford, or a million people, as planned for Precision Medicine Initiative, and you can start to see patterns, experts at the meeting said. Both researchers and clinicians can ask and answer specific questions about which factors are associated with which outcomes.
Such population-level approaches to preventive health care are inspiring more interest in health care inequity and biases in research, some speakers said. Although about two-thirds of people living in the United States are of European ancestry, participants in medical research trials were 96 percent European-American in 2009; by 2016, it was 81 percent European-American, according to Usha Menon, PhD, RN.
Menon, a professor and associate dean for research and global advances at the University of Arizona, suggested a number of strategies for increasing diversity within biomedical research, including, for example, engaging citizen advisory boards in the design of studies and targeting messages. “We have been telling people to stop smoking — for how long?” she asked. “The key is to target and tailor to culture, to what is most relevant to that individual.”
Topol and cardiologist Robert Harrington, MD, professor and chair of medicine at Stanford, discussed the possibility that letting data streams and algorithms do much of the work of examining, monitoring and diagnosing patients could give practicing physicians more opportunity to form the intimate bond that doctors once had with patients. Harrington brought up Stanford’s Presence, a center whose focus “is try to bring some of the intimacy and humanity back into medicine,” he said.
Verily, one of Alphabet’s life sciences units, recently launched Project Baseline with the goal of enrolling 10,000 participants who will share their biomedical data. A long-term goal of the project is to set up scalable and standardized tools for acquiring, organizing and analyzing data, said Jessica Mega, MD, MPH, Verily’s chief medical officer, during a talk at the conference. Formerly called Google Life Sciences, Verily has a variety of projects in development, from wearable sensor devices to big data studies.
“We also want to create a test bed for a number of new tools and devices that are out there,” Mega said. For example, glucose and atrial fibrillation monitoring devices can provide a stream of data that’s valuable to patients and their doctors, valuable to researchers and valuable to whole populations that can eventually benefit from the resulting insights.
“But that’s the known signal,” Mega said. “As each of us is sitting here, in this room, we are giving off our heart rate and our galvanic skin response. There’s digital exhaust all around us that we’re not capturing. Some of it may be actionable, some of it may not be. But until we look we won’t know.”
Project Baseline, Mega said, will serve as a test bed to think about the next generation of tools needed to understand our human physiology and human variability. The holy grail, she said, is to get in front of disease to know ahead of time that an individual needs help taking preventive measures.
Stanford University President Marc Tessier-Lavigne, PhD, exhorted the audience, “as hard as you are working on these problems, double down again.”
“Our charge, our responsibility,” he said, “is to make sure we get to precision health tomorrow and not 10 years or 20 years from now. We know it will be a reality eventually. Our job is to make sure we accelerate the development and implementation of precision health.”
Source: Stanford University