Health eHeart Study: Using Big Data to Treat Heart Disease

reprinted from Issue 17, Fall 2013 of Frontiers of Medicine (PDF)

From left: Gregory Marcus, MD, Jeffrey Olgin, MD, Mark Pletcher, MD, MPH, co-principal investigators of the Health eHeart Study

Heart disease affects one in three Americans, and is the number one killer in the country – claiming more lives than all types of cancer combined. Now the UCSF Division of Cardiology is harnessing the power of smartphones, mobile health apps and other technology to learn more about what causes heart disease, and to develop better strategies to prevent and treat it. The Health eHeart Study is an ambitious initiative, which will gather “big data” from up to 1 million participants worldwide.

The study’s team of physicians and researchers, led by co-principal investigators Jeffrey Olgin, MD, chief of the Division of Cardiology, Mark Pletcher, MD, MPH, a cardiovascular epidemiologist, and Gregory Marcus, MD, the director of clinical research for the Division of Cardiology, will analyze this wealth of information. Their ultimate goal is to reduce deaths from heart disease.

Olgin sees the Health eHeart Study as a “Framingham 2.0” – a modern take on the Framingham Heart Study, the landmark investigation that began in 1948 and recruited more than 5,000 adults in Framingham, Massachusetts to identify the common factors that contribute to cardiovascular disease.

“With the Framingham Heart Study, we got most of the low-hanging fruit, learning that smoking, high cholesterol and a family history of heart disease are bad, and exercise is good,” says Olgin. “Yet we all know people who have horrible family histories but never developed heart disease, as well as people who are diligent about exercise and diet, but who have a heart attack. We hope the Health eHeart Study will allow us to refine our current, very crude estimates of developing heart disease into a better way to predict the occurrence and progression of heart disease that is much more individualized.”

The study has already generated buzz in the mobile health app world, and received recent coverage in the Wall Street Journal, San Francisco Chronicle and the popular BBC television series, Horizon. It also enrolled more than 1,000 people in the first five days, and by press time had enrolled 4,000 adults ranging in age from their 20s to their 80s. Participants hail from all 50 states, and from every continent except Antarctica. “We’re relying on people’s altruism to contribute really meaningful data that will help a lot of people for many years to come,” says Marcus. “The power of this study will be realized only if we have a huge number of people who participate.”

Big Questions, Big Data

Participants complete online questionnaires with basic demographic information, personal and family medical history, and information about lifestyle habits such as exercise, sleep, smoking history and alcohol consumption. The questionnaires take about an hour, but can be filled out in segments at the participants’ convenience.

Participants can also choose to share additional data, using some of the latest technology. The Health eHeart Study team is developing apps that allow participants to collect information using their own devices, such as smartphones, web-enabled scales, blood pressure machines, Using Big Data to Treat Heart Disease Continued from front page glucose monitors and physical activity monitors. The study team has partnered with several mobile health startup companies to create even more tools. These include:

  • EKGs: AliveCor (www.alivecor.com) makes a device that snaps onto a smartphone and records an electrocardiogram (ECG or EKG) when the user touches his or her fingertips to the electrodes. The smartphone then automatically uploads the EKG to the participant’s record in the study database.
  • Heart rate: The team has partnered with Azumio (www.azumio.com) to develop smartphone apps that measure heart rate and sleep. To measure heart rate, the participant holds his or her finger over the smartphone camera, which sees the pulse and uploads that information to the database.
  • Blood pressure: A wireless blood pressure cuff that is connected via Bluetooth allows participants to easily measure and upload their blood pressure.
  • Disposable sensors: The team is working with companies that are developing small sensors that attach like a Band-Aid to a participant’s chest, and measure breathing rate, heart rate, EKG, physical activity and sleep. As the study advances, selected patients – such as those with heart failure – could be invited to wear these sensors around the clock. When the monitor falls off, participants just replace it with another.
  • Exercise and fitness tests: Participants can volunteer to wear small devices, such as the Fitbit, which monitors exercise and activity. The six-minute walk test measures how far a patient can walk in six minutes, and has been shown to be highly predictive of heart disease. A nurse usually administers this test, but the study team is developing an app that allows participants to perform this test at home, with their smartphone measuring the distance walked.
  • Behavior and health: The study is also partnering with Ginger.io, a company that collects and analyzes passive (phone sensor) and active (patient-reported outcome) data from mobile devices to study behavior and health.
  • Diet: The team is developing an app to measure diet. Participants take a picture of everything they eat, and an image-matching program determines the content of the food.
  • Body composition: A measurement app uses the smartphone camera to take a panorama of a participant’s body shape, which can predict fat and muscle composition.
  • Social media and wellness: The team is developing a way for participants to share information about their use of social media sites. They are interested in learning whether indicators such as a high number of connections on these sites corresponds with more social connectedness and better health, or with less exercise and less time spent preparing healthy meals.
  • Genetics and environment: Many participants may be asked to mail in a DNA sample using saliva test kits (“spit kit”) allowing the Health eHeart Study to measure interactions between genetics and environment – learning more about how “nature” and “nurture” factors impact heart health.
  • Participants without smartphones can still take part in most aspects of the study, using a web portal accessible from any computer. Also, everyone will be invited to share access to his or her electronic medical records.

    Discovering Patterns of Heart Disease

    “The Health eHeart Study will allow us to collect very dense data on people, and really model health-related behavior, disease progression and incidence on a very individual level,” says Olgin. “A good analogy is the way that banks follow your spending very precisely to detect credit card fraud. If there is any deviation, you get a phone call saying, ‘Hey, we think your credit card has been stolen.’ And those predictions get smarter over time.”

    The Health eHeart Study team hopes to be able to tease out similar patterns and early warning signs indicating that someone has developed heart disease, generated just prior to a participant’s hospitalization, the team might be able to develop ways to predict and intervene when a participant’s heart disease is getting worse. For example, perhaps a heart failure patient who has stopped answering her phone and is sleeping an hour less per night could receive an alert or a phone call to take an extra water pill, possibly averting a hospitalization.

    “This is extremely valuable data, and is exactly the sort of thing that no other study could collect,” says Marcus. “Ultimately, if this could be translated into clinical practice, that would broadly change the way we predict and prevent exacerbations of certain cardiovascular diseases.” By sorting through the reams of data generated just prior to a participant’s hospitali-zation, the team might be able to develop ways to predict and intervene when a participant’s heart disease is getting worse. For example, perhaps a heart failure patient who has stopped answering her phone and is sleeping an hour less per night could receive an alert or a phone call to take an extra water pill, possibly averting a hospitalization.

    The Health eHeart Study could also provide useful information for relatively healthy patients who may be at risk of developing heart disease later in life. “I’m a general medicine doctor, and one of the dilemmas I face in clinic is what to do with young adults who have high cholesterol,” says Pletcher. “Is it worth taking a medication early in life in order to prevent a heart attack 30 to 40 years later? The answer may be ‘yes’ if the effects are large, and if the patient doesn’t mind taking the medicine for 30 to 40 years. But those are big ‘ifs,’ and we need to figure out better ways of studying those prevention effects.”

    Developing Better Treatments

    Part of the motivation for the Health eHeart Study is to reduce the cost of doing randomized controlled trials. As researchers learn more about the causes of heart disease, the Health eHeart Study will make it easier to design and test interventions that could help. Currently, it can take several years and considerable expense to enroll enough patients in clinical trials to draw meaningful conclusions. “The thing I’m most excited about with the Health eHeart Study is the ability to launch new studies quickly and easily, and to reduce the marginal cost of doing a research study,” says Pletcher. “It has gotten so expensive and burdensome to do any new patient-related research data collection. We’re building the Health eHeart Study to enable researchers to do sub-studies – enrolling participants into smaller studies where we can do interventions and test whether they work.”

    For example, if a researcher wanted to test whether an exercise intervention could prevent heart attacks in people with coronary disease, he or she could look at the number of people with mild to moderate coronary disease who are enrolled in the Health eHeart Study. The study would electronically deliver an invitation to those participants to join the exercise study, explaining its goals, risks, potential benefits and alternatives. Participants could sign the consent form online, then be randomly assigned to either the exercise group or a control group with no intervention. Smartphone apps could remind participants to exercise, and could also measure how much and when they are working out. Because participants already provided demographic data and medical history for the Health eHeart Study, they would not need to resubmit that information.

    “Having this nimble and efficient infrastructure to conduct clinical research is paradigm-shifting,” says Marcus, noting that the team has already received inquiries from researchers studying other diseases who may eventually want to adapt the Health eHeart Study model to their own investigations.

    Pletcher says the team may also develop a “patient-powered research network” to help brainstorm future research ideas. Some study participants have firsthand experience with heart disease, and are attuned to the small, day-to-day changes that may impact their conditions. “As researchers, we have a lot of expertise in physiology and epidemiology, but we can’t imagine all the possible ways we might improve cardiovascular health,” says Pletcher. “Participants have a lot of wisdom to share with us about what research questions might be interesting, and what interventions might lead to better health. Tapping into that wisdom could lead to lots of great new ideas, and to engagement that’s different than a traditional research study.”

    Privacy and Accuracy

    Because the UCSF Division of Cardiology is part of a medical institution, the data gathered for the Health eHeart Study is covered by strict federal privacy and security rules. By contrast, for-profit companies may be bought or sold, and their data subjected to varying privacy policies.

    “The beauty of doing this as a research study is that we are obligated to have incredible security and privacy – the data is treated the same way medical records are treated,” says Olgin. “All of our data is stored within UCSF. Participants sign a consent, and the data has gone through a security overview and is protected and encrypted. By law, we are not allowed to sell the data to anybody.”

    What if some participants try to game the system, claiming to exercise more or eat better than they actually do? “That’s always possible, but we have the same problem whether it’s an electronic cohort or not,” says Olgin. “If we are studying 1,000 people, one or two people who do that could mess things up, but when you have 1 million people, that’s just noise.” Also, because participants are submitting data to help find a cure for heart disease rather than to see how they measure up against others, there may be less of a motivation to submit answers that are not truthful. In addition, the study has built in multiple ways of measuring most variables, so researchers can internally validate the data.

    The Right Time and Place

    The UCSF Division of Cardiology is ideally positioned to lead the Health eHeart Study, since it brings together the strengths of a leading research university with the technological innovations of San Francisco and Silicon Valley. “As I talked to a bunch of start-up companies and venture capitalists in the mobile health world, it became clear that nobody is doing anything like this,” says Olgin. “It’s the right timing to do this. I think even two years ago, you couldn’t have done it, but now everyone is connected.

    “One of the really neat things about this study is that since we will be enrolling people in our own clinic along with everyone else, we have a population where we can validate everything we’re doing,” says Olgin. “It’s very easy for us to have our heart failure patients in clinic take a nurse-administered, gold standard six-minute walk test, and compare their results to the six-minute walk test administered by the smartphone app. We can prove that what we have developed actually works.” Most technology companies do not have a team of heart experts or easy access to large numbers of patients who can help them test the accuracy of new tools.

    The UCSF Cardiovascular Care and Prevention Center at Mission Bay, where most clinic appointments are held, is also located in the epicenter of innovation. “Nine out of 10 companies we’ve talked to or are working with have offices in the South of Market neighborhood, just three blocks away,” says Olgin. “There is nothing like sitting down over coffee or at a meeting, face-to- face, to get things going.”

    As more participants enroll in the study, Olgin hopes to hire data scientists to analyze and learn from the accumulating wealth of data. “I’m really interested in developing new faculty within cardiology who can develop machine-learning algorithms, crunch the data, and figure out how to present that data back to the individual in a new, meaningful way,” says Olgin. “It’s analytics, informatics, interface and visualization – all the stuff that happens at Google and Silicon Valley companies that hasn’t happened in medicine yet. We need that talent, because the Health eHeart Study is opening up a whole new science.”

    Gift Fuels Rapid Development

    The Health eHeart Study was launched with generous support from a private funder. “This initial seed money has allowed us to rapidly develop an interesting idea into an innovative, robust study that takes advantage of the latest technological tools,” says Jeffrey Olgin, MD, one of the principal investigators of the Health eHeart Study. “This allowed us to progress at warp speed, which is just not possible with traditional sources of grant funding – the funding cycle alone from application to start of project is often one year.

    “Philanthropic gifts are essential to laying the groundwork for this kind of out-of-the-box research endeavor,” says Olgin. “This funding has allowed the Health eHeart Study to develop at the pace of a Silicon Valley start-up. We are incredibly grateful for what this philanthropic gift has allowed us to accomplish so far, and look forward to partnering with other supporters as we continue working towards our vision of enrolling 1 million participants worldwide.”

    TO LEARN MORE about the Health eHeart Study or to sign up, please visit www.health-eheartstudy.org.

    To find out how you can support the next phase of the Health eHeart Study, please contact Director of Development Eileen Murphy at (415) 502-0746 or emurphy@support.ucsf.edu.




    Frontiers of Medicine Main Page