Friday, 24 May 2019

Personalized Health Care and Artificial Intelligence Could Improve Your Life—At the Cost Of Your Privacy – Newsweek

Andres Rubiano first got the news that his blood pressure was too high in the 1990s, when he was in his late 30s. It didn’t come as a complete surprise—his father had had chronic hypertension at an early age too. His doctor prescribed medication and encouraged him to get more exercise and cut down on the amount of salt in his diet. Rubiano, though, wasn’t very diligent about following this regimen. Each time he returned for a checkup, doctors gave him the same advice and Rubiano disregarded it.

Four years ago, something caused Rubiano to turn himself around. His doctor convinced him to enroll in a pilot project in digital health care. Once a day, Rubiano slipped on an automatic cuff that wirelessly sent blood pressure readings to his smartphone, which in turn relayed the data to a team of clinicians at Ochsner Health System, an academic medical center based in New Orleans. His Apple Watch relayed heart-rate and physical-activity readings. Soon, Rubiano was getting text messages reminding him to take his pills and emails suggesting ways to cut down on salt and boost physical activity. Each month, his doctor’s office called to discuss his readings, tweak his medication dosage and examine new diet and exercise strategies.

The digital prodding paid off. Rubiano now takes his pills, goes to the gym three times a week and eats less salt. “I don’t even like the taste of it anymore,” he says. His blood pressure dropped from 150 over 100 to a reasonable 130 over 78.

Rubiano is an early beneficiary of a new data-driven approach to health care that is starting to take hold in the medical industry. As the fate of the Affordable Care Act plays out in the courts and politicians and patients wring their hands over the rising cost of drugs and treatments, a growing army of researchers, practitioners, entrepreneurs and executives think tech could be a simpler way to fix much of what ails the U.S. health care system. Essentially, they want to do for medicine what Facebook and Google did for social media and internet search. By gathering real-time data on patients and using it to prod, poke and persuade them to do the right thing, they believe that doctors could improve outcomes dramatically for patients with a broad range of chronic diseases and conditions, including heart disease, cancer, diabetes and Alzheimer’s. “We can’t just fix chronic disease; you’ll have it until you die,” says Richard Milani, a cardiologist who serves as chief clinical transformation officer at Ochsner. “We need to surveil you closely and catch you when you’re heading out of bounds.”

There are, however, clear risks. If you think Facebook and Google invaded your privacy, imagine what hackers could do with a minute-to-minute log of your disease symptoms, behaviors, locations and even your appearance and conversations. The key to digital medicine is, after all, the gathering and sharing of copious data about nearly every aspect of a patient’s life. Are hospitals, researchers, Big Pharma, insurers and tech companies worthy stewards of such sensitive data? “If a system can detect signs of cognitive decline in patients’ voices, they don’t want it being used against them in job situations or health insurance,” says Christine Lemke, co-founder of Evidation, a health data company.

If experts can figure out a way to lock down privacy, digital medicine could pay off big time. Its biggest impact is likely to be in chronic care. At present, U.S. health care is designed for crisis-oriented transactional care—you break your arm, the doctor sets it and you go home. That mode works poorly in early detection and treatment, which is what’s needed to keep a patient with heart disease or diabetes from slipping into a medical crisis. The Centers for Disease Control and Prevention reports that 60 percent of the U.S. population has a chronic disease and that 40 percent has at least two. Heart disease, diabetes, cancer, Alzheimer’s and stroke together account for almost three-quarters of all deaths in the U.S. and 86 percent of health care costs. Chronic disease is “the leading health burden on Americans,” says Milani. “And it’s where the digital piece can make its greatest contribution.”

Digital medicine is a key tool in a broader trend toward “value-based” care, in which health providers are compensated not for each procedure they perform, which drives up costs, but for keeping a population healthy. That puts the emphasis on prevention, which is exactly what digital medicine is suited to. A whole new field of data-driven medical research has arisen to explore the links between health and genes, the environment and behavior. Researchers use artificial intelligence (AI) to digest this data and uncover important patterns.

There are indications that people would be willing to take the risk and trust the medical establishment with their vital signs and other intimate details. Millennials, despite Big Data hacks and abuse by tech companies, seem blasé about privacy but don’t much like their health care. According to a recent Accenture survey, people aged 22 to 38 are up to three times more likely than older patients to express dissatisfaction with the health care system, and half of them are already using digital tools to self-manage some aspect of their health. Even Rubiano, now 58, says he doesn’t mind having his vital signs digitized and shared in real time. “It’s really made a difference knowing that someone is always watching over me through these readings, instead of waiting one or two years for anyone to notice,” he says.

FE_DigitalHealthcare_02_925542192A digital health industry has spent billions on developing new tools and projects in recent years. To improve patient care, they are trying to create an ecosystem of apps and devices that link hospitals and doctors offices to data from portable devices like the Fitbit and an insulin injector for diabetics. FS Productions/Getty

Beyond the Fitbit

A digital health industry has sprung up quickly in recent years. Investors inked almost 700 deals last year for nearly $7 billion, according to a study by Health Data Management, an industry journal. Billions more flowed from health care institutions to develop digital tools and start pilot projects and from tech giants like Apple, Amazon and Google, which are determined to capture market share.

Much of the investment is aimed at finding ways to collect health data that go well beyond conventional patient electronic medical records. Google’s parent company, Alphabet, for example, runs a company called Verily that’s partnering with Duke and Stanford universities to collect a broad range of data on 10,000 mostly healthy people. And Mindstrong, a Silicon Valley startup co-run by former National Institute of Mental Health leader Thomas Insel, is aiming to track people’s behavior in order to catch early signs of mental-health and other brain-related problems.

Much of that data comes from the wireless gadgets that we have become accustomed to interacting with all day long. The worldwide market for smart watches and other activity trackers is now $6 billion and growing at 20 percent a year. Health care–specific trackers are also on the rise; entrepreneurs and investors aim to build an ecosystem of apps and gadgets that patients can pick and choose from to manage their medical data. Everything about your body, your behavior and your environment can be mined for insights. Scientists at Cedars-Sinai Medical Center in Los Angeles, for example, are developing a device that sits against the abdomen to listen to what goes on in the colon to understand digestion. A new watch from Withings monitors blood pressure and takes electrocardiograms. Data from Fitbit trackers alone has inspired 675 studies in health-related journals. Even the 50 million security cameras in the U.S. could be harvested for data on how fast, frequently and robustly you walk or run, and whether you look happy, sad, pained or confused.

FE_DigitalHealthcare_04_579056257A man checks his heart rate on a Fitbit Charge HR wearable activity tracker and monitor. Data from Fitbit trackers alone has inspired 675 studies in health-related journals. Lisa Werner/Getty

Health care professionals call it “quantified biology.” For years, medical researchers have been studying the genome, which is the sum total of an individual’s DNA, and the proteome, which encompasses all the proteins a person’s cells produce. Now, they are looking at the exposome, which contains data about the environment—the sounds, chemicals, people, sunlight and anything else that an individual encounters in life—and the behaviorome, which encompasses behavioral tendencies while eating, sleeping, exercising and beyond.

The data on exposomes and behavioromes is currently sparse, but that’s changing. Yale biologist Caroline Johnson, who specializes in metabolomics—the study of chemicals that insinuate themselves from the environment into our bodies—notes that researchers have found 750,000 metabolites that are linked in some way to health issues. Emory University researchers have developed tests that can identify these metabolites in a urine sample at a rate of 1,000 per minute. The National Institutes of Health (NIH) has set up a $1.5 billion program, called All of Us, that’s planning to gather all sorts of tracking, environmental and other health-related data from at least 1 million people. One goal is to look for clues about how variables like exercise, diet, sleep and heart rate are linked to disease.

To get a sense of how this new “omic” data might improve health outcomes, consider Walter De Brouwer, a computational-linguist-turned-entrepreneur from Mountain View, California, who not long ago was having trouble sleeping through the night. Lack of sleep is linked to high blood pressure, heart disease, diabetes, Alzheimer’s and even to cells’ ability to repair errant genes. The sleep-tracking app on De Brouwer’s Apple Watch confirmed that he was waking up twice each night. He found the first culprit by checking video from the security cameras scattered around his house: His son was making noisy late-night visits to the kitchen. A data dump from his smart thermostat revealed the second: a middle-of-the-night temperature spike in his bedroom.

De Brouwer channeled his data obsession into a startup called, which makes a medical app that automates the process of assembling its users’ exposome and behaviorome. You start by taking a selfie, which the app’s AI uses to estimate your age, height, weight and gender (the idea is to reduce the amount of information a user has to enter manually). Next, you enter your current and previous ZIP codes, which the software uses to query databases of air pollution. In the future, the company plans to add location-based data on water pollution, disease prevalence, mosquito counts and other data. Then, if you trust app, you grant it access to all your online health information—your health care provider’s patient portal, labs where you had blood or other tests, the results of a 23andMe gene scan and so on. The final step is to connect the app to your phone’s GPS, fitness-tracking watches and other smart devices. Doc.ia keeps all the personal information on your phone, says De Brouwer, not in the cloud, to keep it away from potential hackers. is just one of many startups that aim to be the go-to app that patients can use to manage their health data. The company signed up 35,000 people in six months, enough to run some statistically meaningful trials. It is partnering with health care providers, insurers and research institutions to offer its users opportunities to put that data to work in managing their own health, or contributing to research, or—as is typical in medical studies—both. Each user decides on a case-by-case basis who gets what data for which study or service. So far, has offered its early users access to an allergy research study in partnership with Harvard Medical School and health insurer Anthem—the 2,000-person study filled up in three weeks, with another 8,000 on a waitlist. Studies on Crohn’s disease and colitis are live, and epilepsy is coming.

FE_DigitalHealthcare_10_1023297246A scientist studies data about a person’s genetic makeup. Today everything about your body, your behavior and your environment can be mined for insights. Westend61/Getty

Converting Data to Health

Although the field is new, health care organizations are already exploring ways of applying data to care. Ochsner’s blood pressure program has enlisted more than a thousand patients, and the numbers are encouraging: 79 percent manage to keep their blood pressure in safe territory, compared with 31 percent for the hospital overall. In addition, Ochsner has set up a digital program for patients with chronic obstructive pulmonary disease, nearly a third of whom wind up in the emergency room at some point gasping for air, which is traumatic and imposes enormous costs on the system.

To head off these crises, doctors at Ochsner wired their patients’ inhalers to alert them to patterns of usage that signal growing respiratory distress. Those patients get a text or call at home suggesting a change in medication or prompt a checkup. Patients also get text alerts when air particle readings are dangerously high in their communities, urging them to ease up on outdoor activity. “We can wirelessly manage the symptoms of thousands of patients closely,” says Milani. “That’s beyond the capabilities of a non-digital physician’s office.” 

Obesity is another target. Wake Forest Baptist Health in Winston-Salem, North Carolina, for example, has injected a strong digital component into its highly regarded clinical weight-loss program. Patients get a digital scale for home that transmits daily weigh-ins to an app, which makes the data accessible to doctors. It also tracks physical activity, food choices, water intake, sleep and other factors. “We use all that data to provide constant feedback to the patient,” says program director Dr. Jamy Ard. “We can modify the plan, suggest different foods and activities, and set up a video visit if the patient opts for it. Because of that data, we can have a meaningful conversation.”

Drug companies are developing smart drug-dosing devices that collect patient data while they’re being used. Eli Lilly is piloting insulin injection pens and pumps that can gather information on how much insulin a diabetic patient injects, and when