Abstract
P
If you don't believe it, ask yourself: what is your hematocrit, today? What is your cardiac output, right now? Or, are any of the patients you operated on last week developing the early signs of sepsis or worsening cardiac failure, at home alone and unmonitored, as you read this? The likely answer to real-time medical questions is “I don't know.”
Such inattention to real-time details would be unacceptable in other aspects of our daily lives. We demand that satellites and sensors monitor worldwide and local weather, even when the weather is good, so that we are neither surprised nor injured by hurricanes, heat waves, or blizzards. We expect our cars to use continuous sensor output to reduce accidents, or even to deploy protection when accidents are inevitable.
Yet, we still remain surprised when our body contracts a cold or flu, and no sensor gives us advance warning when our bodies urgently require preventative action. You probably know more about how your phone is doing (alerts you when its battery is low) than you do about how your heart is doing (no alerts when your cardiac output is low).
But, this is now changing.
Wearable, swallowable, and implantable sensors are here, and they are hitting the mainstream.
Within 5–10 years, a billion people's bodies will be connected to the Internet, and continuously monitored every day. Your body will have an Internet address. It will be instrumented with sensors, and quantified every second.
Consider how great a sea change this is.
Just a generation ago, I sat in an intensive care unit (ICU) at Harvard watching a premature newborn infant's toe glow red as a then-new experimental monitor, a pulse oximeter, read the oxygenation of blood through the skin using photons, saving the infant from a painful blood draw and subsequent transfusions. At that moment, perhaps only a few hundred patients worldwide were continuously monitored using such optics. This grew to hundreds of thousands of patients monitored at any one moment, as pulse oximeters grew in use by 2010. But almost overnight, wearable monitors have become popular among fitness buffs and medical patients alike. Now, tens of millions of people continuously monitor their heart rates day and night using light sensors, electrical impedance sensors, or MEMS accelerometers embedded in wearable wrist bands, watches, patches, and rings. And the use of these “wearables” is growing rapidly.
Wearables will change medicine in fundamental ways.
First, they bring real-time data to outpatient medicine, allowing patient-specific interventions. Continuous monitoring allows us for the first time to understand a range of what is normal for each and every individual, instead of relying on static reference normals for a population measured years ago. It enables caregivers to more sensitively track progress or detect deterioration; for example, consider that blood pressure monitored once a year in a physician's office is a poor predictor of cardiovascular risk. In contrast, continuous monitoring is able to strongly predict stroke and heart attack risk. 1,2 Once you have a continuous measure for which intervention is known to change outcome, 3,4 then you can begin to change behavior or treatment. No longer does heart disease appear to patients as something in the distant future, because they can see the immediate effects of what they do (new diet, new lifestyle, different medications), bringing awareness and management into the present, and engaging them in their own self-care, ultimately changing outcome.
Second, these wearables enable “Big Data” as a new paradigm for healthcare and health research. Rather than standard small studies defining what is normal or testing for associations and patterns, wearables effectively conduct the world's largest clinical study in history, a study in which a majority of the population can participate 24 h a day, every day. Consider that between 1900 and 2000, approximately 425 articles per year were published in medical journals on heart rate, respiratory rate, and blood pressure in normal and diseased subjects, perhaps including a total of 100,000 patients a year (PubMed currently lists 42,400 articles with “heart rate” in title or abstract since 1900, or an average of about 425 per year).
These studies have encompassed our entire medical understanding of normal ranges for heart rates. However, as noted, nearly 20,000,000 people are now having their heart rates (and in some, respiratory rates and blood pressure) monitored continuously, every single day, collecting each day data from nearly 200 times the number of subjects previously studied and published in a year over the past 100 years. Even the United States Food and Drug Administration (FDA) has noticed this potential, and has just asked for input on how to understand data from wearable devices. 5
Now, patterns of disease and relationships that were imperceptible in formerly large 10,000-patient studies will leap to significance. Big Data mining approaches that have been recently applied to a wide variety of data sets will now become medically focused. Even artificial intelligence systems, such as IBM's Watson, can begin to explore medical questions that we have not asked before. 6 We will almost certainly find new relationships between lifestyle and disease, and new and better predictors. The patterns emerging from these data will result in new insights and new treatments that will lead the way, rather than trailing behind changes in patient care.
Last, the types of tests we can perform will continue to expand. Today, we can measure vital signs such as heart rate, respiratory rate, and temperature, as well as sleep length, sleep state, tissue perfusion, and arrhythmias, at the wrist. Soon, a new generation of sensors, many of them optical, but also including electrical, ultrasound, electrochemical, temperature, pressure, movement and position, shock, touch, and even identity sensors, will add new determinations, from spot tests of blood chemistries to continuous blood pressure to real-time monitoring of emotional states. Already, in my own group, we demonstrated a no-touch noncontact technology that can read vital signs and certain blood chemistry levels at a distance, which makes Star Trek's tricorder look positively medieval.
So, where will this lead?
After all, this is all about improving the health of people everywhere, and the information is useless if patients can't benefit from it. Every patient already knows that eating too much, and not exercising is bad, but many of us ignore this knowledge. But here is where the wearables will shine.
The first studies showing that these devices change behavior and outcome are just being published. For example, athletes that monitor their sleep perform better the next day, 7,8 and people who monitor their steps actually take more steps and get more exercise.
And other changes can be seen.
Last month, a Stanford intensive care specialist became the chief medical officer of the $3 billion consumer wearable electronics company, Jawbone. That medical specialist was me. At Apple, they have hired an anesthesiologist as its chief medical officer. And earlier this year, Fitbit, a wearable tracker startup, floated a public offering that brought in nearly $700,000,000 and a multibillion dollar valuation. Even Google, Microsoft, and Amazon have serious health monitoring Big Data initiatives. Remarkably, these are mass-market consumer companies entering into what used to be the exclusive domain of medical companies.
And physicians and surgeons are joining these programs, alongside the physicists and biomedical engineers. Recently, I met with Dr. Bruce Tromberg, the Director of the Beckman Laser Institute at University of California at Irvine. He is an optical physicist and bioengineer; however, what he wanted to talk about was heart rate and respiratory patterns.
In the end, this will benefit patients. All too often, medicine acts after a disease is underway, and after damage has set in. What if Apple's CEO could have been warned, years before the symptoms were apparent, that he had signs of cancer, through a wrist-worn band? What if we can screen mass numbers of people noninvasively for malaria using sensors in lighting systems or cell phones? What if we could detect all impending heart attacks based on monitored drops in cardiac output? What if your car could sense your emotional state, or assess you after an accident? These examples are not science fiction, they are all possible today.
This remarkable story is a tale of sensors and engineering, and it is unfolding as you read this. Lasers and other light-emitting sensors are leading the way into the world of the quantified self. Just watch what happens as the source of patient data moves from originating in the physician's office to coming from the patients' wrist.
In the end, it will come down to a living connection to the broader network of sensors blanketing the globe. The Internet of Things will give way to the Internet of You. And it has begun.
Medicus, sentire teipsum. 9