Illustrations by Marco Goran Romano

Introduction


How can we transform health care—the doctor’s appointments, finding specialists, managing medications—from a paper-heavy headache into a streamlined, personal process?

There’s more than one answer. It’s widely agreed that we need first of all a broad paradigm shift in our national approach to health care. But we also need to develop and deploy revolutionary new technologies. For example: data-collecting sensors that can monitor patients’ health in real time; data analytics that can help doctors diagnose difficult cases; and machine learning, which allows cloud computing to modify its own programming in order to form conclusions and predictions from vast amounts of data collected from millions of patients over time.

The cloud forms the basis for all those technologies, which will transform our health-care system from the waiting room to the operating room. Many of the changes will be obvious, such as electronic records that save patients’ and providers’ time. But others, all but invisible to patients, will enable a life-saving revolution in medical care.

The cloud will have an effect on every stage and iteration of the health-care process, turning it into an ever-more-watchful, ever-smarter system. Nowhere is the potential for that change clearer or more hopeful than for premature babies, among the health-care system’s most fragile, precious patients.

83 Percent

Percent of IT health-care organizations employing cloud services 1

92 Percent

Percent of health-care providers who consider cloud services to be valuable to their operations 2

65 Percent

Percent of providers who planned to increase their spending on data analytics in 2015 3

CHAPTER 1

Always-On Care with the Internet of Things


The Internet of Things is a nexus of smart sensors and advanced data analytics that can produce previously unattainable insights. Applied to health care, it can capture every heartbeat, every breath, every half-degree change in body temperature—and make that information known where it is needed, in real time.

When babies are born, their bodies find themselves independent for the first time, suddenly required to regulate bodily functions from breathing to eating to maintaining the right balance of fluids and chemicals in the body. Premature babies especially can have a hard time with this transition. They’re hooked up to IVs to maintain their physical equilibrium, and blood tests are performed regularly to identify any deviations from the norm.

As their bodies adjust to the environment, they also become more vulnerable to infection. Smart devices that are gaining traction in hospitals can help prevent that in many ways, from monitoring individual patients’ biometric markers to tracking hand-sanitizer use throughout the building.

GOJO uses smart sensors that measure foot traffic through hospitals and the use of antibacterial dispensers to track hygiene compliance.

$117 Billion

The health-care Internet of Things market segment is poised to hit $117 billion by 2020. 4

2.8 Million

Number of patients globally who were using a connected home-monitoring system at the end of 2012 5

CHAPTER 2

Advanced Streaming Analytics


What actually happens in the cloud? Huge quantities of data, streaming in from the Internet of Things’ multitude of sources, can be analyzed almost instantly, with far more computing power than hospitals have on site, and can sound alarms as soon as problems appear.

1+ Million

Number of health-care data streams capable of being analyzed at once by the cloud 6

By ingesting and analyzing such real-time data as the potassium levels in a baby girl’s blood, her heart rate, and the temperature in her incubator, the computing power of the cloud is capable of delivering medical insights that humans simply can’t. Sifting through enormous amounts of data virtually instantly, the cloud monitors readings from its connected devices to make sure they stay in the healthy range—and to alert the medical staff if they aren’t.

Cloud-based analytics can change the way health care works outside of the hospital as well. With someone who needs supervision and care around the clock, for example, not even the most diligent doctor or nurse could possibly monitor moment-by-moment changes in the patient’s medical condition. The cloud, however, can hone in on irregularities almost instantly.

Aerocrine’s medical devices can measure the content of asthma patients’ exhaled breath. Because they’re so sensitive to the environment, they have to be closely monitored: The cloud allows Aerocrine to proactively determine when devices need replacement.

CHAPTER 3

Smarter Care with Machine Learning


In conjunction with streaming analytics, the cloud also utilizes machine learning to draw conclusions from deep wells of historical data. It’s capable not only of recognizing patterns in the readings of many devices simultaneously but also of putting them in the context of the outcomes from millions of past readings.

With machine-learning capabilities, parents can rest assured that the health-care readings of their newborns are being continuously analyzed against the experiences of millions of premature babies: The slightest deviation from normal measurements will prompt alerts to the relevant nurses and doctors. And machine learning means that the acuity of the analytics will increase and improve over time. As it collects more data from more babies, it adjusts the “thought process” of its analysis to account for new information—the latest published medical research on newborns, for example—and develops an increasingly sophisticated understanding of patterns in their health.

With the ability to alert medical professionals to health-related issues in minutes if not seconds, the combination of data analytics and machine learning will save lives. It can also help people whose chronic conditions complicate and burden everyday life.

500 Petabytes

In 2012 there were about 500 petabytes of patient health-care data being stored by hospitals. 7

25 Thousand

By 2020 that number is expected to increase 50 times, up to 25,000 petabytes of health-care data. 8

12+ Thousand

There are more than 12,000 possible medical diagnoses. 9

$26 Billion

Almost $26 billion could be saved annually by preventing avoidable readmissions to the hospital through data analysis. 10

Using eye-tracking technology to measure students’ eye movements during reading, Optolexia uses machine learning to identify dyslexia with 95 percent reliability and provide assessments to teachers about students needing extra support.

CHAPTER 4

Clinical Decision Support System


Ultimately, the gathering and analysis of data by smart machines would mean nothing if it couldn’t be put to use. That utility relies on a clinical decision support system, which presents structured data and insights from the cloud to the medical professionals who must decide what to do with them.

Because the physical development and immune systems of babies born prematurely are so fragile, they are at risk of a long list of illnesses—anemia, chronic lung disease, and pneumonia, to name a few—each of which is complex in its own way. The clinical decision support system gives the neonatal ICU’s health-care team a way to make clinical sense of the enormous quantity of data streaming in on every one of its babies—and to deal effectively and quickly with that range of complexities.

Doctors and nurses are provided with an array of potential diagnoses and treatments for the newborn, with probabilities attached to each one and a discussion of their merits and demerits. Armed with the smartest, fastest research assistant imaginable, medical experts can make ever more informed decisions for each of their newborn patients.

Past experiments in neonatal care with clinical decision support systems have found that the time taken to calculate drug doses was significantly reduced, as were errors associated with that calculation. Such systems promise to revolutionize not only neonatal care but every other kind of critical care as well.

20 Percent

Nearly 20 percent of Medicare patients admitted to the hospital for chronic conditions end up going back within a month. 11

12+ Million

Number of Americans estimated to suffer from misdiagnoses, costing time and money 12

Tacoma Hospital compares patient data to a database of heart failure outcomes to analyze their risk of readmission and provide patients and providers with relevant recommendations to prevent it.

Looking Ahead


Our youngest patients need the most detailed, instant, and acute diagnostic information possible. The cloud makes that kind of care a reality. It can take the form of a round-the-clock nurse, tracking vital signs and health data at all times, or a diagnostic assistant, with the knowledge and ability to help doctors make clinical decisions. And in the future, as the cloud collects and analyzes more and more health-care data, it will be able to disclose ever deeper insights about medical treatment and possible cures. The cloud will never replace doctors and nurses, but it is going to make our health-care system smarter and more watchful than it has ever been before, which will be an especially precious gift to the patients who need it most.