News > i3

Robot, MD: How Artificial Intelligence Will Revolutionize Health


For decades, science fiction has inspired rich forecasts of what the future will look like. Technologies like virtual and augmented reality, digital voice assistants and connected devices were first imagined by science fiction authors and futurists.

Many modern-day medical technologies also debuted in science fiction. Dr. McCoy’s tricorder in Star Trek served as the inspiration for DxtER, a health analytics device which uses noninvasive sensors to track vitals, body chemistry and biological functions. Elder care robots like SoftBank’s Pepper and Luvozo’s SAM — and social robots like ElliQ and Jibo — draw inspiration from the portrayal of companion and care robots in science fiction films like Big Hero 6 and Robot and Frank.

Mechanized exoskeletons like Iron Man’s armor and Ellen Ripley’s power loader in Aliens helped create functional blueprints for exosuits being used in military and manufacturing applications, as well as for robotic devices like ReWalk which assists paralyzed individuals to sit, stand and walk. As for the U.S. military’s research and development agency, the neural-interface technology developed by the Defense Advanced Research Projects Agency (DARPA) has helped amputees to experience touch with a prosthetic hand or control an advanced prosthetic arm with their mind. That should remind Return of the Jedi fans of Luke Skywalker’s bionic hand.

Artificial intelligence, or AI, is a key ingredient in making each of these real-life applications viable. Machines leveraging AI tools are making inroads into nearly every aspect of our lives. Fusing together the physical and digital realms, AI’s impact on society will be far reaching and profound. Impact evaluation and market sizing from McKinsey estimates that AI will generate value between $3.5 trillion to $5.8 trillion annually across nine business functions in 19 industries.

According to a recent Accenture survey of health care executives, a majority (85 percent) agree that every human will be directly impacted daily by an AI-based decision within the next three years. Accenture expects the AI health market to reach $6.6 billion by 2021, growing 40 percent annually, and potentially generating $150 billion in health care savings by 2026. Another way to gauge AI’s health care potential is by looking at how the investment world views it. Research from CB Insight, a financial analysis firm, shows that AI startups have raised $4.3 billion since 2013, topping all other industries.

From sifting through piles of data and generating insights, imaging and diagnostics, drug discovery to remote patient monitoring and purpose-built digital assistants, AI will also impact every aspect of health care. Frost and Sullivan, a consulting firm, predicts that by 2025, AI systems could be involved in nearly all aspects of global health care, from AI doctors responding to specific patient care needs to managing entire health care systems.

I’m a Doctor, Not a Robot

While the convergence of several trends is responsible for the sky-high expectations from AI’s involvement in health care, none is more impactful that the exponential increase in digital health data. An International Data Corporation (IDC) report estimates that in 2013 the volume of global health care data was nearly 150 billion gigabytes, a number it expects will grow 15 times by 2020 to roughly 2.3 trillion gigabytes. At an individual level, IBM says each person generates one million gigabytes of health-related data during their lifetime.

It’s not hard to imagine that the volume of data collected about our bodies will only surge in the future. This year alone, according to some estimates, the wearables data generated by Android Wear, Fitbit and Apple Watch will likely produce two trillion health measurements. And much of this extraordinary amount of data is “unstructured”, which is AI lingo for data that’s not part of a database with detailed labeling and definitions.

Even if we had the time and resources to integrate and analyze the data, the overwhelming nature of data collection rapidly outpaces our capacity to process, associate and find patterns and knowledge. Enter artificial intelligence. Its ability to crunch vast amounts of information into insights far outpaces human minds or any conventional analytical systems. Armed with this extraordinary ability to draw correlation and connections between seemingly distinct and unrelated data sets and tools for natural language processing, AI systems can tackle challenges for which we previously had no other means of recourse.

The use of AI now and in the future, along with robotics and emergent immersive realities, will deliver rich solutions which will place greater emphasis on improving health outcomes and preventative care. As AI gains sophistication, it will get better at doing what humans do, and do so more efficiently, quickly and cost-effectively.

Many powerful AI tools are already embedded in mobile devices and wearables running Android or iOS. By harnessing connected health technologies, consumers can collect data on their bodies to proactively manage their health and wellness. And consumers for their part express a greater willingness to do so now than ever before. Accenture’s 2018 Consumer Survey on Digital Health found 75 percent of U.S. adults believe it’s important to use digital technology to manage their health, with nearly half using their mobile devices and another third using wearables to do so. About one in five consumers also have already used health-related AI technologies and find value in these products and services.

AI Health is Your Greatest Wealth

One of AI’s biggest potential benefits is to help people stay healthy so they don’t need to seek professional care as often. The use of AI in a wide range of connected health applications is already helping people to manage their own health care and engage in healthier lifestyles.

Examples in this area include the WELT smart belt which debuted at CES 2016. By dynamically tracking waist size and tension, the smart wearable determines when the user may have overeaten and sends an alert via its smartphone app. The Fitbit coach app, paired with a Fitbit wearable, has one of the largest health databases, and takes advantage of complex algorithms to extract meaningful insights from user data, to provide highly-tailored workout recommendations. Apple’s Health app and Apple Watch work in a similar fashion. Meanwhile, YAZIO calorie planner integrates data from multiple wearables/apps and analyzes other aspects like diet, nutrition and activity to get a bigger picture in terms of personalization.

Another wellness area ripe for AI assistance is mental health. Addicaid is a “counseling assistant” app which leverages clinical research, machine learning and adaptive AI to predict when a person might be at risk of falling into addictive behaviors and offers personalized treatment options. Another product in development is WoeBot, built by Stanford psychologists and AI experts. It tracks a user’s mood through brief daily chat conversations and offers curated videos and word games to assist people in managing their mental health.

These AI applications and others encourage healthier behavior in individuals and help with the proactive management of a healthy lifestyle. It puts consumers in control of their health and wellbeing. Health professionals also gain a better understanding of the day-to-day patterns of their patients and can provide better feedback, guidance and support.

AI is also being used to diagnose and treat diseases such as cancer earlier and more accurately. For example, Houston Methodist Research Institute’s use of AI software reliably interprets mammograms, assisting doctors with a quick and accurate prediction of breast cancer risk. The AI software intuitively translates patient charts into diagnostic information at 30 times human speed and with 99 percent accuracy. Or consider Cyrcadia’s iTBra, a wearable sensor used to screen for breast cancer at earlier stages. Early trials of the wearable among 500 patients registered an 87 percent successful cancer detection rate, even with patients with dense breast tissue, which often hides small tumors.

Another AI-driven cancer detection system built by researchers at the University of Heidelberg in Germany, pitted an AI system against a team of dermatologists and found that the AI achieved a 95 percent accuracy in identifying skin cancers, compared to the 86.6 percent achieved by trained professionals.

Heart health is also receiving strong interest. Researchers at Oxford University developed AI to interpret ultrasonic scans of the heart. The goal is to assist cardiologists in detecting subtle changes and identify signs of heart disease more accurately. Ultromics, a British startup, has developed a machine learning platform which aims to reduce errors diagnosing coronary artery disease by as much as 75 percent. The proliferation of consumer wearables and other medical devices combined with AI is also being applied to oversee early-stage heart disease, enabling doctors and other caregivers to better monitor and detect potentially life-threatening events. An AI-based Apple Watch app, Cardiogram, has been able to detect an abnormal heart rhythm with 97 percent accuracy. The same app can also detect sleep apnea with a 90 percent accuracy and hypertension with an 82 percent accuracy.

A new wave of AI chatbots could help ease pressure on overstretched health care professionals by reducing the need to make in-person appointments. The goal of apps like Babylon, Ada, Your.MD, and Dr. AI is to make seeking advice about a medical condition as simple as using a more-reliable WebMD. Unlike self-diagnosis online, these apps lead the user through a clinical-grade triage process. In the future, not only will these tools parallel humans in diagnosing diseases, they will perform them on timelines that are impossible for humans to match. This efficiency is of tremendous value in a field where time is quite literally a matter of life or death.

Machines built with AI will likely challenge the traditional role of the doctor or nurse. But they are unlikely to replace the need for high-trained clinical minds any time soon. AI systems currently work on a narrow range of tasks and will need human input, particularly in the form of professional expertise, for years to come. Its real benefit lies in freeing up health care workers from dull and repetitive tasks enabling them to spend more meaningful time with their patients. As these technologies take shape, they will lead the current health care model from primarily treating illness to keeping patients healthier longer. After all, an ounce of prevention is worth a pound of cure.

Sayon Deb

Tagged

Related