AI and active health

November 16, 2019
"Active health" is a technology-enabled step towards proactive and preventative health management that could define the very notion of health care in the future.

Much of Asia’s AI development emphasis is on increasing the capacity and capabilities of medical professionals and facilities, as well as solving specific local health burdens. But increasingly, policymakers, academics, and technologists across the region are investigating the potential for AI to radically redefine the very notion of health care. A new concept, becoming known as “active health,” is emerging whereby health authorities will compile and categorize disease markers, risk factors, and other intelligence on disease and health conditions, and use it to make predictive recommendations for what people should do to improve their health.

CAICT’s Xu Shan points to health management and aging measures as main concerns for China over the coming years. China’s Ministry of Science and Technology’s 2018 Notice on a National Key Research and Development Plan on “Active Health and Aging Technology Response” is “one of several key special project declaration guidelines China will be introducing over the next four years in areas that are most promising.” She identifies several health-adjacent technologies, including deep learning applications for AI, virtual assistants, wearable monitoring, and multi- heterogeneous data analysis, as well as data processing (labeling and quality control), as industries that will grow quickly in China over the coming two to five years.

Lessons from Bian Que

Xu believes that this new technology focus will catalyze a leap from reactive interpretation of presented data (that is, suggesting a diagnosis based on the review of an MRI scan) to AI applications which, empowered by deep neural networks and even deeper data lakes, make predictive recommendations on patient behavior and treatment for optimal health.

She illustrates this objective using a folk tale of Bian Que, a legendary physician in the Zhou dynasty (circa 500 B.C.) who is often considered China’s first notable medical professional. Bian Que had two brothers who were also doctors, and according to legend, Xu explains, “A king asked Bian Que, the most famous of the three, which one was the best doctor. Bian Que replied that he himself was only average, and only the most famous because he treats people who are already very sick, so the effects of his treatment are well-observed. His second brother, Bian Que explained, was better, because he treated people at the first sign of a disease, when they were only a little sick. His eldest brother was the best doctor, because he saw what could make them sick, and treated them before they even felt bad.”

“Active health” is a technology-enabled step towards proactive and preventative health management that could define the very notion of health care in the future. The ongoing convergence of AI and wearable devices will accelerate health-care providers’ capabilities to predict and prevent the development of non-communicable diseases.

Wearables + AI

Across Asia, wearables are seen as part of this new focus on active health. The region is already the world’s largest market for wearable devices, with revenues estimated at $7.3 billion, dominated by China (at $4.6 billion) and India (at $1.4 billion), and growing at roughly 4.4% annually. Asia’s manufacturing sector is also the epicenter of the world’s largest consumer electronics ecosystem. As a result, there is a burgeoning health-care sector in Asia that is focused on building AI-enabled tools and devices. The DFree toilet timing assistant developed by Japanese startup Triple W is an example of how wearables and AI technology converge to create an active health tool.

Xu believes it is in the ability to predict and advise based on indicators and conditions prior to medical events “that the true transformative power of AI applications in health care will be realized.” This shift from reactive or interpretive AI will partly be driven through deeper insight into existing conditions. Already, intelligent wearable devices can transmit information to analytics systems, which can recommend exercise, diet, and medication regimes, and keep constituents abreast of how effective they are in following those regimes.

Other examples of Asian innovation in wearables include Hong Kong-based Well-Being Digital that has over 50 patents for its sensitive and accurate heart-rate monitors that can be used in earphones and any number of wearable devices, and Singapore’s Health Promotion Board partnering with Fitbit to launch a nationwide healthy living campaign which will see free health monitors distributed to citizens.

No substitute for a human

Yet even with this convergence, the transition to active health AI will be difficult, observes Nishikawa at METI.

“I am optimistic that AI can help make people healthy; in Japan and all over the world people can easily deploy a combination of wearables, applications, and analytic systems to give incentives for making healthy lifestyle choices. However, I am not optimistic that AI will serve as a reliable substitute for the diagnostic services of a doctor, for a combination of reasons: accuracy, credibility and, most importantly, responsibility. A medical doctor must be responsible for their own diagnosis, including the responsibility to manage a patient’s anxiety levels” when presented with an unfavorable diagnosis.

AI should not fully substitute the diagnostic services of a doctor for three reasons: accuracy, credibility, and most importantly, responsibility.

“In this sense, we must be conservative in applying AI capabilities as a complete substitution. Diagnostic support systems today must supplement and support doctors.” The next logical step toward more active AI, Nishikawa believes, is in leveraging AI to fully support health-care professionals in specific contexts. He points to Tricog, a Singapore-based health-care company that has attracted investment from the University of Tokyo’s Edge Capital venture fund. Tricog’s technology links electrocardiography machines operated by clinics and hospitals in over a dozen countries in Africa, south Asia, and southeast Asia to its cloud analytic platform, sending electrocardiogram results for analysis and interpretation by its own medical team. The company claims that this combination of remote expert- and AI-enabled diagnostic support can provide customers with results in six minutes. The value of Tricog’s approach, Nishikawa says “is in providing doctor-to-doctor predictive services, to other professionals in countries where professionals have less experience.”

This is an extract from “AI in health care“, a report by MIT Technology Review Insights.