After the Pandemic, Health Care Will Still Be Broken
Artificial intelligence can help fix it.
Even before the COVID-19 pandemic, health care services around the world were due for a digital transformation; with resources relatively scarce and diagnostics increasingly complicated, the sector needed new technology to allow doctors to provide the best possible care. The pandemic has kicked that transformation into gear. Since it began, artificial intelligence-driven diagnostics, viral spread analysis, and telemedicine have become routine. Yet the transition will need to go much further if it is to alleviate the increasing strain on health care workers and infrastructure.
The pandemic, of course, has presented an enormous burden on the system. But the pressure won’t go away once the crisis ends given aging populations and proportionally fewer working-age adults to generate tax income. To future improve health care, it is time for us to push beyond simple digitization of records or use of phones for telemedicine. It is time to embrace artificial intelligence in health care.
AI has had a crucial, but often unrecognized, role in fighting COVID-19, from mapping the spread of the virus to its diagnosis. A World Health Organization report from February 2020 highlighted how AI has excelled at identifying deep patterns in huge swaths of data, generating actionable insights off the back of them. For example, Facebook’s suite of global disease prevention maps works on sharing anonymized location data, which can be used to determine where people are congregating and therefore transmitting the virus. Data like this not only helps to identify where the spread is likely to occur but can also be used to help policymakers implement restrictions on specific public places and allocate resources more effectively.
AI has been instrumental in identifying where the virus is spreading, not only in society but also within the body. China’s Alibaba, for example, has generated an AI program that can identify the coronavirus in seconds with 96 percent accuracy based on CT scans of patients’ chests. The same job would take a physician approximately 15 minutes and requires the visual analysis of more than 300 images. The quickest doctor money can buy is not a human but an AI program. AI machines may not have great bedside manner or have the intuition of a human, but they can certainly empower doctors to be more efficient.
Both governments and the private sector have started to understand the benefits of AI in health. Yonatan Amir, CEO of the Israeli health tech company Diagnostic Robotics, told Wired that “COVID-19 has been a huge accelerator for the digital health care space with mainstream adoption of different technologies, which many assumed would take a decade, happening in a matter of months.”
Many countries have recently invested in AI within their health care services. For example, the United Kingdom has announced a $346 million investment and the United States a $1 billion investment earmarked for nondefense AI funding, much of which could be focused on health care given the pandemic.
However, it would be less than ideal for this effort to be entirely reactive to COVID-19, which is not a long-term investment in the value of AI as a service-improving technology. The uptake of AI could decelerate as the world comes out of the pandemic. Yet ongoing support for AI in health care is essential if the world is to adapt to aging populations and fewer working-age adults. According to a 2019 United Nations report, the number of people over the age of 65 globally will double by 2050. The challenge may be most acutely felt in Japan, where a large majority of people see the aging population as one of the nation’s most pressing problems.
Ironically, improvements in health care have allowed many people to live longer and, in turn, endanger the very health infrastructure that created this state of affairs. At the same time, changing societal norms combined with a growing middle class mean that people are simply having fewer children in some countries. The latter is offset in many parts of Europe and North America with a migrant population that has a higher fertility rate. But countries like Italy and Japan do not have the same benefit.
These broad changes have led to an unbalanced dependency ratio, that is, the ratio of economically independent citizens to economically dependent ones. Spain is an example of this; by 2050, it is projected to have more dependent citizens than independent ones. This will have implications on government spending on health care and could lead to higher taxes for working populations, which may drive inflation, further squeeze the middle class, and stifle productive investment—therefore slowing economic growth. This is a downward spiral that benefits no one.
Yet AI could increase the value per dollar spent on health care, allowing more care to be provided to more people with the same number of doctors. Of course, there is no AI for impeccable bedside manner or human touch. However, technologies do exist to enable better health care delivery when those are not needed. For example, in the U.K., 50 percent of National Health Service appointments are now done remotely, and Bangladesh’s a2i project has made health diagnostics available through text messaging.
But besides the ability to provide precision and speed, the real power of these technologies is in the amount of data they collect, which can be used by machine-learning programs to identify a potential health issue much faster and earlier than physicians. In short, it allows individuals to prevent health care problems before doctors have to fix them. And despite privacy concerns for personal data, there is an appetite for this preventative approach. A report from pharmaceutical company GlaxoSmithKline found that the vast majority of Europeans would be prepared to monitor their own health through, for example, wearable technology.
Wearables are just the beginning. More sophisticated technologies can help delay or even cure complex health issues. For example, AI programs can identify the development of Alzheimer’s disease in older patients much earlier than doctors. It is predicted that by the mid‐21st century, the number of U.S. citizens ages 65 and older with Alzheimer’s dementia may grow to 13.8 million people. Detecting the condition earlier will enable patients to access treatments sooner, leading to better outcomes.
Preparing for the coming demographic shift will be one of the biggest long-term challenges for governments around the world. The biggest blockage to adoption of AI will likely be cultural, especially by health professionals who may feel undermined or threatened by the technology. The next generation of doctors and nurses need to be digitally native when it comes to these technologies. Medical school curriculums, which are sometimes poor at keeping up with technological trends, need to be future-proofed to make using AI as intuitive as using a computer—and make clear that AI will free up doctors to give the kinds of care they want to be giving.
The biggest benefit of AI in health care is that it will allow people to be productive for longer. By detecting the early onset of health problems and by allowing people to take a more personalized and proactive approach to their own health, AI can be instrumental in allowing people to be physically, mentally, and economically active for longer. That will unleash the untapped potential of older populations and ease looming strains on the world’s health care systems.