Team Biden needs to recognize that health care innovation using AI is just what the doctor ordered

During the middle of the 20th century, scientists and social theorists began to fear the problem of overpopulation, predicting a period of mass starvation. 

Famously, Stanford’s Paul Ehrlich, in his 1968 book, ‘The Population Bomb’ predicted ‘the battle to feed all of humanity is over…hundreds of millions of people will starve to death in spite of any crash programs embarked upon now.’ 

At the time, his pessimistic thinking was not isolated. Simultaneously, Norman Borlaug became a pioneer in wheat production with his work in genetics powering new ways to grow crops. His ‘Green Revolution’ for which he received the 1970 Nobel Peace Prize, is credited with saving over a billion lives. 

Innovation, a tried and tested wire cutter, defused the population growth bomb. The same is true about the Biden administration’s pessimism-driven regulatory obsession with artificial intelligence that aims to replicate these past mistakes.

Much has been written about the innovation in the life sciences sector with new gene therapies repealing death sentences and medical devices transforming hospital-based surgeries into outpatient procedures. But little attention has been paid to the lack of innovation in health care delivery itself.

AI offers our country the potential to put health care back in the hands of the patient. 

Analysis by the Bureau of Labor Statistics demonstrates that private community hospitals exhibited negative labor productivity growth for over the preceding two decades, with productivity declining 5.6% in 2020. In addition to suffering from the ills of monopoly, health care is suffering the absence of a key gene at the heart of the life sciences industry: innovation.

Despite the fearmongering present in Washington, AI offers the opportunity to unleash innovation in service delivery. With many physicians spending 87% of their day bent over a keyboard, AI can function as stenographer, generating clinical notes and allowing physicians to focus on the patient in front of them.

Automation of the mundane is one of several potential patient-facing innovations. Clinical care requires split second real-time decision-making, with AI-driven technologies akin to lane departure and radar-directed cruise control in cars driving safer and more effective care.

Automation of clinical practice can also expand access while improving quality, benefitting the poorest Americans the most. The U.S. cannot train physicians and nurses fast enough to meet our needs. Upskilling and transforming how clinicians work is critical to a twenty-first century delivery system. AI is already being used to read electroencephalograms, digital cytology, and diagnose diabetic retinopathy.

Despite the true promises of the technology, Washington is focused on top-down regulations with many calling for a new independent agency to oversee artificial intelligence and digital platforms. Yet, artificial intelligence is a platform technology, not a tenet of policymaking. Flexible performance-based oversight enacted through issue-specific agencies, or in the case of health care, the FDA, offers a more pragmatic approach.

Such performance-driven policies do not require consumer-facing transparency and direct, constant consumer control. Occupants riding in a car do not decide before the moment of impact whether they desire the support of an airbag. 

Similarly, patients and physicians need to know that AI-driven technology performs as expected in a range of environments whether an integrated insulin pump and glucose monitor or clinical decision support software recommending adjustment to ventilator for an intubated patient in the operating room.

Yet, the Biden administration’s recent 804-page health technology rule undermines this objective and instead focuses too heavily on algorithmic and AI transparency in health care bypassing performance. While transparency is important, the rule is nonsensical at its core as it burdens those who would benefit most from AI with administrative burden and will ultimately stifle the use of innovative AI-driven products. 

Surgeons will not stop operating in order to read the evidence underlying AI-based technology or clinical decision support. With over 1 million new medical papers published annually, physicians and patients do not have the time to read a government-mandated research summary.

Patients and physicians instead must depend upon product performance as a policy goal. An independent network of technical standards development organizations and testing labs can support AI applications in health technology along with a light-touch oversight environment at FDA. Flexible guidance on training datasets and population representativeness, coupled with testing parameters for a variety of clinical situations can help ensure that technology development remains vibrantly decentralized and disruptive.

AI offers our country the potential to put health care back in the hands of the patient. From more time with their doctor to automated diagnosis and treatment of basic health conditions to supporting medication adherence and health behavior change, AI has enormous positive potential to provide access to low cost, high quality care.

At a time of deep political division, Americans remain unified in their dissatisfaction with our health care system. The story of Norman Borlaug offers us a timely reminder that technology-driven innovation, not Washington pessimism, must be the beating heart of our health care system’s disruption.