If we’re serious about healthcare equity and access, we must support autonomous AI

The current COVID-19 public health crisis has converged with the racial and socioeconomic injustices that plague our society, highlighting vast differences in health care access. Though the U.S. spends more on health care than any other developed nation, access to critical preventive care remains particularly challenging for many racial and ethnic minority populations, as well as lower-income and rural Americans.

For many, visits to a specialist for a routine diagnostic exam represent a time-consuming and costly venture. As a result, many serious conditions go undiagnosed until they are advanced when treatment becomes more expensive and invasive, and outcomes are less favorable. In 2019 a record number of U.S. adults (33 percent) said they put off receiving medical care due to cost, according to Gallup’s annual Health and Healthcare poll. In 2020, even more people are delaying specialist care during the pandemic.

It doesn’t have to be this way: People should not suffer needlessly from preventable diseases. For better health outcomes, early diagnosis and timely treatment for all — rather than only privileged subgroups — are key at both patient and population levels.

While not a magic bullet, autonomous artificial intelligence (AI), which makes medical diagnoses without requiring human input or interpretation, can bring affordable, accurate diagnostic exams for many serious conditions directly to patients, without a specialist visit. In this model, patients receive diagnostic exams for skin cancer or diabetic retinopathy wherever it is convenient for the patient — whether a primary care office, public health clinic, or another care environment entirely. Fully autonomous AI diagnosis systems require only electricity and a non-specialist operator. 

One example is the first FDA-authorized autonomous AI diagnostic system, IDx-DR. It diagnoses retinopathy and macular edema, both leading causes of blindness, and its safety has been extensively validated in clinical trials.  A recent John Hopkins School of Medicine study published in the Journal of the American Medical Association reinforces this, showing that diabetic eye exams are effective and cost-saving for patients. 

To fully realize its potential, autonomous AI needs to be financially sustainable for providers and patients – and it needs to have the public’s trust. Tremendous progress has been achieved in developing an ethical foundation for autonomous AI and ensuring accountability by making the creator of the autonomous AI — n the IDx-DR case, the company that developed it, Digital Diagnostics — liable for its safe performance, ensuring accountability. 

The federal government has also taken a giant step forward in the acceptance and adoption of autonomous AI in health care. In August, the Centers for Medicare & Medicaid Services (CMS) made a historic decision: to reimburse the use of autonomous AI in health care, for the first time. This proposed rule comes at an especially important time as health care providers look to catch up on their backlog of pandemic-delayed testing. This milestone should be applauded, as reimbursement is key to bringing autonomous AI into more communities, expanding its benefits to more patients.

That said, reimbursement is unrealistically low as currently proposed. It will not sustain continued research and development to bring these technologies’ full value to patients, particularly those in underserved communities. Just as parity in reimbursement for telehealth was necessary to expand care access during the pandemic, higher reimbursement rates will be needed to foster autonomous AI and encourage adoption by primary care and frontline providers. 

It may not be intuitive, but adequate reimbursement rates for autonomous AI will drive down our nation’s health care costs immediately, increase access, and provide needed support for our struggling primary care system — which has been particularly hard hit. 

To incentivize the use of fully-vetted, transparent, FDA-authorized autonomous AI diagnostics, we urge CMS to consider raising the proposed reimbursement for autonomous AI. In doing so, together, we seize this historic opportunity to bring meaningful improvement to the health care system and increase health care access for millions of Americans, especially those who need it most.

Former Sen. Bill Frist, M.D., is a heart transplant surgeon, former U.S. Senate Majority Leader, and the Vanderbilt Transplant Center, founder. He serves on the board of autonomous AI company Digital Diagnostics.

Michael D. Abramoff, M.D., Ph.D., is a retina specialist at the University of Iowa Hospitals & Clinics and the Iowa City VA Health Care System. He is the founder and executive chairman of Digital Diagnostics, the first company in any field of medicine to get FDA de novo authorization for autonomous AI.