Artificial intelligence Breakthroughs happen in computer science labs or tense board games on television between a person and a machine. The recent advancement in medical AI has less glamorous origins: the depths of the U.S. government’s bureaucracy.
The U.S. Centers for Medicare & Medicaid Services (CMS) recently announced it would use two AI systems: one that can be used to diagnose a blindness complication and one that will notify a specialist if a brain scan suggests that a patient has suffered a stroke. The decisions aren’t just remarkable for Medicare and Medicaid patients – they could help advance the use of AI in healthcare significantly.
Both products are already approved by the Food and Drug Administration and are used by some vendors. However, new devices and treatments are generally not deployed until the US government approves payments for Medicare and Medicaid patients. Private insurers often decide whether to cover a new invention of CMS, although they usually pay higher rates.
Last month, CMS began paying for the use of ContaCT AI software from San Francisco startup Viz.ai as part of a program that encourages the adoption of new technologies. ContaCT is installed in a hospital emergency room to alert a neurosurgeon when algorithms on a CT scan see evidence that a patient has a blood clot in the brain. Speeding up stroke diagnosis and treatment, even by minutes, can dramatically reduce a person’s disability and recovery time.
The WIRED guide to artificial intelligence
Supersmart algorithms don’t do all of the work, but they learn faster than ever and do everything from medical diagnostics to showing ads.
The agency has also announced that it will soon pay for IDx-DR software, which analyzes photos of a person’s retina to diagnose diabetic retinopathy, a complication of diabetes that can lead to blindness. In August, CMS proposed paying for the software developed by Digital Diagnostics of Oakdale, Iowa.
Viz and Digital Diagnostics both received FDA approval in 2018, pioneering efforts to convince regulators that AI can improve health outcomes. IDx was the first AI product approved to diagnose disease, a clinical appeal previously only made by human doctors. CMS’s belief that tax dollars should be spent on medical AI could be seen as a similar milestone.
“This is very important for everyone in AI,” says ophthalmologist Michael Abramoff, CEO of Digital Diagnostics. The proposal to pay for IDx would also cover other AI tools that diagnose diabetic retinopathy.
The government’s willingness to pay for using AI tools could be good news for other companies working on medical AI products. Analysts at CB Insights report that AI Health Care invested $ 4 billion in startups in 2019, up from nearly $ 2.7 billion in the previous year.
CMS rated the brain and eye scans very differently – highlighting the complexity of the U.S. healthcare system and a challenge for new AI technology. In any case, the agency was faced with a question with philosophical dimensions: How do you value the work of software that by its nature performs a task that is normally carried out by a highly qualified person? There were two very different answers.
In the case of Viz’s ContaCT, CMS ruled that hospitals should receive up to $ 1,040 for use of the software for certain patients, and cited evidence that doing so significantly improves stroke management. The agency considered whether AI software, which merely speeds up the work normally done by humans, is new enough for a program reserved for new technologies, noting that “human intelligence and human processes are not approved by the FDA approved or approved technologies are “.
Chris Mansi, CEO and Founder of Viz, says the approval of CMS has already encouraged more hospitals to sign up for ContaCT. It has previously been used in around 500 large hospitals that have operated without the promise of government reimbursement because faster identification of stroke patients can increase the number of lucrative, time-sensitive surgeries they perform. “The people who benefited from it were important hubs,” says Mansi. “Now more hospitals can use the software and get paid for it.”
“It’s very important to everyone in AI.”
Michael Abramoff, CEO, Digital Diagnostics
CMS has suggested that hospitals pay much less for retina checking software like IDx. Abramoff says it would generally be less than $ 20, though the amount varies depending on geography and other factors.