February 08,2024

Wyden Statement at Finance Committee Hearing on AI in Health Care

As Prepared for Delivery

This morning the Finance Committee meets to discuss the use of artificial intelligence or AI systems in health care, with a focus on how this technology is being used in federal health programs like Medicare and Medicaid. 

There’s no doubt that some of this technology is already making our health care system more efficient. But some of these big data systems are riddled with bias that discriminate against patients based on race, gender, sexual orientation, and disability. It’s painfully clear not enough is being done to protect patients from bias in AI. 

Just as I worked to ensure tech innovation would improve patient care in the 1990s with laws empowering telemedicine and digital signatures, Congress has an obligation to encourage the good outcomes from AI and set rules of the road for new innovations to deliver better care for Americans.

Today we’ll also discuss the role Congress and this committee must play in helping strike a balance between protecting innovation and protecting patients and their privacy with legislative proposals like my Algorithmic Accountability Act, which would tackle these concerns head on.

There are a lot of reasons to be optimistic about the potential of AI to improve health care. Today the industry is facing a host of challenges, all made worse by the strain the COVID-19 pandemic put on our health care system. There’s an ongoing workforce shortage, existing providers are facing high rates of burnout, health care costs are rising faster than wages, and there’s an ever-growing gap between the care that’s needed and the care actually being delivered to many Americans.

Already, AI tools are being deployed to reduce some of these pressures and ease strain on the industry and providers. Some doctors are using this technology to pre-populate clinical notes and emails to reduce workload, submit bills to insurers to reduce administrative waste, and even help with diagnostics. Primary care providers can use these tools to screen for certain diseases and connect patients with specialists for treatment, saving patients’ time and money, and leading to better, more timely care.

There is no doubt these technological innovations can improve care for patients in Medicare and Medicaid, while also improving workload for providers, many of whom are already stretched thin. But addressing these challenges with new technology shouldn’t mean worse patient outcomes, and sacrificing patient privacy. 

Unfortunately, there are clear, glaring examples of AI tools being developed with data that perpetuate racial biases, and deployed in ways that bypass important doctor expertise, leading to inadequate care for patients. 

The committee is lucky to have here today Dr. Ziad Obermeyer, who in 2019 discovered harmful racial bias in an AI tool developed by the health care company Optum – a subsidiary of UnitedHealth Group – and used by providers across the country to offer care management services. 

He found that the tool, on average, required black patients to present with worse symptoms than white patients in order to qualify for the same level of care.

This algorithm was available to thousands of doctors across the country, potentially impacting millions of patients. How could such a flawed system make its way into general use? The answer is simple: there was nobody watching. No guardrails were in place to protect patients from flawed algorithms and AI systems.

To make matters worse, the technology that insurance companies or health systems use can play a role in what care patients receive, and what services are approved or denied. And the Department of Health and Human Services does not – yet – oversee the use of these systems.

I think most of us here would agree there are many ways this technology can be used to improve health care and patient outcomes. However, as we increasingly rely on technology like AI to make decisions in every facet of our day-to-day lives, this committee has a responsibility to ensure there are guardrails in place to protect patients, particularly in Medicare and Medicaid, and I do not believe that current laws go far enough to achieve that goal.

My Algorithmic Accountability Act lays the groundwork to root out algorithmic bias from these systems. As applied to health care, my bill would require health care systems to regularly assess whether the AI tools they develop or select are being used as intended and aren’t perpetuating harmful bias.

I’ll close with this: I believe the same protections in my Algorithmic Accountability Act must apply to patients in Medicare and Medicaid. Here’s what’s needed most: transparency in how these tools are developed and used to foster trust, and accountability for how these tools are used in health care. These tools should also preserve the privacy of patients. Lastly, these tools should further equity in health care, not perpetuate harmful bias or disadvantage hospitals and providers who serve low-income patients or communities of color.

The Food and Drug Administration and the Office of the National Coordinator for Health IT have proposed new rules to address some of these issues. That’s a step forward. But they don’t go far enough. It’s clear more is needed to protect patients from flawed systems that can and will directly affect the health care they receive. I look forward to working with my colleagues on the Committee to identify ways we can protect patients and improve care going forward.