AI and Health Equity in California

This April, the California Health Care Foundation (CHCF) hosted a discussion on artificial intelligence (AI) and its effect on health care delivery in California. CHCF is an independent, nonprofit philanthropy working to improve the state's health care system and provide individuals with the care they need; especially those from communities that experience significant barriers to care.

The session’s moderator, Kara Carter, senior vice president of Strategy and Programs at CHCF, called out the potential of AI to transform health care in a positive way. According to Carter, AI can help doctors sift through mountains of data to diagnose diseases faster and more accurately, while making real-time updates to electronic medical records. This will allow physicians to reduce what Carter refers to as “pajama time” (time spent charting and reviewing patient notes early in the morning or late at night), while reducing burnout, and ultimately increasing job satisfaction. “Providers will be able to prioritize patient interactions even more,” Carter said.

Carter also emphasized the real risks AI poses, including concerns about data privacy, irresponsible oversight, inequitable access to the benefits it can bring, and algorithms that perpetuate bias and health inequity. Thus, Carter argues it is essential policy leaders ensure that AI is responsibly deployed to the benefit of the public. “The goal of AI should be to support and augment the workforce instead of replacing it, and to build trust instead of suspicion in the health care system,” she said.

AI Opportunities in the Safety Net

Following her address, Carter welcomed keynote speaker, Mark Ghaly, MD, MPH, secretary, California Health and Human Services. Dr. Ghaly’s discussion focused on the impact of AI on California’s safety net: a grouping of programs and providers across the state that specialize in coverage and care for patients with low incomes, no insurance, and/or special medical needs.

According to Dr. Ghaly, adopting new technologies is typically delayed for providers within the safety net when compared to larger organizations with more resources. Even today, not all providers within the safety net are using an electronic health record (EHR). Thus, Dr. Ghaly emphasized that this trend must not be replicated with the adoption of AI.

Dr. Ghaly believes it is important to balance both the fears and the optimism of emerging technologies. “One of the areas that we have significant fear around,” Dr. Ghaly said, “is whether the government can actually create a regulatory framework that supports and secures equity, quality, and access in all the ways we can imagine.” These are valid concerns, that invite and enlist policymakers to pave the way for responsible and equitable use of AI across all communities.

Dr. Ghaly also provided a useful case for AI in the form of a community health worker. Daily, they may use AI to capture notes during visits with clients, to make better sense of the data collected during those visits, or to develop culturally competent and linguistically appropriate follow-up. “[It] requires us to look at tools like AI and other technologies not with fear, but with confidence that together, we can imagine ways to use it more effectively,” he noted.

To prevent the misuse of AI, Dr. Ghaly spoke to the need for core values that guide the deployment of AI tools, including addressing health inequities, promoting accountability and transparency, generating demonstrated value, empowering human capacity, strengthening community, cultivating innovation, and improving accessibility. In response to health care workers’ worry that AI will replace their roles, Dr. Ghaly called on policymakers and leaders to guide those conversations and be sensitive to the workforce impacts of AI. “This technology shouldn’t be a threat to jobs, but an aide for health care workers to free up their time for the more fulfilling, patient-facing aspects of their roles,” he said.

Dr. Ghaly closed by reiterating the importance of collaboration as the health care continuum works to leverage emerging technologies to build a better future and improve outcomes for all patients.

Framing AI

Next, panelists Susan Ehrlich, MD, MPP, CEO of the Priscilla Chan and Mark Zuckerberg San Francisco General Hospital and Trauma Center, and professor of medicine, UCSF; Carolina Reyes, MD, associate clinical professor, Maternal Fetal Medicine, School of Medicine, UC Davis Health, chair of the board of directors of the California Health Care Foundation, and board member of Common Spirit Health; and David Lubarsky, MD, MBA, FASA, vice chancellor of Human Health Sciences and chief executive officer, UC Davis Health, were invited to speak about their experiences using AI in their respective programs.

Dr. Lubarsky discussed trust as a significant barrier to the practical implementation of AI. “The speed of trust determines the speed of change,” he said. “You can’t get anywhere unless people actually believe AI is a good thing.” He also expressed that this resistance is often healthy and necessary, especially when providers demand proof that new technology will not negatively affect patient safety and care.

Dr. Lubarsky then reframed AI as “augmented intelligence” rather than artificial intelligence. “It is not artificial,” he explained. “It simply processes large amounts of data more quickly and accurately than a human can do on their own.” He echoed Dr. Ghaly’s sentiment that AI should be viewed as an assistant—not a replacement—for health care workers, capable of drawing their attention to what needs to be done.

Equity Concerns

Carter then posed the question of how the health care community can ensure these benefits of AI reach all communities in California, not just the wealthy ones. Dr. Ehrlich acknowledged that AI requires vast resources in the form of expertise, time, and money. She also expressed that her hospital, Zuckerberg San Francisco General, has witnessed the effects of the disparity in resources firsthand. “We’re watching our colleagues at UCSF move far ahead of us in their ability to take advantage of these technologies—AI, machine learning—just because they have more resources than we do,” Dr. Ehrlich said.

Dr. Reyes added that in the discussion of AI implementation, it is crucial for providers to keep their guiding principles at the forefront. If a new tool or method doesn’t allow them to treat patients with dignity and give them the best care possible, it should not be used. She was also hopeful that the introduction of AI will allow patient interactions to be much more focused on the whole health of the person, rather than a single symptom: “It’s really about the provider having that full data around that 1 individual patient…How do you incorporate the data knowing where that patient lives, what their situation is like, and their health condition to know what has worked and what hasn’t worked for that patient." 

Building Trust in the Community

According to a recent survey by CHCF, more than 50% of Californians have expressed concern about AI being used in their health care. In response, Dr. Lubarsky emphasized the need for providers to be always transparent and accountable. Further, he believes that organizations who have successfully used AI need to share their results with everyone. “It’s not a competitive advantage to take better care of a patient,” he explained. “It’s a moral imperative.”

Dr. Reyes concluded by highlighting that the importance of every level of the workforce, from clerical workers up to the executive level, to have the cognitive and ethical skills to make judgment calls where AI is concerned—rather than assuming everything an AI tool produces is accurate.