Healthcare in the United States has long been a major business journalism beat, not least because the sector accounts for almost 20% of the U.S. economy, but with rampant inflation in healthcare costs over the past few years, it is increasingly top-of-mind for many American consumers.
So on July 9, the Society for Advancing Business Editing and Writing (SABEW) convened a panel in partnership with the National Institute for Healthcare Management Foundation to address what journalists need to understand about healthcare affordability and what changes or innovations – including artificial intelligence – are driving healthcare costs.
Panelist Larry Levitt is the Executive Vice President for Health Policy at the Kaiser Family Foundation (KFF). Marty Steffens, the SABEW Chair for Business and Financial Journalism at the University of Missouri, moderated.
Background
Levitt first spent some time contextualizing the environment of healthcare affordability, saying KFF’s latest polls have revealed 64% of voters are worried about being able to afford healthcare, and 41% of adults have some kind of debt for medical or dental bills, totaling over $200 billion in healthcare debt.
“Medical debt certainly has financial consequences. I mean, it causes people to cut back on other household needs. It can hurt people’s credit reports. But it also leads to poorer access to care,” he said. “In many ways, I think that medical debt is kind of the canary in the coal mine of our healthcare.”
Levitt also explained some of the biggest challenges people currently face with healthcare affordability, including the future impacts of H.R. 1, or the omnibus spending bill passed by Congress in July 2025, which is projected to cut Medicaid by $911 billion over the next decade. The complexity of navigating health insurance, from obtaining insurance to begin with to dealing with prior authorization – where some treatments or medications need to be approved by health insurance before receiving care – also places a burden on patients who may forgo necessary care to avoid financial strain.
“A small percentage of people use the vast majority of healthcare, and it’s for chronic and serious illnesses,” he said. “We don’t emphasize prevention and primary care nearly to the extent they do in other countries. That’s part of the reason we spend almost double on healthcare per person compared to other high-income countries.”
Where journalists should be looking
One of the biggest reasons Levitt thinks journalists struggle to cover healthcare affordability is the idea of timeliness, as these challenges have been around for decades. Levitt and Steffens discussed how, even in the 1970s, then-President Nixon called healthcare costs unsustainable.
“Healthcare is kind of the slowly boiling frog,” he said. “Healthcare costs have risen dramatically over time, much faster than inflation, much faster than workers’ wages, faster than we can afford, but it’s like a steady drip…If something’s been a crisis for decades, it’s hard to make it a current story.”
In addition, healthcare is such a vast and varied system that it’s difficult to narrow down into a neat story.
“Gas prices, you can go to the corner, you can see the gas prices posted, there is nothing like that in healthcare,” Levitt said. “Healthcare is multiple different kinds of services, multiple different industries. There is no single price, there’s not really transparent prices yet, it varies from patient to patient based on the services they’re getting, so there’s not always that one number that allows you to write about it clearly.”
Despite the complexity, Levitt said there are still interesting avenues for journalists to take, especially considering overlap with politics, labor, and technology.
“The government plays such a big role in healthcare,” he said. “Much of healthcare spending is now made by the government. So even what may appear to be a healthcare business story is still a politics story, because you can’t take the politics and policy out of healthcare.”
Artificial intelligence
But how has such a long-standing ecosystem been adapting to the rise of artificial intelligence? While the use of AI in the industry is on the rise, Levitt said it’s still too soon to know exactly what impact it will have.
“I think we’re going to soon be into issues around what type of care can AI provide. There’s a pilot going on in Utah now where AI can renew a prescription. I think we’ll very quickly be into a debate about whether an AI app can actually write a prescription,” he said.
Levitt provided an example of how artificial intelligence might be improving the patient experience by allowing doctors to be more present and attentive in an appointment and allow technology to handle the administrative side. However, he added that the same AI that enables this connection can backfire.
“I think it’s very likely that AI is increasing costs right now, not decreasing,” Levitt said. “Those systems are very effective at upcoding or revenue cycle management, basically increasing costs.”
Upcoding refers to a kind of medical billing fraud that inflates the severity of a patient’s diagnosis or treatment, letting billing departments get greater reimbursements than the actual care provided. This leads to higher healthcare costs for patients through higher insurance premiums, further impacting later access to affordable care.
Insurers are also now using AI to root out inappropriate claims and identify care that may not be needed, and providers and patients are now using it to write appeal letters when their claims are denied, turning the prior authorization process into “bots talking to bots.”
KFF surveys are also seeing a growing use of AI by patients, which Levitt said is partially driven by worries about healthcare affordability, though it’s unclear yet how that trend will impact medical costs.
While many aspects are still unclear, Levitt said this unpredictability is a good aspect for journalists to explore: “I think it’s a huge story in drug development, being able to identify potential new candidates for drugs. Healthcare is, I think, one of the biggest areas where AI is being applied but still very early on, I think, as is true in many other sectors.”
