Q&A Session Following the Opening Keynote at the 2026 WCRI Issues and Research Conference
Following his opening keynote at the 2026 Workers’ Compensation Research Institute Issues and Research Conference, Harvard health economist Dr. Benjamin Sommers sat down with WCRI CEO Ramona Tanabe for an extended Q&A session that drew pointed questions from the audience. If his keynote laid out the landscape of what’s changing in American health coverage, the discussion that followed explored the practical consequences — for states, for hospitals, for workers’ comp carriers, and for the patients caught in between.
A Stealth Partial Repeal, Not an Outright One
Tanabe opened by asking Sommers where he sees the coverage landscape heading over the next several years. His answer was measured but direct. Sommers noted that when Republicans controlled Congress in 2017, full repeal of the Affordable Care Act came within a single Senate vote — the famous thumbs-down from the late Senator John McCain. This time around, nobody has seriously proposed full repeal. The ACA, he observed, has grown more popular over the past decade, making outright elimination politically untenable.
What the One Big Beautiful Bill Act represents instead, Sommers argued, is a quieter undermining of the law’s coverage gains. Work requirements don’t repeal the Medicaid expansion, but they erode it. The subsidy expiration doesn’t eliminate the marketplaces, but it prices millions of people out. As he described in his keynote, roughly 40 million Americans gained coverage through ACA-related programs at the peak. Sommers now expects the country to “backslide” by 5 to 6 million on the Medicaid side through work requirements and another 3 to 4 million through the marketplace subsidy loss — landing somewhere between the pre-ACA world and the historic coverage highs of 2023.
“Some would call this a stealth partial repeal,” Sommers said. “I think that’s a fair reading of it.”
Arkansas Was Supposed to Be the Cautionary Tale — It May Be the Best-Case Scenario
When Tanabe pressed on the evidence base behind work requirements, Sommers expanded considerably on the Arkansas data he had presented in his keynote. Arkansas remains the only state that fully implemented a Medicaid work requirement, and the results were sobering: 18,000 people disenrolled, uninsured rates climbed, and employment didn’t budge. But Sommers revealed that two other states offer corroborating evidence, even though their programs never fully launched.
New Hampshire, he explained, had received federal approval and built out its verification infrastructure. State officials were confident they could avoid Arkansas’s mistakes. But when they examined their own numbers just before beginning to remove noncompliant enrollees, they discovered they were performing just as poorly — large numbers of likely-eligible people hadn’t completed the paperwork. New Hampshire hit pause rather than proceed.
Georgia took a different approach, pairing a partial Medicaid expansion with work requirements as a front-door condition. The state initially projected 100,000 enrollees in the first year, later revised that down to 30,000, and ultimately enrolled fewer than 5,000 — a fraction of the estimated 300,000 to 400,000 people who could have qualified under a standard expansion. Interviews on the ground told the same story: people didn’t understand the rules and couldn’t navigate the system.
Perhaps most striking was Sommers’ reassessment of what the Arkansas experience actually tells us. He had long described it as a cautionary tale, but the more he examined it, the more he came to believe Arkansas may represent something closer to a best-case scenario. Arkansas had relatively strong data systems — it could cross-reference Medicaid rolls with SNAP enrollment and medical frailty designations to automatically exempt many beneficiaries. Many other states lack those capabilities. The people who lost coverage in Arkansas were disproportionately those the state couldn’t automatically verify, and even then, more than half of them were removed. States with weaker data infrastructure, Sommers warned, could see significantly worse outcomes.
The Two Big Misunderstandings
Asked what the public most misunderstands about work requirements, Sommers identified two things. The first is the assumption that Medicaid is full of able-bodied adults choosing not to work. The data show otherwise: roughly 40 percent of the affected population is already employed, another large share has health-related limitations, and only about 3 to 5 percent are neither working nor exempt for an obvious reason. The policy targets that narrow sliver while imposing paperwork burdens on everyone else.
The second misunderstanding is more subtle. Many of the legislators who voted for the One Big Beautiful Bill Act, Sommers said, told him they supported the concept but wanted to make sure vulnerable populations — people with disabilities, those in treatment for substance use disorders, pregnant women — would be protected through exemptions. The problem is that writing exemptions into law doesn’t mean they function in practice. The pandemic unwinding and the Arkansas experience both demonstrated that eligible people lose coverage in large numbers when they can’t navigate administrative processes, regardless of what the statute says should happen.
He predicted that by 2027, media coverage of real people losing Medicaid despite qualifying for exemptions will force policymakers to revisit the issue. Public support for work requirements drops sharply, he noted, when people learn that eligible individuals are losing coverage as a side effect.
Workers’ Comp as a “Soft Target”
The most direct workers’ compensation connection came from an audience member — Joe Paduda of Health Strategy Associates — who framed the issue bluntly. With workers’ comp representing just 0.74 percent of total U.S. medical spending, roughly $31 billion, and with hospitals deploying increasingly sophisticated revenue cycle management tools, Paduda argued that workers’ comp is “uniquely incapable of fighting back” and represents a soft target for providers seeking to maximize reimbursement as other coverage sources shrink.
Sommers agreed, drawing an important distinction. The traditional cost-shifting hypothesis — that hospitals simply raise prices on private insurers when public coverage shrinks — doesn’t hold up well in the economics literature. Hospitals generally negotiate the highest rates they can regardless of their payer mix. But claim shifting to workers’ comp is a different matter entirely. When a worker loses Medicaid and faces a gray-area injury that could plausibly be filed under either program, the incentive to route it through comp becomes much stronger. That dynamic, as Sommers documented in his keynote, has been consistently observed in the research: coverage expansions reduce comp claims, and coverage contractions increase them.
The Cost of Running Work Requirements
Paduda also raised the question of implementation costs — how much states actually spend to administer work requirement programs versus what they save by covering fewer people. Sommers acknowledged that covering fewer people does reduce spending, but described the administrative overhead as substantial and largely unrelated to healthcare delivery. Georgia, he noted, spent hundreds of millions of dollars building eligibility verification infrastructure that enrolled only a few thousand people. The federal government has set aside implementation funds for states under the new law, but Sommers said preliminary estimates suggest it falls well short of what will be needed. States will face a painful choice: implement the program poorly and watch eligible people lose coverage, or divert already-strained budgets toward administrative compliance.
Rural Hospitals, the Data Gap, and International Context
Several other questions rounded out the session. On rural hospitals, Sommers noted that while the Rural Hospital Transformation Program created under the new legislation will provide some financial relief, its formula distributes funds relatively evenly across states rather than targeting the expansion states where coverage losses will be most concentrated. Preliminary analyses suggest the program won’t offset the revenue losses hospitals will face from newly uninsured patients. He pointed out that rural hospital closures over the past decade have already been disproportionately concentrated in non-expansion states, and expansion states may now begin experiencing similar financial stress.
On data collection, Sommers made an impassioned case for robust federal monitoring of work requirement implementation. During the pandemic unwinding, the Centers for Medicare and Medicaid Services required states to report detailed data on who was losing coverage and why — information that enabled the kind of research he presented in his keynote. Whether the current administration will impose similar reporting requirements remains unclear, and Sommers described the research funding environment as increasingly difficult.
An audience member asked about GLP-1 medications and whether insurers are too hasty in curtailing coverage. Sommers, pivoting to his clinical perspective, called the drug class genuinely impressive — the evidence keeps getting stronger with each new study, covering cardiovascular disease, sleep apnea, liver disease, and addiction in addition to obesity. But he cautioned that most effective healthcare interventions still cost more than doing nothing, because you treat many people to prevent adverse outcomes in a few. The real challenge, he said, is that insurers making coverage decisions today rarely capture the long-term savings, because patients switch plans and employers every few years. That misaligned incentive structure, he noted, leads to chronic underinvestment in prevention across American healthcare.
Asked what the U.S. could learn from other countries, Sommers offered a wry observation: plenty, but American policymakers are “really uninterested in hearing about other countries.” The U.S. remains an outlier in both its uninsured population and its prices, paying more than any peer nation with outcomes that are no better.
The Physician’s Perspective
Tanabe closed by asking Sommers to speak as a doctor rather than an economist. He described a patient — a veteran who had experienced homelessness and was enrolled in Medicaid with nominal copayments of a dollar or two per prescription. The man had 15 medications and asked Sommers to rank them in priority order because he couldn’t afford them all each month. Sommers told the audience he got the patient to at least number six on the list, but that clinically, the patient needed 12 or 13 of those prescriptions.
He also highlighted a finding from his recent research that carries particular relevance for the workers’ comp world: confusion itself is a health risk. During the pandemic, his team found that many Medicaid enrollees who were still covered believed they had lost their insurance because they hadn’t heard from their state in over a year. Those people behaved like uninsured patients — they delayed care and avoided doctors’ offices — even though their coverage was intact. As work requirements and more frequent eligibility checks layer additional complexity onto an already difficult-to-navigate system, Sommers warned that misinformation and confusion will drive coverage losses beyond what the policy itself intends.
For an industry that sits at the intersection of employment, healthcare access, and injury, the message from this session was hard to miss: the changes coming to American health coverage won’t just affect Medicaid rolls and marketplace enrollment numbers. They’ll ripple directly into workers’ compensation — through claim shifting, through sicker workers with less access to preventive care, and through a healthcare system under increasing financial strain looking for every available dollar.