Researchers from Weill Cornell Medicine and Boston University School of Public Health have been awarded more than $950,000 from Arnold Ventures to create a “Medicaid Atlas” — a national, data-driven web platform that will illuminate how health care use and spending vary across Medicaid programs, plans and populations. The project will be one of the first major efforts of the Medicaid Policy Impact Initiative, a new cross-campus Cornell program aimed at supporting evidence-based policymaking in Medicaid.
The project launches at a moment of growing urgency for Medicaid policymakers. As states face increasing fiscal pressure due funding cuts enacted under the One Big Beautiful Bill Act, leaders need actionable data to understand where spending is high, why it varies and where opportunities exist to improve care.
Dr. William Schpero
“Medicaid is a massive program. It’s one of the biggest line items in states’ budgets. And yet we still lack great visibility into what drives spending variation,” said Dr. William Schpero, an assistant professor of population health sciences at Weill Cornell Medicine and co-lead on the project.
Understanding those drivers has been difficult for researchers because Medicaid is highly fragmented across states and often delivered through private managed care plans. Two patients with similar health needs may generate very different spending depending on where they live and which plan they are enrolled in. At the same time, information about health care spending, use and quality has historically been siloed across states and insurers, forcing policymakers to rely on time-intensive, custom analyses to answer even basic questions.
This approach does not always support the rapid-cycle policy environment many Medicaid programs operate in. “With the atlas, we want to make it possible to go from months-long studies to making helpful data insights one click away,” said atlas co-lead Dr. Sarah Gordon, associate professor of health law, policy & management at Boston University School of Public Health and co-director of the BU Medicaid Policy Lab.
That goal has become more feasible in recent years with the release of high-quality national Medicaid claims data from the federal government, which for the first time enable consistent analysis across states. In 2022, Drs. Gordon and Schpero helped start the Medicaid Data Learning Network, a national consortium of researchers dedicated to developing best practices for analyzing federal Medicaid claims data. The research produced by this network has demonstrated the potential impact of national Medicaid data analyses to inform policy.
Dr. Sarah Gordon
“These data have proved to be a major catalyst for conducting timely Medicaid research that can directly inform state decision-making,” said Dr. Schpero, who is also an associate director at the Cornell Health Policy Center, which houses the Medicaid Policy Impact Initiative. “Our hope is that this tool can be a valuable resource for Medicaid leaders to understand variation within their state and benchmark to other states, all in a one-stop, easy-to-use dashboard.”
Over the first two years, the research team will develop a set of 10–15 measures capturing major drivers of Medicaid spending and utilization. These measures will be selected in collaboration with state Medicaid leaders to ensure they help answer high-priority policy questions.
“The goal of this project is to put Medicaid policy and financing insights at the fingertips of policymakers, researchers, journalists and the public,” Dr. Gordon said.
The atlas will allow users to examine variation across states, counties, plans and enrollee populations — and to identify patterns that may signal opportunities to improve efficiency and care delivery. States could use the atlas to benchmark spending, evaluate managed care plans and identify high-value models of care.
There’s precedent to think this approach will work, Dr. Schpero said. A decades-long project called the Dartmouth Atlas of Health Care used claims data to identify opportunities to lower spending while maintaining or improving quality in the Medicare program. Some of the project’s findings laid the groundwork for components of the Affordable Care Act.
“There’s a rich history of using national claims data to inform policymaking,” Dr. Schpero said. “To date, much less of this has happened in Medicaid. There is a lot that we can learn about Medicaid policy from these data and that states can learn from each other.”
The atlas is part of a broader effort at Cornell to strengthen the role of data and evidence in Medicaid policymaking. The Medicaid Policy Impact Initiative, directed by Dr. Schpero, brings together Cornell faculty, data infrastructure and state partnerships to produce research designed for real-world decision-making. The initiative focuses on how Medicaid programs are designed, financed and delivered — and translates those insights into practical guidance for policymakers, health plans and providers.
