How Should Children’s Medicaid Eligibility Be Monitored?

Background on Medicaid Continuous Coverage

Medicaid is a critical source of health insurance for children in the United States, covering nearly 4 in 10 (38.8%) in 2023, second only to employer-based insurance (48.7%).¹ However, coverage through Medicaid and the Children’s Health Insurance Program (CHIP) can be disrupted from year to year by annual requirements to ascertain, or to “redetermine,” ongoing eligibility for the program. Where enrollees can’t have their eligibility automatically renewed based on data states can access (eg, state wage databases), they need to provide information to the state substantiating their eligibility. During the COVID-19 pandemic, as part of a federal law intended to strengthen the safety net, states suspended these redeterminations pursuant to a requirement to keep Medicaid enrollees—children and adults alike—continuously covered until the end of the public health emergency, and they received additional federal funds to do so.² With the termination of the continuous enrollment provision effective March 31, 2023, as part of the Consolidated Appropriations Act (CAA) of 2023,³ early that year states began to roll back Medicaid and CHIP continuous coverage. During the “unwinding” of continuous coverage and resumption of eligibility redeterminations (which typically occur on an annual basis), states were obligated under the act to report metrics on terminations and renewals resulting from Medicaid eligibility redeterminations—the first time that reporting and publishing of such outcomes in any standardized way has been required.³

Unwinding jeopardized health insurance for millions of Americans, attracting significant attention from policymakers and researchers alike. Research suggests that gaps in health insurance coverage—and even transitions to other kinds of coverage without a gap—can leave people exposed to financial risk and affect access to care. In particular, people who experience gaps in coverage are more likely to delay needed and preventive care, skip prescription refills, and use the emergency department more often.⁴ During unwinding, experts voiced concerns about avoidable “administrative churn” (disenrollment among people who remained Medicaid-eligible but failed to navigate the renewal process); early projections suggested that nearly half of those disenrolled would be people who remained eligible.⁵ Other experts highlighted evidence that people who lose eligibility for Medicaid due to increased income—but become eligible for subsidized HealthCare.gov Marketplace coverage—often fail to successfully transition, resulting in avoidable coverage gaps.^6,7

Following this experience, there has been considerable interest in strategies and best practices to improve Medicaid—and CHIP—redetermination processes. But researchers only recently started measuring redetermination outcomes to rigorously identify and evaluate such strategies. Failure to continue monitoring redetermination outcomes—which varied demographically and geographically—would stymie progress to improve redetermination processes and inhibit policymakers’ ability to identify and implement policies that could reduce unnecessary lapses in coverage.⁸

Fallout of Medicaid Unwinding

As of September 2024, 31% of individuals (representing about 25 million people) who had undergone eligibility redeterminations had been disenrolled from Medicaid and CHIP during the unwinding process, which began in April 2023.⁹ As of September 2024, most terminations (over two-thirds) were “procedural,” meaning the state could not definitively ascertain eligibility.^9,10 Some people whose terminations fell into this group might have acquired other health insurance but not withdrawn from Medicaid.^11,12 Some might have believed they no longer qualified for Medicaid due to changes in income or circumstances and so did not respond to renewal paperwork. Others—the group that concerned policymakers, advocates, and researchers most—might have lost Medicaid because they never received or could not complete renewal paperwork, despite their eligibility. Overall, enrollment in Medicaid among adults declined by 19% between March 2023 and December 2024.¹³ The drop in Medicaid and CHIP enrollment among children during this period was about 11%.¹³

Redetermination outcomes varied considerably across states during this process.⁹ These uneven results reflected a combination of implementation choices during unwinding, extant differences in state programs and policies, and variation in the capabilities of state eligibility and enrollment systems that process redeterminations.^14,15 They also reflected noncompliance issues, which particularly impacted child enrollees. First, a number of states initially undertook eligibility redeterminations at the household, rather than the individual, level. This matters because children may remain eligible for Medicaid (or CHIP) even if their parents do not, given more generous income eligibility thresholds for children. Second, at redetermination, some states failed to consider eligibility for both Medicaid and CHIP; children may lose eligibility for one program while becoming eligible for the other, depending on family income (CHIP has more generous income thresholds).

The heterogenous redetermination outcomes represent an opportunity to learn about what helps—and what does not help—to keep eligible individuals continuously enrolled in Medicaid and CHIP.¹⁶  However, the lessons from this experience are also limited in a number of ways. Few states separately evaluated outcomes for adult and child enrollees or considered CHIP redeterminations separately from Medicaid redeterminations. Moreover, interviews with state officials suggest that the unwinding strategies implemented with children in mind were limited due to the overwhelming workload Medicaid agencies faced during the unwinding process.¹⁶

A Nascent Data Ecosystem

While Medicaid disenrollments and churn were common before unwinding, our understanding of these outcomes and their drivers were severely constrained by data limitations. Federal Medicaid enrollment and claims data available to researchers do not contain information on reasons for plan exit. One study attempted to obtain pre-2020 performance indicator data on procedural terminations in Medicaid and CHIP from the Centers for Medicare and Medicaid Services (CMS) with a Freedom of Information Act request, but only 13 states had data of sufficient quality to support analysis.¹⁷

As mentioned, the federal mandate for states to report redetermination metrics as part of the CAA of 2023—the legislation that also ended the continuous enrollment provision—marked the first time these outcomes were collected and published in any standardized fashion.³ Starting in April 2023, the law required states to provide CMS with the number of Medicaid and CHIP enrollees who were procedurally terminated, the number found to be truly ineligible for coverage, the number who completed renewals through the ex parte process versus using a form, and the number whose coverage was terminated and applied at, or had their information transferred to, the HealthCare.gov Marketplace.¹⁸ States also reported the number of redeterminations that had been initiated but not completed by the time of the report (“pending” cases).¹⁸ These data enabled researchers, policymakers, and the general public to access information in a relatively straightforward way about the total number of people redetermined in a given month and the share of people who were disenrolled in each state.

Many states publicly posted their reports contemporaneously with their submission to CMS, allowing for near-real-time evaluation of unwinding outcomes. CMS harmonized the data across states, publishing complete records with a lag of about 3 months, consistent with other enrollment data the agency has historically published.¹⁹ In early 2024, the number of renewals reported began to reflect a combination of those attributable to unwinding and those for enrollees who entered the program after the continuous coverage provision had sunset.

The CAA specified that states would need to continue reporting redetermination metrics through June 2024.³ However, in May 2024, CMS announced that it was exercising its regulatory authority to continue collecting and publishing these metrics to “maintain transparency into Medicaid and CHIP renewal outcomes at the national and state level.”²⁰ Because doing so constituted a discretionary regulatory action, current and future administrations could quietly rescind the requirements. Enrollment snapshots reflecting redetermination outcomes, which had been released on a roughly monthly cadence, were temporarily suspended from mid-January 2025 until the end of March 2025.²¹ In addition to hindering insights into the consequences of eligibility redeterminations for coverage and access, suspending publication of data on redetermination outcomes makes it more difficult to understand the extent to which states are spending taxpayer dollars on avoidable administrative costs of disenrolling and reenrolling people—and on the costly, but perhaps inefficient, systems that drive those outcomes.²²

Expanding Data Collection

Ongoing reporting of Medicaid renewal metrics is a critical development that allows researchers and policymakers to assess not only how but why renewal outcomes vary—analyses that could inform strategies to improve redetermination processes and minimize unnecessary coverage gaps and administrative costs. However, there are several ways that CMS could amend the reporting obligations to offer more detailed and actionable insights.

Some demographic groups—such as Hispanic children and children with parents who do not speak English—are more likely to lose coverage due to administrative burdens.

At present, CMS only requires that states report aggregate metrics that combine Medicaid and CHIP program outcomes, irrespective of demographic or eligibility group (eg, children, persons with disabilities). However, some demographic groups—such as Hispanic children and children with parents who do not speak English—are more likely to lose coverage due to administrative burdens.²³ Moreover, different Medicaid and CHIP eligibility groups may be disproportionately affected by renewal processes. CMS should require key metrics reporting, especially for specific groups subject to different eligibility rules. At a minimum, policymakers would benefit from results stratifying enrollees into the following groups: children (in Medicaid or CHIP), non-elderly adults who qualify for Medicaid on the basis of income, and adults who qualify for the program on the basis of age or disability. Doing so would offer more useful insights into enrollment dynamics following eligibility redeterminations.

Children are a particularly salient eligibility group of interest, as they historically have had some of the worst rates of administrative churn—that is, disenrollment while remaining eligible—but, under the CAA, receive 12 months of continuous eligibility.³ Currently, 16 states conduct interim data checks (in addition to annual redeterminations) by accessing available databases to evaluate whether enrollees’ eligibility status may have changed before their renewal date; children, under the new continuous eligibility rules, are exempt from such checks.²⁴ Unlike annual eligibility redeterminations, wherein enrollees typically have at least 60 days to submit required paperwork, states can give enrollees as few as 10 days to respond to requests for information substantiating their eligibility if the data check suggests their status may have changed.²⁴ Under regulations finalized in 2024, states will be required to offer at least 30 days for responses starting in June 2027,²⁵ although rescission of these regulations is under consideration by Congress; the new presidential administration could also undo these regulations in the absence of legislative action. In addition, observing continuous 12-month eligibility for children in Medicaid and CHIP, but not for adults in the same household who might simultaneously qualify for Medicaid benefits, could generate confusion about enrollment requirements and coverage status, as evidenced by the so-called “undercount.” Specifically, the gap between the number of people reporting Medicaid coverage in federal surveys and the greater number enrolled per administrative data worsened during unwinding, particularly for children,^26,27 suggesting that parents may be unaware of children’s continuous eligibility. Additional research is needed on the best strategies to publicly communicate ongoing Medicaid enrollment for child enrollees.

Enrollees who qualify for Medicaid on the basis of old age or disability status are another eligibility group of interest, as they must meet asset limit requirements in addition to having income below a specified threshold. The resulting administrative burdens—verifying that these enrollees continue to meet the asset limit at each annual renewal—can be a particularly onerous process. Given the nature of their eligibility, these enrollees are likely to have intensive health care needs and to be particularly at risk for disruption to care in the event of lapsed coverage.

Existing measures of redetermination outcomes are also coarser than they could be. The “procedural terminations” metric, for example, captures a wide variety of potential reasons for termination; as an aggregate measure, it offers little information about which policy interventions could best improve outcomes. It would be helpful to know, for example, the number of enrollees who experience a procedural termination after failing to be renewed ex parte due to an “over income” determination vs the number for whom the system was unable to determine income (or some other key eligibility criteria). States should also track and report the number of procedurally terminated enrollees for whom they received returned mail.

Lastly, CMS could consider requiring that states report data on Medicaid reentry within 6 or 12 months—which previous research has shown is common, especially for children^28,29—similar to how HealthCare.gov enrollments are currently reported.³⁰ Such a policy would be particularly relevant for child enrollees, who are likely to remain eligible for Medicaid or CHIP even if their parent earns above the adult income threshold.

Conclusion

The Medicaid unwinding from April 1, 2023, through May 2024³¹ was singular in some ways. National Medicaid enrollment swelled by over 30% from February 2020 to January 2023 while continuous coverage was enforced, meaning states were undertaking their largest eligibility redetermination efforts in history.³² Larger caseloads created capacity challenges, which some states may have exacerbated with policy choices unique to unwinding (eg, front-loading “likely ineligible” cases or condensing the process over 6 months).^33,34,35 There were also concerns that, after 3 years without redeterminations, enrollees would be more likely to have outdated contact information on file with the state, in consequence of which they might never receive notices and required paperwork.³⁶

Yet, in other ways, the Medicaid unwinding was utterly ordinary. States have always been required to conduct Medicaid eligibility redeterminations on an annual basis (and, in some cases, opted to do so more frequently). Researchers have long suspected that these processes, which can be onerous, impose administrative burdens that screen eligible individuals out of coverage. However, data limitations have made these dynamics—and their consequences—difficult to study. The new reporting requirements, intended to provide transparency and support federal oversight during unwinding, offered unprecedented visibility into Medicaid renewal outcomes.

These reporting practices may change with shifting political tides. Some members of Congress have stipulated that they might increase eligibility redetermination cadence as part of a suite of policies to fund an extension of expiring tax cuts; revenue would come from reduced program enrollment.³⁷ It seems plausible that any single administration would not want to publicize the volume of enrollees—especially children—losing Medicaid eligibility on a monthly basis under these circumstances. The unwinding of the Medicaid continuous enrollment provision has led to a number of policy lessons—and the importance of publicly available redetermination data is a key one. Suppression of these new data would be a loss for policymakers and the Medicaid enrollees they serve.