The LIS Landscape

The policy conversation about low-income Medicare beneficiaries almost always defaults to dual eligibles. That population is important, heavily studied, and reasonably well-served by an infrastructure of D-SNPs, FIDE SNPs, and state integration contracts designed to wrap services around their needs. But there is a population that is larger, less studied, and far less well-served by existing policy infrastructure: the low-income Medicare beneficiaries who receive Extra Help for Part D, or who qualify for Medicare Savings Programs, but who are not full dual eligibles. These are more than 13 million Americans navigating Medicare costs without the full protection of Medicaid, often unaware of the programs that exist to reduce their burden.

The programs themselves are not obscure. The Low Income Subsidy provides premium, deductible, and copay assistance for Part D prescription drug coverage. Medicare Savings Programs cover Part A and B premiums, deductibles, and cost-sharing at various levels depending on income. Together, they represent the most significant cost-reduction infrastructure available to low-income Medicare beneficiaries outside of Medicaid. The problem is not that these programs do not exist. The problem is that millions of eligible beneficiaries are not enrolled, the enrollment processes are fragmented across federal and state agencies, and the policy conversation treats these populations as an afterthought relative to the dual eligible cohort that attracts the most integration investment.

The LIS/Extra Help Program

#

The Low Income Subsidy, administered jointly by the Social Security Administration and CMS, provides Part D prescription drug cost assistance to beneficiaries with limited income and resources. The program operates at two levels. Full LIS covers the Part D premium (up to the regional benchmark), eliminates the deductible, and caps copays at nominal amounts that in 2026 range from $0 for institutionalized beneficiaries to $5.10 for generics and $12.65 for brand-name drugs depending on income category. Partial LIS reduces the premium, applies a sliding-scale deductible, and caps cost-sharing at 15 percent of the drug cost. SSA estimates the average annual value of Extra Help at approximately $5,700 per person.

Full LIS is available to beneficiaries with incomes up to 135 percent of the federal poverty level and limited resources. Partial LIS extends to 150 percent of FPL. In 2026, resource limits for full LIS are $16,590 for individuals and $33,100 for couples, indexed annually by the Consumer Price Index. Auto-enrollment covers the populations for whom eligibility is already established in administrative data: full dual eligibles, SSI recipients, and beneficiaries enrolled in a Medicare Savings Program. Everyone else must apply through SSA, and the application gap is substantial. Approximately two million beneficiaries who qualify for LIS are not enrolled, according to NCOA estimates drawing on CMS enrollment data and Census income tabulations. The eligible-but-not-enrolled population skews older, female, widowed, and concentrated in states without Medicaid expansion where the income band between Medicaid eligibility and LIS eligibility is wider.

The Part D interaction is operationally significant. LIS beneficiaries who do not actively choose a Part D plan are auto-assigned to a benchmark plan in their region. The benchmark is the weighted average premium for basic Part D coverage, and it varies by CMS region. In 2026, regional benchmarks range from approximately $0 to over $40 per month. A beneficiary enrolled in a plan priced above the benchmark pays the difference out of pocket, a cost that can accumulate to over $100 annually and that many LIS beneficiaries do not anticipate. Auto-assignment does not optimize for formulary fit. A beneficiary taking three medications may be assigned to a benchmark plan that covers two of them, requiring either a formulary exception, an out-of-pocket payment, or a medication change that no one discussed with them.

The LINET program provides temporary Part D coverage for up to two months to low-income beneficiaries who qualify for Extra Help or Medicaid but have not yet enrolled in a Part D plan. Administered by Humana under contract with CMS, LINET fills a coverage gap that would otherwise leave newly eligible beneficiaries without drug access during the enrollment processing period. The program is operationally important but not widely known among the beneficiaries it serves.

Medicare Savings Programs

#

The four MSP tiers provide different levels of Medicare cost-sharing assistance, each with its own eligibility rules, enrollment mechanism, and administrative apparatus.

The Qualified Medicare Beneficiary program is the most comprehensive. QMB covers Part A and B premiums, deductibles, copays, and coinsurance for beneficiaries with incomes at or below 100 percent of FPL and resources below $9,660 for individuals in 2025. More than a dozen states have eliminated asset tests entirely, including New York, Connecticut, Arizona, Oregon, and Louisiana. As of 2023, more than 8 million people were enrolled in QMB. But CMS data from 2022 showed that more than half a million eligible beneficiaries were not enrolled, and independent estimates from MACPAC suggest the actual participation rate for QMB-eligible adults is approximately 53 percent. Nearly half the people who qualify for the most protective low-income Medicare program are not receiving its benefits.

QMB carries an additional protection that is important and routinely violated: providers cannot charge QMB beneficiaries cost-sharing amounts. Federal law prohibits balance billing for QMB enrollees. Yet compliance is uneven, and many beneficiaries do not know they have this right. The billing violation problem is compounded by the fact that QMB status is not always visible to providers in their billing systems, creating a documentation gap that produces erroneous charges.

The Specified Low-Income Medicare Beneficiary program covers Part B premiums only, for beneficiaries with incomes between 100 and 120 percent of FPL. The Qualifying Individual program also covers the Part B premium but operates under a capped federal appropriation, meaning enrollment is effectively first-come, first-served within each state’s allocation. MACPAC data shows SLMB participation at roughly 32 percent of eligibles and QI participation at approximately 15 percent. The Qualified Disabled Working Individual program, the least common tier, covers the Part A premium for disabled beneficiaries who have returned to work and lost premium-free Part A eligibility.

A CMS rule finalized in 2023 and effective October 1, 2024 required state Medicaid agencies in 36 states and the District of Columbia to auto-enroll SSI recipients into QMB when they meet the eligibility criteria and are enrolled in premium-free Medicare Part A. This was the most significant structural change to MSP enrollment in years. But the One Big Beautiful Bill enacted in July 2025 placed a moratorium on several provisions of the CMS streamlining rules, creating uncertainty about whether the auto-enrollment mandate will continue to be enforced in all states. The moratorium does not prohibit states from implementing auto-enrollment voluntarily, and advocates are pressing states to maintain the policy regardless of the federal moratorium. Whether they do will determine whether the QMB enrollment gap narrows or widens.

The LIS Benefit Cliff

#

The transition from full LIS to partial LIS to no LIS creates a benefit cliff that has measurable consequences for medication adherence. A beneficiary receiving full LIS who crosses the income threshold loses premium protection, faces a deductible, and sees copays rise from nominal amounts to 15 percent of drug costs. For a beneficiary taking a branded specialty medication, that shift can mean hundreds of dollars per year in new out-of-pocket costs appearing without warning when SSA redetermines eligibility.

The interaction between LIS and other means-tested programs compounds the problem. SNAP categorical eligibility creates a pathway that links food assistance and drug coverage for some populations, but the connection is not automatic and depends on state-level implementation. A beneficiary who loses SNAP eligibility may also lose the categorical pathway to LIS, even if their income has not materially changed.

Marriage creates an additional penalty. When two LIS-eligible beneficiaries marry, their combined income and resources are assessed against couple thresholds that are not double the individual limits. A marriage that would have no effect on either person’s Medicaid eligibility can push both above the LIS resource threshold, eliminating drug coverage assistance for both. This is an underexamined policy design problem that affects elderly couples making end-of-life financial and personal decisions.

The BALANCE model, if it proceeds to implementation, raises a specific LIS question. If GLP-1 drugs become covered under Part D through BALANCE’s anti-obesity medication coverage framework, the cost-sharing structure for LIS beneficiaries will depend on whether those drugs appear on benchmark plan formularies. Full LIS auto-enrollees are assigned to benchmark plans. If GLP-1 coverage is available only through non-benchmark plans, LIS beneficiaries would need to actively switch plans to access the benefit, a step the auto-enrollment design does not facilitate.

What Plans, Advocates, and Programs Can Do

#

The enrollment gap is not a mystery, and the interventions that reduce it are well documented. Automatic enrollment works. The states that implemented data-sharing to auto-enroll SSI recipients into QMB before the federal mandate saw enrollment rates rise measurably. California’s DHCS began auto-enrolling all SSI/SSP members into QMB with assignment of the QMB aid code effective January 2025, removing the application burden entirely for that population. The model is replicable.

Outreach also works when it is funded and sustained. NCOA’s Center for Benefits Access and the BenefitsCheckUp tool have demonstrated that proactive screening at community sites, tax preparation events, and social services offices can identify eligible beneficiaries who would not otherwise apply. State Health Insurance Assistance Programs operate in every state with counselors trained to navigate MSP and LIS applications, but SHIP funding has been flat or declining in real terms for years, and the counselor workforce has not kept pace with the growing Medicare population.

MA plans and Part D sponsors have specific obligations to LIS beneficiaries that create compliance requirements around formulary access, premium notification, and auto-assignment. The compliance gaps are real. Plans are required to send notice when a LIS beneficiary’s premium will exceed the benchmark, but the notices are often confusing, arrive late relative to enrollment deadlines, and do not explain alternatives in language that beneficiaries can act on.

The broader infrastructure question is whether proactive LIS and MSP eligibility identification can be built into the enrollment systems that beneficiaries already touch. Social Security field offices, Medicare.gov, and the 1-800-MEDICARE call center all represent interaction points where screening could happen but generally does not. AI-assisted navigation tools that cross-reference income data with program eligibility represent a potential pathway, but that pathway depends on data access and system integration that does not currently exist at scale.

The LIS and MSP populations are not invisible because they are small. They are invisible because the policy infrastructure treats them as residual categories within a dual-eligible-centric analytical frame. That frame misses the 13 million beneficiaries for whom these programs are the primary source of cost protection, and it understates the enrollment gap that leaves millions more without the assistance they qualify for.

Related Reading

#

MCR-09_05 Medicare Savings Programs: The Invisible Benefit Cliff MCR-01_05 BALANCE: The GLP-1 Gambit MCR-06_14 The Human Advocacy Layer: ADRCs, SHIP, AAAs, and the Benefits Enrollment Ecosystem