Progressing medicaid enrollment and eligibility
Erica Olenski was with her two-year old son, who was receiving treatment for brain cancer, when a nurse gave her some news. His Medicaid coverage had just been terminated. There had been no prior warning—no letter or notice. Without coverage, her son’s treatment would have to stop. She tried calling her state’s Medicaid office, but no one answered.
That’s just one story of nearly half a million other children who faced similar disruptions to their Medicaid during last year’s Medicaid unwinding. And it wasn’t just kids–7.5 million people lost coverage, despite still qualifying. These terminations were “procedural”, which is bureaucratise for such issues as paperwork errors, missed deadlines, or outdated contact information. Entirely preventable mistakes.
As health technologists, we spend most of our time improving the systems that deliver or pay for care. But that focus has a blind spot: we rarely inspect what it takes to get and keep access to care in the first place.
I spoke with a Medicaid enrollment officer who shared a nightmare scenario with me. Picture Maria, a part-time receptionist with chronic back pain, applying for Medicaid online. She checks the disability box, and unknowingly triggers a requirement to apply for another benefit: Social Security Disability. Her condition isn’t on the approved list, so she has to go through months of clinical evaluations. Meanwhile, since her Medicaid application is in limbo, she runs into another problem: Medicaid’s asset limits mean she has to either spend down her savings or get an attorney to restructure her finances.
Maria is one of millions lost in Medicaid’s bureaucratic labyrinth. But it’s nothing new. Sara Rosenbaum— founding commissioner of MACPAC and chair of George Washington’s Milken Institute School of Public Health –said “for over half a century, states have struggled with the Medicaid eligibility process, one driven by voluminous federal regulations and subject to countless lawsuits challenging unlawful denial or termination”.
We can’t fully diagnose the root causes here, but at least two factors drive this: the policy bloat in eligibility and administrative rules, and the way states build technology to check who qualifies. We’ll have to save this for another post. These are chronic, systemic issues decades in the making.
These longstanding and systemic flaws in Medicaid will need state-level reforms. In the meantime, however, there are simpler solutions, immune from the vagaries of federal regulations, that can bring us closer to a better experience for applicants and the eligible uninsured.
Take for instance Fortuna Health, a VC-backed startup founded in 2023, who are building a portal to guide applicants through enrollment. Think about how TurboTax transformed how you file your taxes. Fortuna is doing exactly this for folks seeking Medicaid–in fact, they’ve coined themselves “TurboTax for Medicaid”.
Fortuna was conceived to address the extraordinarily bewildering process of applying for Medicaid. For some time, people have suggested that state web portals could be more user-friendly. Income rules are hard to understand and inevitably involve guesswork – which in turn may lead to ‘procedural’ disqualification. Communications from caseworkers, such as requests for more documentation or explanations for coverage denial, are as often as not unclear. To alleviate this, Fortuna’s portal not only simplifies the initial application submission, but also guides applicants through everything that follows. It is in the followup, afterall, where most eligible Medicaid patients fall. As one enrollment officer told me, it’s simply not enough to point people to the application and encourage them to apply, you must also follow-up on their behalf. This is exactly Fortuna’s goal.
They’re doing this by centering their product on two things: an intuitive user experience that feels more Silicon-valley than DMV, and an automation engine that’s encoded hundreds of pages of mind-numbing Medicaid eligibility legalese. Now, when Maria is unsure how to calculate her income because her pay stubs have been irregular and she’s unsure whether child support counts as income, Fortuna’s intelligent chatbot walks her to an accurate estimate. Or when she’s denied because her last tax status doesn’t reflect her recent divorce, Fortuna auto-generates an appeal letter with her divorce papers and citations about household composition criteria.
But Fortuna isn’t just a consumer-facing portal. They’ve integrated their services directly into healthcare workflows through partnerships with hospitals like Alomere Health and Essen Health, Medicaid managed care plans, and payment platforms like Cedar. For hospitals, this means their uninsured patients can be quickly screened and enrolled in Medicaid, often retroactively covering recent care. Plans benefit from more stable membership and reduced churn costs. Rather than simply directing applicants to state resources, Fortuna’s team and system handles the entire process - submitting and triaging applications directly with states and counties, coordinating with county case workers, and managing the back-and-forth that typically falls to overwhelmed hospital staff or confused applicants.
For the thousands who lose Medicaid coverage only because they missed their renewal deadline, Fortuna notifies them well ahead of time. But there is a rather basic difficulty we are overlooking. What happens when we can’t reach these people in the first place? We’ve presupposed a kind of stability that doesn’t reflect the realities of many in low-income populations. When someone’s moved three times in a year or exhausted their prepaid minutes or had to switch carriers entirely, they effectively become invisible to both their Medicaid managed care plans and state agencies. There is a role for technology to play here, as Sachin Jain notes: “technology may enable us to keep track of patients in ways, provided they consent to it, that allow us to stay close to patients who are otherwise very difficult to find”. A Medicaid Director I spoke with put this to practice. Her health plan partnered with credit reporting agencies like TransUnion to track down current phone numbers, and used services like MemberMatch to monitor ER visits. The idea was that when their members showed up at out-of-network emergency rooms, the discharge paperwork might contain new contact information. But the results were modest at best. This feels like a quiet reminder about the limits of tech-focused solutions. Engagement will always need human connection. Sachin Jain would surely agree. Some startups are rethinking how to engage these “hard-to-reach” members. Fabric Health, which provides health services in laundromats - where something like 32 million Americans, many from underserved communities, spend several hours every weekend. Ounce of care provides health services to residents of affordable housing communities. Reema Health focuses on trust, deploying Guides—local community members who meet people where they are, offering face-to-face meetings at their home, coffee-shops, libraries, and hospitals. This past year, these organizations prioritized Medicaid renewals, helping people navigate deadlines. To me, these models are a marked progression in delivering care to people where they are, and have the potential to impact vulnerable populations just as in-home care has transformed access for seniors and retail-based clinics have for working families seeking convenient primary care services.
When Medicaid coverage ends, families face tragedies, just like Erica and her son. These disruptions also take a less visible toll on the larger Medicaid ecosystem, especially as it moves towards value-based care. Providers find themselves in an impossible situation: how can they orchestrate long-term care plans for patients who come in and out of coverage? What makes this ironic is that Medicaid care models are beginning to address factors like housing instability, financial security, and health literacy - the very factors that increase the likelihood of coverage disruption in the first place.
Astrana Health is a leader in bringing VBC to Medicaid, and they’ve made it much easier for physicians to succeed with these contracts. If you are one of the 10K physicians in Astrana’s network, you have access to their technology platform that comes with predictive analytics and care coordination tools - all helping you provide timely care to your patients. You don’t have to deal with the headaches of administrative work like claims processing and revenue management either. If you’re practicing in California, you might be one of the physicians working in their Enhanced Care Management (ECM) program. Since ECM brings together social services and clinical care, you can use Astrana’s platform and services to work with social care workers and community based organizations to help your patients get the social support they need.
But despite the promise of ECM, California has only seen 1% of providers participate in the program. Part of the reason is because ECM’s success–just like any VBC program–depends on providers being able to build long-term care plans and stick with their patients over time. When patients cycle in and out of coverage, this becomes challenging.
Here’s what can happen: A patient loses coverage. They can no longer access Astrana’s carefully tailored care and social supports. Referrals to social services stop. Preventative services stop. Performance payments stop. Everything stops – except the conditions that brought the patient into the system in the first place.
Building the operations to deliver integrated care effectively requires significant resources and time. It takes 8-9 months just for physicians to be certified for the ECM. But providers are already operating on thin margins from low Medicaid reimbursement rates. They can’t justify this investment when their patients might disappear in a few months.
This creates a troubling cycle. Medicaid care models are becoming increasingly integrated with social services, aiming to address the very causes of instability that lead to coverage disruptions. But coverage instability prevents these models from taking root and succeeding at scale.