By Jim Hickman, with assistance from Julie Silas, Homebase

On Thursday, October 15, Kinda Serafi, Partner at Manatt, Phelps & Phillips, LLP, and Michael Freeman, Assistant Deputy Director, Health Care Benefits and Eligibility, California Department of Health Care Services, presented at the Putting Care at the Center town hall in Oakland, “Preparing for Federal Medicaid Eligibility Changes: California’s Implementation Plan.” Afterward, Julie Silas and I talked through the three ways to meet the work requirement (hours, income, or school) and compared how the $580 income requirement plays out across the country. Julie contributed handwritten notes that started the chart below.

80 hours × $7.25 per hour = $580

H.R. 1 sets the monthly Medicaid community engagement requirement at 80 hours a month, $580 of household income a month, or enrollment in school at least half time. The 80 hours can combine work (paid or unpaid), job training, community service, and other qualifying activities. For people who earn income, Congress converted those hours to dollars at the federal minimum wage, $7.25 an hour, unchanged since 2009. A Medicaid applicant or beneficiary can meet the work and community engagement requirement by earning at least $580 a month in household income. Other pathways and exemptions also apply, and for some people the income route will be the easier one.

The benchmark stays at $580 of household income everywhere. Depending on where a Medicaid beneficiary lives, they may reach it working far fewer than 80 hours a month.

The range runs from 34 hours in Washington to 80 hours in the 20 states that use the federal floor. A California worker at minimum wage reaches $580 of household income in about 35 hours a month. The same worker in Texas needs 80 hours of paid work.

The income test counts the household

The test uses household income under the Medicaid modified adjusted gross income method. One example from the town hall involved two household members earning $400 and $200. Together, their $600 clears the threshold.

The implementation question

Will eligibility systems recognize the income pathway automatically, or will eligible workers face extra paperwork?

Starting in 2028, the federal interim final rule requires states to use data they already hold before asking people for documents. The town hall discussion described California’s approach the same way, with data first, documentation second, and a signed attestation last. In 2027, attestation alone was described as earning six months of coverage, while a data match or documentation earns twelve.

A data match can double a person’s coverage period, so payroll and health information exchange connections matter. Plans, providers, and community organizations can ask their state Medicaid agency which data sources feed community engagement verification.

Learn more

Stay Covered Toolkit launch webinars. The Stay Covered Coalition, including Homebase, is hosting a two-part webinar series to launch the California Stay Covered Toolkit, a resource for organizations helping Californians experiencing homelessness keep their Medi-Cal. Webinar 1 is Thursday, November 12, from 9:00 to 10:30 a.m. Pacific. It introduces the toolkit and explains what is changing for people experiencing homelessness. Webinar 2 is Thursday, December 3, at 11:30 a.m. Pacific. It covers the work and community engagement requirements and medical frailty exemptions. Registration is on the Homebase website https://www.homebaseccc.org/stay-covered-coalition.

Onward,

Jim

Sources and contributors

• Town hall 509, “Preparing for Federal Medicaid Eligibility Changes: California’s Implementation Plan,” Putting Care at the Center, Oakland, October 15, 2026. Presenters: Kinda Serafi, Manatt, Phelps & Phillips, LLP, and Michael Freeman, California Department of Health Care Services.

• Julie Silas, Senior Directing Attorney, Homebase, handwritten notes on state minimum wages and hours, and the discussion of variation across the United States that followed the town hall.

• U.S. Department of Labor, Wage and Hour Division, State Minimum Wage Laws, updated July 1, 2026.

• Public Law 119-21, section 71119, and Centers for Medicare & Medicaid Services interim final rule CMS-2454-IFC, 91 Federal Register 33348.

I regularly type and dictate meeting notes and takeaways, then send them to myself as emails. You can imagine how many of those drafts I delete. A few survive, next to the latest streaming recommendations.

One I kept is titled “Favorite quotes from Putting Care at the Center 2017,” from the conference in Los Angeles, co-hosted by Kaiser Permanente. The first line is “process is the product.” The last is “eminently doable.”

I reopened it while looking through my drafts when I got home from this year’s conference, Putting Care at the Center 2026 in Oakland, where Kaiser Permanente was co-host again.

Hope, not need

Kathleen Noonan, CEO of the Camden Coalition, opened with the tagline for the year, “Uncertainty be damned.” The Camden Coalition published its strategic plan in the middle of the uncertainty. My 2017 list has a line for that: “focus on solution/hope not need/problem.”

Funding from Providing Access and Transforming Health (PATH), the state initiative that paid startup costs for community-based organizations, ends in December. Susan Philip, deputy director for Health Care Delivery Systems at the Department of Health Care Services (DHCS), said the federal government has been explicit that there will be no renewal. On January 1, members with unsatisfactory immigration status move from managed care into fee-for-service. The federal interim final rule on Medicaid work requirements is in effect.

Dr. Michelle Schneidermann told the room about a young woman with advanced heart failure who left the hospital against medical advice again and again, until a medical respite program earned her trust. The lesson: “Our expectations had to move at the speed of trust.”

What five years showed

My 2017 list says to make the lessons standard of care: first we have evidence, then we move to standard care. On Friday, Philip joined a five-year retrospective on Community Supports. DHCS reported in June that 12 of the 14 original Community Supports were cost-effective within the study period, and that personal care and homemaker services were associated with a 79.9 percent net cost reduction. Philip said that result struck her most. Helping people stay home may look like a no-brainer, she said, and the numbers still help.

The panel also named limits. A health plan representative in the audience said it is hard to isolate one support when members receive several at once. Katherine Barresi, chief health services officer at Partnership HealthPlan of California, said cuts to federal housing vouchers were not anticipated when these services were designed. Philip said cost-effectiveness is one measure and members’ own account of their outcomes is another, and the evaluation running through 2028 is meant to cover both.

Her biggest lesson was standardization. The program launched across many plans with different contracting systems, referral processes and authorization rules, so an organization working with three plans learned three systems. DHCS has since moved toward standardized referrals and tighter guidance. My 2017 note says “standardization & customizations,” and the question is still what the state fixes and what stays local.

Medical frailty

On Thursday, a town hall on California’s plan for the federal eligibility changes covered how the state will apply the medical frailty exclusion in the work requirement. The speakers were Michael Freeman, assistant deputy director for Health Care Benefits and Eligibility at DHCS, and Kinda Serafi, a partner at Manatt, Phelps & Phillips. I organized public events for the White House working group on welfare reform in 1993 and 1994, which came before the last national round of work requirements. Now they are in Medicaid.

Before the rule, a qualifying diagnosis was enough. Now the condition has to significantly impair the member’s ability to meet the requirement. The session described a tiered code framework patterned on Louisiana’s, which the Centers for Medicare & Medicaid Services pointed states toward. Some codes would qualify a member alone. Others would count only with corroboration, such as hospital records, medications, oxygen or durable medical equipment, and the corroboration does not have to share the diagnosis. A member with no data would need documentation from a clinician. The code list is still being refined and has not been published.

DHCS described attestation this way. In 2027 a member can attest to frailty and get six months of coverage, and documentation or a data match earns twelve. From 2028 attestation can be used once per enrollment period, the state must check data first, and a member who has attested must produce documentation at the next renewal. A member with no regular clinical relationship has the least to draw on at that point.

Audience discussion at the town hall surfaced six ideas for DHCS to consider. Publish the code list as a searchable tool. Put redetermination dates in the monthly eligibility file providers already pull. Use homeless services data systems to carry renewal dates and exemption flags. Let a community health worker (CHW) bill renewal help for a member already in Enhanced Care Management (ECM) in the same month. Widen who can sign off on frailty documentation. Require contracted CHW vendors to offer in-person service. One attendee described a vendor assigning a Sacramento member to a CHW based in Los Angeles who would not drive.

My 2017 list says “small scope/big impact.” A published list and a date in a file are small scope.

Connect what we digitized

Barresi said it on Friday: “We digitized healthcare. We didn’t connect it.” Jamie Almanza, CEO of Bay Area Community Services, said the organization runs five different homeless management information systems and five electronic health records across counties. Another health plan representative in the audience said ECM lead care managers usually cannot see Community Supports notes, and proposed a housing-specific hub built on the homeless management systems as a first step toward one view of a member. My 2017 note has a line for this too: “debating around the edges.”

On the second panel, Hannah Erickson of Impact Health, a network lead for 60 community-based organizations in North Carolina, said contracting with an organization differs from managing the health of a network. Impact Health draws service zones from the number of clients each organization needs per month to reach a target margin. The principle was to treat these organizations as health infrastructure and pay them that way.

Anand Shah of Kaiser Permanente said the organization weighs effectiveness and readiness when it chooses community partners. Readiness is easier to measure, so it becomes the proxy, and that favors larger organizations that are not necessarily more trusted. Shah named learning agility as the capability that matters most.

The tables

Almanza described what changed for clients once the program could help pay rent, find a landlord and buy a bed for someone moving in. My 2017 note says, “needs don’t change communities, dreams do.”

From the audience, I thanked Partnership HealthPlan and the other plans that partnered on the Collaborative Planning and Implementation initiative. The Camden Coalition set up tables and brought trusted voices across from plans and providers to untie the knot, and the tables produced more trust and knowledge transfer than anyone expected. I said we were shock absorbers and some communities will miss the tables after January 1. Some are already keeping them going on their own. A community organization leader from the North Bay said she could not follow the language when she first sat down, and that once she could, she built her team around it.

The strongest objection is that this asks stretched organizations to build billing paths, data links and local tables in the same quarter their startup money ends. That is fair. Erickson warned against going all in too fast and described 60 organizations in the network that had to turn off services within 29 days when funding ended unexpectedly. Massachusetts shows the same risk. Its budget cut care management funding by 50 percent, and the state’s 20 Community Partners programs, serving about 35,000 people, were slated to lose funding by June 30, according to GBH News. The state said it will still require health plans to provide care management. Some of this will not be built by January 1.

DHCS encouraged philanthropy to sustain the tables and connections PATH once paid for. Some funders have answered. Yet a table keeps meeting only when someone owns the convening, and in some communities that owner is still unnamed.

The last line of my 2017 note is “eminently doable.” A published code list is doable. So is a renewal date in the file providers already pull, and so is a local table that keeps meeting after PATH ends. In July I filed a written comment with the Future of Medi-Cal Commission arguing that coordination belongs in the base payment. After five years of Enhanced Care Management and Community Supports, we have enough experience to start making it visible in reporting, funded in rates, and counted as care when it is measured.

Onward,

Jim

Disclosure: I am a senior advisor to the Camden Coalition and sit on the Putting Care at the Center planning committee, and Hickman Strategies exhibited at the 2026 conference.

I spent Tuesday and Wednesday in a ballroom in New Orleans listening to state Medicaid leaders describe a problem I had been writing about in California two weeks earlier. Nobody in the room connected the two, and by the second session I could not get it out of my head.

At the Health Management Associates national conference this week, Scott Partika, director of the Ohio Department of Medicaid, described 1,400 Medicaid-enrolled infants who received zero Medicaid claims after birth. No claims data, no wage data, nothing. They surfaced because a children’s services partner asked for a data pull, not because anyone’s claims analysis flagged them.

Two weeks earlier, the California Department of Health Care Services told its stakeholder advisory committees that 649,405 Medi-Cal members, about a third of the 1.93 million moving from managed care to fee-for-service on January 1, cannot be assessed at all. They have no claims history. Chris Esguerra, the department’s deputy director and chief quality and medical officer, put it carefully and appropriately. This is an absence of data, not necessarily absence of need, and the state does not know what the need is.

Different states, different populations, same finding. The surveillance apparatus that runs American Medicaid sees people through the claims they generate, and the people at greatest risk generate the fewest. The panel in Ohio’s session said the quiet part out loud. Extremely low utilizers are still likely in critical need, and they will not show up in standard claims surveillance.

What it costs when nobody is looking

A physician on that panel described a patient, a 38-year-old man, fully Medicaid-eligible, who lost coverage through an administrative failure. He took three inhalers for severe asthma and lost all three prescriptions. He ended up in the emergency department and was admitted for three days. The hospital absorbed the cost. It took him three months to reconnect with primary care, and the physician estimated six to twelve months to recover fully. Routine primary care would have prevented the entire episode.

That arithmetic is already running. Aaron Todd, chief executive of the Iowa Primary Care Association, reported community health centers in his state seeing a 5 percent increase in uninsured patients in a single year, driven by coverage losses that have nothing to do with the cuts in H.R. 1, which have not taken effect yet. Service lines are closing and sites are shutting down, some in anticipation of a law that has not reached them. One panelist described the safety net as held up by a wet match.

The third group nobody counts

There is a category below the low utilizers, and it came up in the same session. Immigration fear is keeping roughly 30 percent of some patient populations from seeking care at all. Those people hold coverage. They generate no claims by choice, because the cost of being seen feels higher than the cost of staying sick.

California is about to run a live experiment on exactly that population. The members moving to fee-for-service on January 1 are the ones with unsatisfactory immigration status, and the federal public charge rule changed in September. The state will mail a notice in December telling nearly two million people that the way they get care is changing. Some of them will read it as a reason to stop getting care.

What actually finds people

Nothing in either story is solved by better data exchange. Exchange moves records that exist, and for a third of this population there are none.

What worked, in the one example anyone offered, was people. When a state had to remove providers quickly for program integrity reasons, it deployed community-embedded outreach workers to contact affected members directly, using utilization patterns to find who was most exposed. The care management infrastructure proved its value during the disruption, not in steady state.

The operational versions are also in this week’s sessions, and they are cheaper than they sound. Community health centers in one state use an administrative services organization to reach community health workers and eligibility workers without adding them to their own payroll. Michael Owens, chief medical officer at Health Care Los Angeles, described more than thirty health centers running five different electronic health record systems that unified patient visibility across all of them, and now negotiate with payers before penalties arrive instead of after. Aaron Todd’s association put nearly every center in Iowa on one record system and can finally see quality at the population level, which no single center could do alone.

None of that is a technology story. It is a story about organizations that decided to share a back office so the front line could keep its people.

What I would ask for

Three things, and all of them were said this week by someone with the authority to do them.

Fund the analytics floor. A state cannot run this with two people in data analytics. Administrative cost floors need to go up, not down, which is a hard argument in a cutting year and the right one.

Design the recovery, not just the prevention. Re-enrollment is possible the day after a disenrollment. Almost nobody tells members that, and almost no state has a pathway that puts someone back on coverage and back on their medication within a week of a gap being found.

Pay the people who can see what claims cannot. In California that means the community health worker codes that survive the January transition, billed through an enrolled supervising provider, which can be a community-based organization or a county. The money is small and the pathway is real.

Dr. Michael Owens of Health Care Los Angeles put it this way from the stage: the unimaginable is the result of a dearth of imagination. The people we cannot see in the data are not unimaginable. We have never paid anyone to go find them.

Onward,

Jim

Related: No. 22, The Largest Group Is the One No Claim Can See and the DHCS fee-for-service transition page.

I spent last Wednesday in a room of Central Coast providers hearing what happens to their Enhanced Care Management caseloads on January 1, and the question underneath every hand that went up was the same one. Who is going to find these families in January?

Six days later, the Department of Health Care Services (DHCS) convened a special joint session of its two stakeholder advisory committees on that transition. Chris Esguerra, its deputy director and chief quality and medical officer, described roughly 650,000 Medi-Cal members the department has no recent claims data for. “This is an absence of data, not necessarily absence of need,” he said. “We just don’t know what the need is.”

His point was that nobody should read those members as low-need because the state cannot see their recent utilization. Esguerra said the January design needs a way to find these members and to respond when one of them, or a provider, asks for help. That is the sentence to hold onto, because a request-driven channel puts the first move with whoever a family already calls.

The deck DHCS presented is the most complete public picture yet of who is moving and what is ready. One number in it should shape how community organizations spend the hundred days left. The day brings together members nobody can find in the data and providers who know those members and cannot yet bill for them.

Who is moving

The driver is federal. A CMS letter to state Medicaid directors holds that federal funding for members with unsatisfactory immigration status (UIS) is available only for emergency services actually furnished, and a monthly capitation payment to a health plan is not a furnished service. Moving these members to fee-for-service, where the state pays per service, is how California keeps the federal match. The scale of that move is in the policy guide: Medi-Cal fee-for-service held 711,519 members in March, and DHCS expects about 2.7 million on January 1.

DHCS counts about 1.93 million members with UIS as of April, roughly 251,000 under 21, with every figure marked preliminary. The group is wider than undocumented adults. It includes children and pregnant people, green card holders here less than five years, people residing under color of law while an asylum claim is pending, and qualified noncitizens losing federal full-scope eligibility under H.R. 1. DHCS is explicit that these members keep their coverage. How they reach care changes on January 1, and Enhanced Care Management (ECM) and Community Supports, which exist only in managed care, end for them on that date.

DHCS has also sorted the population by transition risk. About 16,000 members need their care moved before January, roughly 279,000 have an active problem that needs management, and about 530,000 need no action, because every service they use is already with an active fee-for-service provider. The room pushed back on that sort. The Local Health Plans of California noted that more than 700,000 of these members had a specialty visit in the past year, 16,000 are on dialysis and 70,000 are receiving behavioral health services, and argued that a high-urgency count of 16,300 undercounts the real risk. DHCS agreed to publish written definitions of its risk categories.

The group no claim can see

The largest group on the slide sits outside all three. DHCS lists 649,405 members, about a third of the population, as unassessed because they have no claims history, and says it cannot tell whether they are healthy or disconnected from care. A share of any insured population goes a year without a claim, so the number on its own proves little. Esguerra’s caution about it is the part that matters.

Finding them takes people who already know them: a promotora who knows the family, a clinic front desk that remembers the name, a community health worker (CHW) already in the home. DHCS’s own plan points that way. The 2026-27 Budget Act includes $8 million for culturally and linguistically appropriate navigation by clinics and community-based organizations, through sub-awards from the California Primary Care Association, targeted to start January 1. It adds $31 million for care coordination, including a statewide vendor for the highest-risk members that takes referrals from call centers, the Nurse Advice Line, community organizations and plans. Spread across the state, $8 million funds a thin layer, and it goes furthest in organizations that already hold those relationships and are ready to bill for the work that follows.

What carries over

ECM stops because it is a managed care benefit. Much of the work underneath it has its own fee-for-service code. DHCS has told plans and providers that care management, case management, care coordination and CHW services stay reimbursable in fee-for-service when an eligible, enrolled provider bills them under Medi-Cal coverage policy. DHCS says a resource document with the code list is still to come. The codes circulated through the plans so far are below, and the community health worker provider manual carries the billing rules for the CHW codes. CPT is Current Procedural Terminology, and HCPCS is the Healthcare Common Procedure Coding System.

The supports arrive with the deadline

The timing problem sits inside DHCS’s own plan. A provider not enrolled in fee-for-service by January 1 cannot bill Medi-Cal, and DHCS said there is no expedited pathway, because program integrity requirements do not bend for a deadline. The care management vendor is still in contracting and is expected to be running in late December or early January, and the navigation money launches in January. Network adequacy standards for fee-for-service specialty care are a 2027 build, and the monitoring program is still in planning. Every support arrives on or after the day the deadline falls, which leaves the gap to the people already in the room with these families.

Owen set that against the program’s own history. The last time roughly 20 percent of Medi-Cal members sat in fee-for-service was 2016. “On January 1,” she said, “we sort of moved backwards overnight.” Her ask of DHCS was to measure actual access, not enrollment in a billing system alone, and to publicly monitor specialty access, appointment availability, geographic and language access, and disruptions in care. “This magnitude of change really demands tremendous humility and vigilance and transparency from us all,” she said.

A protection that covers other Medi-Cal transitions does not reach this one. DHCS states in its policy guide that the continuity of care requirements in Health and Safety Code section 1373.96, which make a plan finish a course of treatment with a departing provider, do not apply to members moving into fee-for-service. What replaces the statute is provider enrollment and the handoffs people make by hand.

Payment is the second pressure. Plans pay their providers more than the fee-for-service schedule, which stakeholders raised as a reason some may not follow their patients across. The National Health Law Program and the Local Health Plans of California proposed the same fix: enroll every managed care provider in fee-for-service automatically. That does nothing for January, and it would retire this problem before the next transition.

Before January 1

Enrollment is the gate. Of roughly 151,000 managed care providers serving these members, DHCS counts about 86,000 active in fee-for-service, 45,000 enrolled but not active, and 20,000 not enrolled at all. Check your status in the Provider Application and Validation for Enrollment (PAVE) system this week. An organization not already enrolled has to apply and be approved by January 1 to be paid at fee-for-service rates.

Marina Owen, chief executive of CenCal Health and chair of the Local Health Plans of California, named the limit of that number at the convening. “Enrollment in the PAVE system, the billing software, is not the same as participation,” she said. “And participation is not the same as capacity.” Meaningful capacity, in her description, means the right provider in the right geography and specialty, with the right language capabilities, who is actually willing to see the member.

Ask your plan partners which of your ECM and Community Supports members are moving. Plans must keep care coordination running through December 31 and help members learn whether their providers will continue in fee-for-service.

Match each service you deliver to a code, confirm the ordering and supervising requirements in the Medi-Cal provider manual, and name the person who will submit your first claim in January.

Know when the money starts. CHW services cannot be billed for a member who is enrolled in ECM, and plans pay ECM through December 31, so the CHW line opens in January, in the same month the ECM line closes. Build the workflow this fall and bill from the first week of the year.

Work the authorization calendar. DHCS accepts requests for January dates of service from October 1 and will match them against the authorization data plans have already sent. It asks that October to December referrals go to providers already enrolled in fee-for-service, and plans must still authorize any ECM or Community Supports referral received through December 31.

Get ahead of the notice. DHCS mails one member notice by December 1, to whatever address the county has on file, and runs a text campaign in every threshold language. Remind members to update that address and to refill prescriptions before January 1.

Keep a line on public charge. The federal rule that took effect September 18 reaches some households here and not others, the fear travels further than the rule, and individual questions belong with immigration legal services.

Who holds the enrollment

The policy guide holds a detail that reverses the obvious answer. Claims for CHW services must come from an enrolled supervising provider, and DHCS defines that list: a licensed provider, a hospital, an outpatient clinic under 42 CFR 440.90 including a tribal federally qualified health center, a pharmacy, a community-based organization, or a local health jurisdiction. Non-tribal federally qualified health centers and rural health clinics cannot serve as supervising providers in fee-for-service, and CHW services are not billable for them.

So a community-based organization can enroll and hold this itself, which is worth reading twice in a sector used to billing through a clinic. A county health department can hold it, and so can a hospital or a clinic under that federal definition. The health centers DHCS is asking to absorb more patients cannot bill these codes. A small organization can still work through a county or a hub, and that decision belongs in October, because whoever holds it needs approval by January 1.

I am recommending all of this, and I do not know whether it happens at scale. An organization can enroll and still find that claims, denials and audit-ready documentation cost more than a thin margin carries, and a hub or a county has to volunteer to hold enrollment and supervision for other people’s staff in the same quarter its own revenue changes. We will not know until the system is tested. What would move me is the January claims run: who enrolled by January 1, and whose first claims paid.

The relationship has to last past January. Six-month renewals start with the March 2027 eligibility month, and July 2027 brings the end of adult dental coverage and premiums for members 19 through 59. Whoever finds a family this fall is who that family calls in March.

The strongest objection is that this asks stretched organizations to build a billing path for a population that may stop coming through the door. That objection is fair, and it is also the case for continuity. A family that leaves managed care, steps back from public programs and loses its care coordinator in the same season turns up in an emergency department with nothing arranged. The organizations that already hold its trust are best placed to keep it.

This week

DHCS took written public comment only on the September 22 session, at SACinquiries@dhcs.ca.gov, and every comment lands in the meeting summary. Its fee-for-service transition webinar runs September 30 from 10 to 11 a.m., and demand was strong enough that the first registration cap filled and DHCS expanded it. A California Children’s Services Advisory Group meeting on October 16 takes up children’s coverage, and the committees meet again October 28. If your organization serves UIS members through ECM or Community Supports today, tell your plan partner this week what you intend to do on January 2.

One more thing, from someone who has watched a version of this before. Expect 2027 to bring back a vocabulary this field spent a decade trying to retire: the Emergency Medical Treatment and Labor Act (EMTLA), frequent utilizers, charity care, uncompensated care. What is old is new. This time we have the experience, and better tools and practices.

Onward,

Jim

Related: No. 21, The CalAIM Budget Question

Jim Hickman is Founder & Chief Catalyzer of Hickman Strategies LLC. He serves as a Senior Advisor and CalAIM subject matter expert with the Camden Coalition, as a Senior Advisor to PCCI, and as a Venture Partner at Suncoast Ventures. CalAIM Field Notes is independently written and edited.

On August 18 we spent an hour with 211 people on the California Improvement Network’s webinar, most of them running community organizations inside CalAIM. Eight questions came in through the chat and the survey that the hour could not reach. One of them was the plainest question anybody asked all day: what is the reduction in CalAIM funding in the California budget?

It took going back to the enacted budget to answer, and the answer is that the question has three.

We are budgeting against the wrong number

Talk to a community organization this month and you will hear a version of the same sentence. CalAIM is being cut, so we are planning for less. It is a reasonable thing to believe. It is also, on the figures, the smallest part of what the enacted budget did.

Three separate reductions land on the same members starting January 1, 2027, and they are not the same kind of thing. Reading them as one number produces a plan built against the wrong constraint.

The three reductions, on one basis

A caution before the figures, because it is what makes this comparison hard to do correctly. The budget states some of these reductions in General Fund dollars and others in total funds, and the two are not comparable. Everything below is General Fund, so the numbers sit on one basis and the gaps are real gaps.

CalAIM benefit refinement. Enhanced Care Management $41.4 million and Community Supports $26.9 million, so $68.3 million in 2026-27. About $150 million ongoing. It narrows the doors. Every service survives.

Members with unsatisfactory immigration status to fee for service. $524.9 million in 2026-27, $1.3 billion ongoing. It moves roughly two million people out of the delivery system that carries the benefits.

Work and community engagement requirement. $90.3 million in 2026-27, $2.4 billion by 2029-30. It attaches a functional-capacity test underneath an income test.

Source: 2026-27 Budget Act, Department of Health Care Services Highlights, July 1, 2026. May Revision figures are still circulating in decks around the state and they do not match this document. The combined $150 million ongoing figure is mine, not the state's: the Budget states the two ongoing savings separately and I have added them, which is a reasonable inference and not a published number.

The gap is in the out-years, not in this one

Look at the 2026-27 column and the story is not what anyone expects. The CalAIM refinement is $68.3 million. The work requirement, the change that has absorbed most of the attention in every room I have been in this year, is $90.3 million. In the year everyone is budgeting for right now, those two are roughly the same size.

Then look at the second column. The refinement settles near $150 million ongoing. The work requirement reaches $2.4 billion by 2029-30. Two changes that look comparable this year separate by roughly sixteen times inside four years, on my addition of the two CalAIM lines.

An organization that reads the first column and concludes these are one problem of a similar size will be right for about twelve months and then badly wrong. The three do not scale together, and only one of them is a benefit cut at all.

The second is the one nobody is planning for, and it is the largest thing happening next year. Moving two million members to fee for service takes nothing out of the CalAIM benefit set. Every service definition survives, and the 2026 trailer bills extended the demonstration to December 31, 2031. What changes is which delivery system these members sit in, and Enhanced Care Management and Community Supports exist only inside managed care. The Department of Health Care Services says so in its own July materials. A provider can hold every contract it has, watch the benefit get extended by five years, and still lose a large share of the people it serves.

Four things that follow from reading it this way

Ann Middleton took an earlier version of this list back to her own roster, reordered it, and added a move I did not have. Her reordering is the first item below and it is better than mine was.

Use the access you already have, and ask for what only your plan can tell you. Contracted providers can usually check referral status directly in the plan portal or transaction viewer, and most are not using it. The formal notice is built for the referring entity. If you inquire on a referral you are owed acknowledgement and a status within one business day, and the Department says it does not intend to issue exceptions. The clock runs on your inquiry, and it does not reach every referral. The noticing requirements fall away where the member or family placed the referral themselves, and where the plan generated it from its own data. Then ask your contract manager what no status report answers, which is what level of enrollment and contact keeps you in good standing. Across my interviews this summer an informal floor near 200 enrolled members came up more than once, and one conversation put minimum viable volume closer to 50 and rising. Everyone was working to a threshold and nobody had seen one published. None of it is a state number and DHCS has published nothing.

Write down functional limitation, not just diagnosis. The medical frailty exclusion turns on what a member cannot do without assistance, in the language of the six activities of daily living, with a date on it. Most care plans capture the condition and miss the limitation. This is the highest-yield documentation change available before January, and the federal rule states plainly that a thin claims record cannot be used to deny the exclusion. Work the state version of the exclusion list while you are in there, because California’s runs broader than the federal one and reaches the parent, guardian, caretaker relative or family caregiver of someone with a chronic condition, a disability or a functional limitation.

Lisa Chan-Sawin of Transform Health made the point on the August issue that the frailty rules are in litigation, and that the distinction matters in eligibility work. She is right. California is a co-lead plaintiff in the multistate challenge, and two of the provisions under challenge sit directly under this advice: the requirement that a frail condition also significantly impair the ability to comply, and the twelve-month lookback. The preliminary injunction was denied July 29 for failure to show irreparable harm, so the merits are untouched and the court is moving to decide them before January 1. Congress set the January date, not CMS. None of that changes the instruction, because a dated record of functional limitation is worth having whichever way the rule lands.

Build the Enhanced Care Management to Community Health Worker step-down. Community Health Worker services cannot be billed alongside Enhanced Care Management, and they can run before it begins and after it ends, which makes them the reimbursable way to stay in contact with a member who falls outside the refined January criteria. A community-based organization can be the supervising provider without a licensed provider on staff, and since April 2025 a statewide standing recommendation covers every member who meets the eligibility criteria, so nobody waits on an individual referral. What still gates it is enrollment and signature. Claims come from the enrolled supervising provider, so the organization has to be through the Provider Application and Validation for Enrollment system, and the plan of care needs a licensed provider on it. The clearest write-up of the sequencing came out of Intrepid Ascent’s work on other reimbursable services in Contra Costa. They are one of the organizations we deliver technical assistance alongside.

Get into the reinvestment conversation now, because the window just moved. The initial Community Reinvestment Plans covering 2027 through 2029 were due to the state September 1. On August 26 the Catalyst for Community-led Reinvestment project reported that the Department of Health Care Services revised plans’ 2026 obligation amounts in August and moved the plan deadline to October 15, with final amounts posting in September. DHCS’s own Community Reinvestment FAQ still printed September 1 as of its last revision, so confirm the date with your own plan. If it holds, the six weeks to mid-October are drafting weeks, and a costed proposal that arrives while a plan is still writing lands very differently from one that arrives after it files.

Four regions built the same thing without talking to each other

In the interviews I ran this summer the same structure kept appearing. One person, or one organization, holds a cross-sector table together, and that labor sits outside every CalAIM rate we have.

In San Diego a wellness collaborative has run a Medi-Cal Changes steering committee since January, seven sessions in six months. Their goal, in their own words, is that nobody loses their benefits because they did not know or were not connected. Sacramento, Los Angeles and Contra Costa each stood up provider-led workgroups to outlive Providing Access and Transforming Health. None of the four coordinated with the others and none of them is funded.

All Plan Letter 25-004 carries the opening almost nobody is acting on: the state explicitly encourages plans in the same county to align and pool their reinvestment. That is an invitation to build a county-level coordination backbone with money attached, in a cycle running through 2029. Four convened structures already exist, unfunded, and one funding mechanism explicitly rewards what they are doing for free. They have not met, and the extension is what makes introducing them possible this year instead of next.

What none of the three numbers pays for

Not one of the three reductions touches coordination, because coordination is not a line item in any Medi-Cal rate. It is the work every one of these programs assumes and none of them purchases. When PATH ends the function dissolves, because no entity owns it and no rate carries it.

There is one place the state did put money against navigation, and its size is the argument. The enacted Budget allocates $39 million for care coordination and for clinic and community-based navigators to help members through the January transition. Across roughly two million members that is under $20 each for the year. The intent is real and it is not a funding model.

The reinvestment cycle is the first mechanism in years that could pay for the function properly, and it is being drafted right now by people who mostly have not been asked what to put in it. Worth being in that conversation before the plans file, and worth being in it again after they post, when there is something specific to respond to.

Where this argument is weakest

An accounting distinction does not pay anyone’s bills. If your Enhanced Care Management rate falls because your encounter intensity is low, the state’s savings being modest in General Fund terms is cold comfort. You experience the refinement at your own door.

The General Fund framing is itself a choice. The state manages to General Fund. A provider and a county experience total funds, and on that basis the work requirement in 2026-27 is $357.6 million and not $90.3 million, which is a different and equally real picture. I used General Fund because it is the only basis on which all three reductions can be stated.

Fee for service is also not the same as no coverage. It still pays for physicians, hospitals and pharmacy, so what these members lose is the care management layer and not the care itself. A plan medical director will make that point in the first minute and it is fair. My answer is that the layer is what reaches the people who do not present on their own.

None of which changes what to do this month. Use the portal access you already have and ask your plan the question only it can answer. Write the functional limitation down and date it. Start the reinvestment conversation while there is still a draft to influence. Those hold whichever basis you read the budget on.

~ Jim

Jim Hickman is Founder and Chief Catalyzer of Hickman Strategies. If it would help to walk the readiness grid against your own plan contracts, we will do that with you. Thirty minutes, no charge.

Since 2014, Medi-Cal expansion has mostly asked adults under 65 one question: what do you earn?

On January 1, it starts asking another: do you meet the new community engagement requirement, or do you qualify for an exception?

For many people, the answer may depend on information already sitting in a clinical or care record. That brings providers, care managers and community organizations into an eligibility process they were never built to administer.

The 80-hour requirement gets most of the attention. The harder implementation question may be what counts as evidence to qualify for an exemption.

Diagnosis is not enough

Federal law exempts several groups, including people who are medically frail or have special medical needs that significantly impair their ability to comply, people participating in drug or alcohol treatment, certain caregivers, veterans rated totally disabled, and people who are pregnant or postpartum.

For medically frail members, a diagnosis may not settle the question. Evidence may also need to show how the condition affects the person’s ability to comply.

CMS identifies asthma, hypertension, obesity, ADHD and type 1 and type 2 diabetes among conditions that generally would not establish medical frailty on diagnosis alone.

CMS also names conditions states may reasonably treat as serious or complex, among them cancer, end-stage renal disease, HIV and AIDS, COPD and sickle cell disease. Appearing on that list does not settle the question either. The impairment test still applies. Neither list is exhaustive, and the document that governs is the state’s own condition list rather than the federal example.

California is challenging that interpretation in federal court. The litigation continues while implementation moves forward.

States are expected to use information they already have before asking members for more. That can include claims and encounter data, wage information and records from other public programs.

For medical frailty, some of the most useful evidence may be functional.

Activities of daily living include bathing, dressing, transferring, walking, toileting and eating. Instrumental activities such as shopping and transportation are different. The daily living standard is written into one frailty category rather than all of them, so a member with a serious or complex condition is not required to show it.

Serious illness is rarely static. A person receiving chemotherapy may function well one week and poorly the next. Heart failure, COPD and autoimmune disease can move between stability and acute episodes. Someone recently diagnosed, newly insured or inconsistently connected to care may have little documentation at all.

The care record is now the eligibility record.

An Enhanced Care Management note documenting help with bathing may provide evidence of functional impairment. A note recording only grocery assistance may tell a different story. A member with sickle cell disease may have a full year of emergency and inpatient care in the claims record and nothing in it that speaks to function.

What gets documented could help determine whether a member qualifies for an exception.

Four things worth doing now

1. Start with caregiver status

Family caregivers do not always describe themselves as caregivers.

Ask in language people recognize:

Do you regularly take care of someone in your family or household who needs help because of a disability?

2. Document function, not just services

Care plans often record what an organization did. They do not always record what the member could or could not do.

When clinically appropriate, document limitations in activities of daily living and episodic impairment. An encounter record can miss what happens between visits.

3. Build a community service documentation pathway

Qualifying community service will need to be verified.

Organizations supporting it need a practical way to track activities, dates and hours, and someone who can verify them.

That is an operating workflow.

4. Screen appropriate members for SSI and SSDI

Some Medi-Cal members with disabilities are likely eligible for Supplemental Security Income or Social Security Disability Insurance and not enrolled. That is a second eligibility pathway, and it does not depend on the community engagement requirement at all.

Some conditions have named Social Security listings with specific qualifying criteria. Sickle cell disease is one, with pain crises, hospitalizations and hemoglobin levels each treated as a separate route to a determination.

Timing is the constraint. An initial decision takes months, and many first applications are denied. A member identified in January is not a member covered in January.

Who owns the work?

Functional documentation is not routine in much of community practice. It takes time and often sits outside the work clinicians and care teams are paid to do.

Yet eligibility may increasingly depend on information those teams are well positioned to observe.

That raises a financing question.

If coverage depends on coordination, documentation and verification, someone has to own that work and be paid to do it.

Leaving it in organizational overhead is unlikely to work at scale.

The administrative burden compounds

Community engagement is only one eligibility change coming in 2027.

Many expansion adults will also move from annual to six-month eligibility redeterminations. More eligibility events mean more opportunities for missed notices, incomplete documentation and coverage loss.

California’s Legislative Analyst’s Office estimates the community engagement requirement alone could reduce Medi-Cal enrollment by about 233,000 people by June 2027, although the state is still refining its estimates, including the number of people who may qualify as medically frail.

The implementation challenge is not simply determining who meets the rules. It is making sure the system can recognize them.

Technology can find people. It cannot know them.

Plans and counties will need better ways to identify members approaching renewal, reach them and help them complete the process.

Technology can make that work faster and more targeted. Reaching someone is only the first step.

Some members need a reminder. Some need a person who can help them navigate the process. Some may qualify for an exception, making an hours reminder irrelevant.

The scarce resource is knowing what kind of help the person needs.

Technology enables. Relationships transform.

The organizations closest to members often hold information the eligibility system now needs. The work is getting that information to the right place, in usable form, before coverage is at risk.

That is orchestration over digitization.

Coverage sits upstream

The 1115 renewal matters. Enhanced Care Management design matters. Community Supports, rates and reinvestment matter.

They all depend on enrollment.

The federal changes create work that does not fit neatly inside today’s organizational boundaries. Someone has to identify exceptions, document function, connect eligible people to other coverage pathways and verify qualifying activity.

Counties will administer eligibility. Plans, providers and community organizations will each hold pieces of the evidence.

The implementation question is who connects them.

That is the work to solve now.

Onward,

Jim

Jim Hickman is Founder & Chief Catalyzer of Hickman Strategies LLC. He serves as a Senior Advisor and CalAIM subject matter expert with the Camden Coalition, as a Senior Advisor to PCCI, and as a Venture Partner at Suncoast Ventures. CalAIM Field Notes is independently written and edited.

California's rural transformation application has a required field at its center: what pays for this after the grant runs out. For a rural California project the honest answer runs through CalAIM. In a competition closing August 14, the coordination layer California has built since 2022 is its strongest asset.

In June the NORC Walsh Center for Rural Health Analysis published the Practical Guide to Supporting Innovation in Rural Health, hosted by the Rural Health Information Hub with federal rural health policy funding. It is a toolkit with no regulatory force. But it is plainspoken about a word the field keeps blurring. Innovation in rural health, it says, is coordination that fits the place it serves, and it declines to bless anything that cannot outlive the grant that started it.

The Rural Health Transformation Program, authorized under Public Law 119-21, is moving fifty billion dollars into the states over five years, and $233,639,308 into California for the first budget period. Most states are building their coordination layer from a blank page. California starts somewhere else.

Tuesday, the program stopped being a plan

On Tuesday morning the Department of Health Care Access and Information (HCAI) convened the first meeting of the Rural Health Policy Council, a twenty-member advisory body. HCAI walked them through the initiatives, the grant calendar, and the stakeholder input behind California’s application. That engagement reached sixteen hundred people in eight weeks. They ranked access to care first, at seventy-seven percent. Maternal and child health came in at twenty.

Now look at what the money rewards. Applicants must select at least one Transformative Care Model metric. Two of the six named metrics are maternity: first-trimester prenatal care, and uncomplicated vaginal birth after cesarean. So a constituency that ranked tenth of eleven anchors a third of what the money measures. Call that an opening. It belongs to whoever moves on it.

The head start has a name

California reads this guide fluently, because it spent years building the substrate the guide describes, for Medi-Cal rather than rural health. Enhanced Care Management is care coordination across settings, delivered by people the community already trusts. Community Supports are the whole-person services that keep someone well between visits: housing navigation, medically tailored meals, respite. Those networks are what the guide tells rural states to go find.

The application asks the question directly

Until this month that was an argument about good design. The published application turns it into a field you have to fill in.

Section 7c of the Accelerator Partner Request for Applications is a required attestation titled Financial sustainability. It asks how the project continues after grant funding is fully expended, and it names what it wants to hear: sustainability strategies, partnerships, reimbursement pathways, operations efficiencies, long-term implementation plans. Five hundred words.

The allowable-use rules point the same way. The Grant Guide closes two federal categories to Accelerator subrecipients: Category B, provider payments, and Category J, capital. Federal rules decline payments for clinical services that duplicate billable ones. Accelerator dollars build the coordination layer. Paying for the care falls to someone else, and the application wants to know who.

What that answer looks like on the page

Picture a critical access hospital in a county where no hospital delivers babies. It applies with the health center that sees most of the county’s pregnant patients, a regional hospital ninety minutes away that still has an obstetric unit, and a community organization already working the same valley. The metric is first-trimester prenatal care. The example is illustrative, the mechanisms real.

The grant builds the connective tissue. A shared risk and transfer protocol across the three sites. A tele-obstetrics pathway configured into the record at both ends, so the regional obstetrician sees a patient without either of them driving. Training for the care team and the outreach staff. Scheduling coordination, so a positive pregnancy test produces a confirmed appointment that week rather than a phone number. Data capability to count first-trimester starts. One-time costs, none paying a provider or buying a building.

So if Medi-Cal already pays for care management, what is the grant buying? Timing. In practice a plan identifies a pregnancy through claims, so in the earliest weeks, when the first prenatal visit either happens or does not, it cannot yet see the member. Nothing it funds can reach her there. Finding her in that window is work no one pays for, and it is the difference between a first-trimester start and a second-trimester one.

Then Section 7c. The visits bill to Medi-Cal as they always have. For members who meet the Enhanced Care Management Birth Equity criteria, live since January 2024, the plan authorizes and pays care management through a contracted provider. Doula support bills through the Medi-Cal doula benefit. Community Supports carry what the clinic cannot bill. No existing funded position is backfilled, because the grant pays for the build and the plan pays for the benefit. Section 7d makes every applicant attest they will not supplant existing funding. The braid runs on sequence.

That is most of a five-hundred-word answer, and a real one.

The last part is the honest part, and Section 7c asks for it: name what still needs support. Here it is the convening. Somebody keeps three organizations in a room after launch, holds the protocol current, chases the metric. No one bills for that, and the grant that pays to build it ends. It is the smallest line in the budget and the first to disappear. Saying so is more credible than claiming the thing self-sustains on day one.

One more required attestation belongs next to Section 7c. Section 7b asks every applicant to attest it is willing to serve as a regional incubator, support peer learning, document what it learns, and help scale what works. The Grant Guide says why: lessons from Accelerator Partners are meant to inform statewide scaling of the hub-and-spoke framework. So an applicant naming Enhanced Care Management and Community Supports in Section 7c carries one project past its grant and drafts the pathway the next cohort copies. Four more budget periods follow. Whoever answers this first sets the template.

The seam runs between two departments

CalRHT lives at HCAI. CalAIM lives at the Department of Health Care Services (DHCS). The risk in front of California is parallel play.

Tuesday’s roster is worth reading with that in mind. The Council is broad: rural hospitals, rural clinics, county public health, two Tribal partners, two consumer advocates, a health plan, a managed care medical director, and a nurse-midwife running a district hospital midwifery program. DHCS holds one seat, and it is the Value-Based Payment Branch Chief. The Enhanced Care Management network holds none.

Take that as an observation. A rural transformation council should center rural facilities and the people they serve, and workgroups may follow as the work develops. Still, if the sustainability field in every application runs through CalAIM, the room gains from someone who has stood up Enhanced Care Management in a county with one hospital. Members were asked to bring emerging issues forward. This is one.

What this asks of us, and by when

For rural providers, your Enhanced Care Management and Community Supports relationships are an asset the application rewards, and they belong at the hub-and-spoke table now. For community-based organizations, the RFA names your role: outreach and patient navigation alongside a hospital applicant and a Federally Qualified Health Center.

Partnerships are optional, and lacking them can affect scoring. The Accelerator grant guide, request for applications, and budget workbook are published. Applications are due August 14 at 3:00 p.m. Pacific, through the portal the request for applications names at calruralhealth.submittable.com. Write against the published documents and confirm access before the deadline. HCAI submits California’s Budget Year 2 application to the Centers for Medicare and Medicaid Services on August 31, which means next year’s design conversation is happening this month.

The guide handed every state the same definition of innovation. California already built the operating system that definition describes. The remaining work is orchestration across two departments and one rural patient. We have the two rooms. We need the door between them, a trusted person standing in it, and the financing that keeps them there.

Onward,

Jim

Disclosure: Hickman Strategies advises organizations active in rural health, including some that may participate in CalRHT funding opportunities. This is independent analysis, written for no applicant and on behalf of no agency.

Sources: the CalRHT Accelerator Partner Grant Guide and Request for Applications, Version 1.0, June 30, 2026, and materials presented at the Rural Health Policy Council meeting of July 28, 2026. Rural Connection, CalRHT Partner Finder, Rural Connection Coalition (CSRHA, CHA, CPCA), calrhtpartners.org. Stakeholder percentages date to application development. Questions to info@calruralhealth.org.

I joined Wednesday’s fourth meeting of the Future of Medi-Cal Commission by video. Four work groups reported out to co-chairs Dr. Mark Ghaly and Ann O’Leary, who want the whole effort to land on roughly five recommendations. Ghaly was firm about the test each one has to pass. Every recommendation must trace back to a principle the Commission has settled on, equity and access, accountability and outcomes, capability and innovation. In his framing, a slate that does not is a failure.

One thread ran under all four reports. I want to name it, then tell you what I filed in response.

The demand question is closing

DHCS reported nearly 227,500 people in Enhanced Care Management in the third quarter of 2025, up 59 percent from a year earlier, and 191,000 members using Community Supports in the same quarter. Enrollment in both is at a record, and the services themselves are no longer the open question. DHCS now finds ten of the twelve Community Supports it studied cost-effective. What the Commission spent the day circling was whether the system underneath those members can deliver the coordinated care the services promise.

The Rationalized System work group went around on one question and could not resolve it: who is ultimately accountable for a member across the departments of Health Care Services, Social Services, Developmental Services, and Aging. The one place it reached consensus was on sequence: simplification before integration. Fewer moving parts first, then the work of connecting what remains. That is a defensible order, and it puts the harder half second. Strengthening accountability and coordination is that group’s charter, so this is a gap the Commission assigned itself and has not closed. The Member Experience group named lapsed members as its most worrisome population, the ones who fall out of care and never reconnect. Rationalized System, with Technology, Data and Infrastructure, named real-time longitudinal data as the prerequisite for almost everything else. And the Payments group is designing a base-and-performance payment structure with no immediate place in it for the coordination that ties services together.

Four reports, one gap underneath all of them. The care sits in contracts. The coordination between services sits in nobody’s contract. Two of the four reports are really one report. Longitudinal data is what it would take to see a member whole, and accountability is hard to assign while no single view of the member exists, which puts the accountability question downstream of the data question. There is a way to test that. For a member enrolled in Enhanced Care Management and also receiving a Community Support, which of the four departments can produce a current view of everything that member is authorized for? Whoever can already answer is the candidate.

That gap is about to widen, and not by California’s choosing. The federal community engagement requirement arrives as an interim final rule, effective this Friday, July 31, with state implementation required by January 1, 2027. Comments close the same day it takes effect. The rule bars managed care plans from determining whether a member has met the requirement and restricts what they can delegate, which leaves the entities holding the richest claims and encounter data outside the determination that data would inform. Someone has to do that work. The rule names no one, and it carries no money.

What I put in the record

I facilitate CalAIM implementation across California counties and advise the plans, providers, and community organizations doing this work, so I filed a two-page written comment during the public session. It makes one argument: fund the coordination layer. Three actions, sized to the slate of five the co-chairs want. Between them they cover all three principles, each names a legislative, budget, or executive pathway, and each has a version that holds under today’s constraints.

Put coordination in the base payment tier, not the margin. The Payments group is already building the mechanism: base stability payments held separate from performance payments tied to outcomes. Coordination belongs in the base, because plans and providers cannot produce the outcomes the performance tier will measure without it. Funded as a line, it stops competing against direct service every budget cycle. The case for a line rather than a margin got sharper on July 21, when CMS deferred $867.5 million in federal matching funds to California pending documentation supporting the underlying claims. Work that lives in overhead is the hardest work in the system to evidence when someone asks to see it.

Name a local owner. The Rationalized System work group’s unresolved accountability question has a practical answer: designate a local orchestrator accountable to the care model rather than to a single funding stream, with a defined role in the regional structures the group is mapping. That mapping will show where the state’s investments in Enhanced Care Management, Community Supports, and behavioral health sit. Someone then must be authorized to act on it, or the work produces a description of the system instead of a change to it. The federal rule has now put a date on that question. Whoever California names to hold accountability for a member will also have to hold the exemption and verification work when the requirement goes live.

Use coordination as the retention and learning instrument. The organizations closest to members hold the freshest contact information in the system and the relationships that make outreach land, and in the counties I work in it is personal outreach, not mail alone, that moves people through a renewal. That churn is not hypothetical. CHCF’s enrollment tracking shows Medi-Cal down about 5 percent between June 2025 and March 2026, from 14.78 million to 14.05 million, a loss of roughly 730,000 people. More than half of that decline, some 378,000 people, came in the first quarter of 2026 alone. The pace is accelerating, and all of it sits before H.R. 1’s major provisions take effect. Demand for the benefits is at a record while the base beneath them shrinks.

Fund and instrument those organizations, and coordination becomes the retention mechanism for a redetermination cycle moving to twice a year, and a live data source for the oversight the work groups said they lack. It also builds the feedback loop the meeting named as missing, because that is where an intervention’s effect shows up in weeks instead of in a two-year-old evaluation.

The comment carried one exhibit: a readiness assessment I use with frontline organizations across the state, mapping four layers of organizational capacity against four stages of maturity, with the managed care plan contracting floor marked. Most community organizations sit just below that floor, and the distance measures underfunded potential, not performance failure. The framework is not new. I published it in Field Notes No. 17 in May and presented it at the SCAN Foundation’s Bridging Capital and Care convening at UC Berkeley’s School of Public Health.

The asset a plan cannot buy

Underneath all four layers sits the one thing a plan cannot buy: the trust a community organization holds where it works. That is what speakers were describing when they talked about being present at the hardest moments, and about already holding the member data and relationships the work groups kept asking for. Coordination is what makes that trust usable at the scale a state contract requires.

Public comment on Wednesday cautioned the Commission against relying on efficiency gains to fund what comes next. The harder question sits one step back. Efficiency asks whether we are doing things right. Readiness asks whether the organizations we are counting on can do the thing at all, before we align payment and policy on top of them. That question comes first.

I am developing this argument with colleagues at the Camden Coalition, and that fuller version is headed for Health Affairs. A closer read of the four reports, and where I think the order of operations sits, is coming here in two weeks. The Commission is doing serious work under real fiscal pressure. The recommendation I will be watching for is the one that treats provider readiness, and the coordination it enables, as infrastructure worth financing rather than a box checked the day a contract is signed.

Five recommendations are a tight number. If only one of these three fits, it should be the second, because the other two describe how a named owner gets funded and instrumented. I hope one of the five is this one.

Jim

CalAIM Field Notes is a working notebook for the possibilists inside CalAIM. If a colleague on a county, plan, or community team would find this useful, send it their way.

Series note: If you missed SE-12: Two Million People Are Being Moved, read it first. This essay builds directly from that discussion. https://calaimfieldnotes.substack.com/p/se-12-two-million-people-are-being

Walk past a construction site long enough and you learn to tell scaffolding from structure. Scaffolding goes up fast, holds people while they work, and comes down when the funding stops. Structure is what stays. It is load-bearing, designed once, and built to outlast the crew.

For a decade, California has been remarkably good at putting up scaffolding around Medi-Cal, and remarkably reluctant to pour the foundation underneath it.

This summer made the pattern visible again. The 2026-27 budget protected Medi-Cal’s floor against H.R. 1, the federal law signed a year ago on July 4, 2025. It delayed an asset test, postponed dental reductions, funded county eligibility workload, and stood up navigators. Those are real protections and I am glad we have them. They hold the building through a hard season. They are not the structure.

The question I keep meeting in county rooms is simpler. After every protection is in place, who is accountable when a member moves between a health plan, a county behavioral health department, a hospital, and three community organizations in a single month?

Start with what is working, because it is a lot. Since 2022, nearly 453,000 members have received Enhanced Care Management. More than 528,000 have accessed a Community Support, across more than 1.5 million services. In the third quarter of 2025 alone, Enhanced Care Management reached nearly 227,500 people, up 59 percent year over year, with children and youth up 90 percent.

It is proving out on cost as well as reach. The Department of Health Care Services studied twelve Community Supports authorized as In Lieu of Services and found ten already cost-effective against emergency visits, inpatient stays, and nursing facility stays. Environmental Accessibility Adaptations cut net costs 51.4 percent. Housing Deposits, 15.2 percent. Medically Tailored Meals, 10.5 percent.

DHCS drew the right conclusion. In June it said plainly that these services are built into the core structure of Medi-Cal managed care, not temporary pilots, and that permanency lets everyone keep investing in the relationships that make them work.

That concedes my argument before I make it. Permanence is what makes relationship investment rational.

Now look at what did not get that treatment.

The services are in the contracts. The layer that connects them is in nobody’s contract. Accountability for the ecosystem’s long-term performance still sits across financing, governance, and operating structures that were never designed to hold it.

That is not a failure of commitment. It is what happens when we fund the transaction and treat the connective layer between transactions as overhead.

The room writing the next decade has already named this

I do not have to make this case alone. The Future of Medi-Cal Commission, co-chaired by Mark Ghaly and Ann O’Leary, is building a ten-year roadmap for the next governor across care delivery, financing, governance, and infrastructure. Read its spring materials and the diagnosis is unmistakable.

They describe a program assembled in layers instead of designed as a whole, leaving fragmentation, uneven capability, and unclear accountability. They find that money does not reliably deliver the stable funding long-term investment requires, and that approved payments can take years to reach providers. And across every workgroup runs one theme: no single actor is accountable for a member’s whole-person outcome, because the accountable parties sit too far from the people doing the work.

That is not my framing borrowed by the Commission. It is the Commission’s framing.

So the question in front of it is not whether coordinated care works. CalAIM has gone a long way toward answering that. It is how a proven demonstration becomes permanent operating infrastructure.

The field said it first

When the California Health Care Foundation asked the field for bold ideas for the next decade, 132 submissions came back, and the most common theme, in nearly two dozen of them, was payment innovation. Stakeholders see financial realignment as a prerequisite for care transformation.

The budget says the same in dollars. We found money for eligibility caseload and navigators, both of which matter. We did not find a line for the operating model that turns a funded navigator into a retained member. The scaffolding got appropriated. The structure did not.

Name the backbone, and it becomes fundable

The backbone goes unfunded partly because we have not named it as a thing with a design and a price. Overhead gets cut. Infrastructure gets built.

It has a readiness dimension: the capabilities a region needs before coordination can hold, which is why one county absorbs a reform in a quarter while another stalls a year. It has a governance answer: a county orchestrator accountable to the model rather than to any one institution, which answers the Commission’s own finding that no one owns the outcome. And it has a financing answer, where the decade is won or lost. Call it dual-capital orchestration: braiding program dollars with reinvestment and longer-horizon capital so the backbone has a funding source rather than a hope.

The vehicle already exists, and its clock is running

California has already built a channel that could fund it. Under All Plan Letter 25-004, Medi-Cal managed care plans must reinvest a portion of net income into local communities. The Catalyst for Community-Led Reinvestment Project puts first-cycle obligations near $299 million. Initial plans are due September 1 and set the cycle through 2029. Each must align with the local health jurisdiction's community health assessment and carry an attestation from its public health director.

The plans directing nearly $300 million are being written right now. That is not a future debate. It is a live one.

And it is the whole argument in miniature. Reinvestment becomes impact only through an operating model: alignment with local priorities, advisory committees with real influence, trusted partners positioned to deliver. The dollars are the permission. The model is the work.

If you want to know what your county is owed and where planning is underway, the Catalyst Project built a county tool that will tell you. It is a partnership of the Prevention Institute, Golden Policy Partners, and Jeremy Cantor at JPC Health Strategies, backed by the California Endowment. If you have been waiting for a place to start, start there.

One caution, and it is the deeper reason to build structure rather than rent scaffolding. Community Reinvestment is a slice of net income, and only plans in the black owe it. As H.R. 1 squeezes plan financing, that pool can shrink in exactly the years communities need it most. A backbone funded from residual profit disappears in a downturn. Design the financing. Do not wait for the surplus.

Programs demonstrate what is possible. Infrastructure determines what endures.

The window is open, and it does not stay open

Hold onto the asymmetry. The floor was defended this year, in a budget cycle that resets every twelve months. The ceiling is being drafted now, for ten years. The Commission continues its work this fall, delivers its report by the end of the year, and hands the next governor a roadmap in early 2027.

The door is not metaphorical. The Commission meets in public, next on July 23 in Los Angeles and September 21 in Oakland. The July 23 public session runs 9:30 to noon, with public input on the agenda, and you can register to attend in person or by Zoom. The written channel stays open between meetings, and staff host virtual comment sessions every other week. Right now, it is listening.

Two pages: what your backbone costs, and who is accountable for running it. A region that can write them is ready to shape a reinvestment plan now, and to put evidence in front of the Commission next.

We build this together, or we do not build it

The counties, the plans, the hospitals, the community organizations, the eligibility specialist working bedside in an emergency department, each hold a corner of the same structure. The Commission needs to hear, in their words and with their numbers, what it takes to build trust as infrastructure, because trust is the load-bearing element no appropriation can buy and no platform can replace.

So let us bring the evidence to the rooms where the next decade is being drawn. Shape a reinvestment plan before September. Offer two pages and two minutes to the Commission before it writes the roadmap. We spent a decade proving that coordination works. Let us spend this window making sure we fund the thing that makes it last.

The work now is to build the structure, not rent the scaffolding.

~ Jim

Jim Hickman is the founder of Hickman Strategies LLC and a senior advisor to the Camden Coalition. He works at the intersection of healthcare policy, innovation, and system transformation. CalAIM Field Notes is a working notebook for the possibilists inside CalAIM. If a colleague on a county, plan, or community team would find this useful, send it their way.

This weekend a promise turns 250.

Life, liberty, and the pursuit of happiness. We say it like it is about the big things. Freedom. Rights. The vote. It is. But underneath every one of them sits something smaller and more stubborn. Time. You cannot pursue anything without the hours to do it in. The founders promised freedom. What most people run short on is time.

Much of that time goes to work the country never sees. Caregiving. The daughter on her lunch break fighting a billing error for her father. The son who leaves work early to drive his mother to dialysis. More than sixty million people carry it, most of them women, almost none of it on a payroll. Hire that work out and it would cost about a trillion dollars a year. It holds the whole system up, and we have never once priced it.

Give one of those caregivers one hour of the day, and you hand a family back a piece of its life. The hour is the asset.

Here is the part I keep turning over. It is for you, the one working two jobs.

Whose hour comes back?

If you have money, you have always been able to buy your hour back. You hire the aide. You order the groceries. You pay someone to sit with your mother so you can sleep. The wealthy have never had a time problem a check could not solve. That is the liberty most families cannot buy.

And if you are working two jobs, the math is worse. You cannot buy the hour, because there is nothing left to spend. You cannot give it away either, because you need the wage. So you work both jobs and carry the caregiving on top, and the hour comes out of your sleep. Out of your health. Out of a retirement you keep putting off.

A better tool helps the person whose problem is a crowded calendar. That is not your problem. Yours is that you have to be in the room, and you cannot miss the shift.

CALAIM FIELD NOTES

Ground truth from California’s Medicaid transformation.

Written by Jim Hickman, Principal of Hickman Strategies, from the field of California’s Medicaid transformation. The views here are his own.

Subscribe to follow the work, and forward this to someone carrying it.

jim@hickmanstrategies.com · hickmanstrategies.com

Technology Enables · Relationships Transform™

California enacted a $351.7 billion budget this week, balanced across two fiscal years, with record reserves. The headlines are right that it defended most of the Medi-Cal floor. After a year when the Governor’s proposals would have cut deep, the Legislature rejected the worst of it and delayed the rest. That is a real win, and the people who fought for it earned the credit.

So start with what held, because it matters. The budget rejected the Governor’s cuts to In-Home Supportive Services, to the PACE program for frail elders, and to the acupuncture benefit. It delayed the return of the Medi-Cal asset test, which now arrives at $21,000 in July 2027 rather than dropping to $2,000 now. It delayed the dental and immigrant coverage cuts by a year, to July 2027. It put $197 million toward the county eligibility workers who will carry the H.R. 1 workload, $250 million toward public hospitals, and $100 million toward Covered California subsidies for the lowest-income enrollees. It renewed the managed care financing that keeps Medi-Cal whole, and it started a process to make large employers share the cost of covering their own workers. Program by program, the floor was defended.

Then there is the move that did not make the headlines.

The same budget adopts the transition of roughly two million immigrant Californians from Medi-Cal managed care into fee-for-service, beginning January 1, 2027. Federal guidance drives the split: to keep federal funds flowing, the state is required to run a separate delivery system for this population. On paper it reads like a delivery-system technicality. For anyone who works in CalAIM, it is not. Fee-for-service has no Enhanced Care Management and no Community Supports. Those are the two benefits at the heart of whole-person care: the care coordination, and the housing, food, and recuperative supports that keep high-need members stable. Two million people will keep their coverage and lose the coordination activities that come with Medi-Cal managed care. The softening is $39 million for navigators to ease the transition.

This is the tell. Federal rules forced the delivery-system split. Inside that requirement the state still held a choice: backfill the coordination these members are about to lose, or bank the savings. It banked the savings. The move is scored to save roughly $584 million next year. Set against that, the softening is $39 million for navigators, close to fifteen dollars saved for every dollar spent easing the way out. When a budget has to choose, and this one did, the connective tissue is the part that gives. Not because anyone set out to cause harm. Because the backbone is treated as discretionary, the line you trim in a hard year, rather than structure you would never think to touch. Coverage is the card in the wallet. Coordination is what makes the card mean anything. We protected the card for two million people and let go of the thing that makes it work.

And the timing sharpens it. The budget lands the same year the federal clock starts. Outreach notices to enrollees go out across this summer, and H.R. 1’s work requirement takes effect January 1, 2027, asking expansion adults to document eighty hours a month of qualifying activity. Eligibility checks move from once a year to twice. Every new touchpoint is one more place a still-eligible person falls through on paperwork. The floor was defended. The churn is just beginning.

So what do we do with the time the budget bought?

For the millions staying in managed care, the answer is the one this notebook keeps returning to. Ex parte first, always: answer every renewal and every work-status check from data the state already holds, before a packet goes out. Confirm, do not reconstruct: ask members to verify what we know rather than rebuild their lives from scratch. And solve the contact problem first, because a disconnected phone blocks more renewals than apathy ever will.

For the two million moving to fee-for-service, the $39 million in navigators is the seam to watch. Whether those members keep any thread of coordination will not be decided by the appropriation. It will be decided by the operating model built around it: who owns the handoff, who picks up the phone, who sits beside them when the system changes underneath them. This is where the plan and the community health center find each other. In a recent Community Health Centers Unfiltered conversation, the CEO of Medicaid Health Plans of America and the CEO of Advocates for Community Health landed on the same point from opposite sides of the table: through this stretch of federal change, plans and health centers keep access alive together or not at all. Orchestration over digitization is not a slogan here. It is the operating model that decides whether two million people keep a thread of care.

And notice what else the budget did while it deferred the hard fights. It handed the next administration a deadline of March 1 to produce at least one plan to make large employers share the cost of covering workers who land on Medi-Cal. The financing fight moved to next year, the same as the coordination fight. A possibilist reads that as an opening. The same budget that let the connective tissue slip also planted a revenue mechanism a future administration could use to fund it, permanently, out of dollars the state does not yet collect. Groundwork is not a guarantee. It is a place to build.

That is the whole lesson in one budget. The floor was defended because people fought for it, program by program, in a hard year. A backbone we have to defend every budget cycle is not yet structure. The work now, for counties, plans, hospitals, and the community organizations holding the other end of the line, is to make the connective tissue permanent, so the next hard year does not reach for it first.

Jim

CalAIM Field Notes is a working notebook for the possibilists inside CalAIM. If a colleague on a county, plan, or community team would find this useful, send it their way.

The front door is changing

For years, the way into Medi-Cal was a county office, a paper form, a thick envelope. Now it’s a chat window. AI tools that screen eligibility, explain benefits, and start enrollments are racing to be the first thing a member meets.

This isn’t hypothetical. In April, through its Bold Ideas for Medi-Cal series, the California Health Care Foundation published a proposal from Pair Team and BluePath Health that names the shift outright: AI “Advocates” for members, an “AI app store” where the state governs the data and certified vendors build the assistants on top. They’d run on Medi-Cal Connect, the state’s population health data service, linked through the open Model Context Protocol so the data stays where it already lives, watching for rising risk and walking people through enrollment in their own language. It’s a serious proposal, and it makes the front door concrete.

And the promise is real. An AI front door can answer at midnight, in someone’s own language, without making anyone feel small for asking. For people told in a hundred quiet ways that the system wasn’t built for them, a door that opens on the first knock matters.

The question the hype skips

But a front door isn’t care. It’s an entrance, and an entrance is only as good as the room it opens onto.

When the AI says “you qualify,” who picks up the enrollment? When it says “you need a specialist,” who handles the referral, the authorization, the ride? When a member quietly drops off at renewal, does the door even notice?

The companies racing to own the door are good at the door. That’s the easy part. The hard part is everything behind it: the warm handoff, the trust, the money that makes the next step real.

Orchestration, not digitization

The team that wins won’t have the slickest AI coach. It’ll be whoever connects the layer underneath, the part that turns “you qualify” into “you’re enrolled, and someone here knows your name.”

The most serious version of this idea already knows that. The CHCF proposal keeps the state holding consent, keeps a human in the loop, and routes members to Enhanced Care Management instead of replacing them. The authors say it plainly: trust is a precondition for adoption. That’s orchestration thinking, not digitization.

We’ve said it before: technology enables, relationships transform. An AI front door is technology at its best, fast and tireless. But the transformation still runs through people: the community health worker who picks up where the chatbot stops, the data that follows a member across systems, the contract that pays someone to keep the door open instead of just installing it.

David Brailer, who led the federal push to digitize medicine in the 2000s, looked back this month and was blunt about what that era mostly delivered: by his account, its lasting legacy was “the digitization of billing at scale,” not better care. A technology never sets a system’s incentives; it carries them out, faster and at greater scale. He cites one number that captures the risk: AI-enabled hospital coding added $2.3 billion in excess spending in a single year, by making patients look sicker on paper than their records supported. An AI front door bolted onto a system that never built the human and financing layer underneath won’t transform it. It will only digitize the same extraction, faster.

That’s orchestration over digitization: the gap between a front door that generates a lead and one that delivers care.

The rules are catching up

AI found the front door. Someone still has to let it in, and increasingly that’s a question the law is answering. In early 2026, Manatt counted more than 240 AI-in-healthcare bills across 43 states, nearly as many as in all of last year, most aimed at the exact pressure points an AI front door creates.

Consent leads: states want providers to disclose when AI is in the loop, want chatbots to admit they aren’t human, and about a dozen now bar non-humans from claiming titles like “nurse.” Coverage is close behind: after California’s 2024 law, at least eleven states limit AI in prior authorization, most requiring a human to review any denial, and Indiana became the first to rein in downcoding, the quiet swap of a claim for a cheaper code.

Accountability is forming fast, too. California is weighing a nurse-backed bill that would let clinicians override an AI decision without retaliation, and a companion Bonta bill that would keep AI from standing in for a professional’s judgment; both could pass before August. The front door is a governance decision as much as a product one.

Meanwhile, the rails get built

And it isn’t all restrictions. At the federal level, the plumbing is being laid in the open. On June 9, CMS stood up an Office of Health Technology and Products, pulling interoperability, the National Provider Directory, digital identity, Medicaid and CHIP technology, and a first enterprise AI strategy under one roof, with a default toward open-source code. It’s already shipping pieces: Emmy, short for Eligibility Made Easy, is an open-source, consent-based tool to help states verify the new Medicaid work requirements with fewer dropped enrollments. The rails for a national front door are going in. What still isn’t funded is the local layer that turns a national standard and a 24/7 agent into a warm handoff for one person on one block. Plumbing reaches the house. It doesn’t introduce itself to the person inside.

The door is already open, and California is deciding who holds it

We don’t have to imagine the door swinging open. On June 12, California’s technology department reported six high-risk automated decision systems already running across state agencies, scoring recidivism, flagging unemployment claims, judging student work. A year ago the same disclosure named zero; the state had just never looked closely enough. Senate Bill 1248, which would have barred agencies from using an automated system as the sole basis for a decision, was set aside last month, and Poppy, the state’s new AI assistant, goes statewide next month.

At the same moment, a blueprint is taking shape. The Future of Medi-Cal Commission, an independent, foundation-convened body co-chaired by Dr. Mark Ghaly and Ann O’Leary, is drafting a 10-year roadmap, with recommendations due in early 2027. Its draft materials name the question directly: how to use AI responsibly and protect privacy to keep members safe. Commissioners have named the risk in their own terms, that efficiency-first technology erodes equity unless equity is built in from the start, and that no single entity is accountable for a member’s whole-person outcome. (We covered this Commission in Field Notes back in December.)

Read together, these aren’t in tension. They’re an assignment. One shows the front door already in use; the other shows a serious body describing the layer that holds it open. The orchestration layer isn’t ours to invent from scratch. California is drafting it, in public, right now, and the work is to get it into the roadmap before that roadmap finalizes.

So who holds it open?

For anyone building or buying an AI front door this summer, the questions are concrete. Who owns the handoff when the AI says “next step”? Whose data is it, and what consent travels with it? What pays for the human layer, not just the software? When a member falls out, who reaches back?

The CHCF proposal answers a lot of this on paper, and answers it well. The real test is whether the build matches the blueprint, and who’s accountable when it doesn’t. If the answer is “nobody yet,” the front door is just an efficient way to walk people up to a threshold they still can’t cross.

Let’s build it

None of this is a reason to slow down. It’s a reason to build the rest of the system with the same energy we’re pouring into the door. The population is already there. The safety net is already there. What’s up for grabs is the infrastructure that connects them.

There’s a concrete test already on the table. The federal ACCESS model pays for chronic-care outcomes instead of visits, and though it starts in Medicare, the debate has already turned to whether it belongs in Medicaid. The clearest case is maternal health: Medicaid covers roughly four in ten births, and the postpartum months are exactly where the system tends to lose people. Remote monitoring and coaching could close part of that gap. The missing piece has been reimbursement, and an outcomes model could supply it, but only if someone orchestrates the enrollment, the consent, and the handoff back to a human the moment a reading goes the wrong way.

So if you’re a health plan, a county, a CBO, or a company working in this space, bring the front door. Then let’s talk about who holds it open.

AI found the front door. Holding it open is on us.

Join us on August 18

I’m taking this up live with the California Improvement Network, a project of CHCF managed by Healthforce Center at UCSF: Making CalAIM Work in a Constrained Fiscal Environment. Tuesday, August 18, 12 to 1 PM PT.

Three invitations for the people in this network:

Health plan and county leaders: bring your hardest implementation question. We’ll spend the hour on cross-sector collaboration and financing approaches you can put to work in your own region.

Philanthropic funders: the infrastructure here is concrete and solvable. It doesn’t take backfilling billions, just targeted investment in what’s already working. That’s exactly what strategic philanthropy is built for.

CBO and provider partners: what’s the first funded pilot that proves the model in your geography? The infrastructure for the conversation exists. Let’s use it.

Register for August 18 → https://ucsf.zoom.us/meeting/register/l49ONXkfSvSHHRdYub701g

Prefer to talk it through? Email Jim at jim@hickmanstrategies.com, or visit hickmanstrategies.com.

Postscript (June 25, 2026)

Updated after catching up on the June 11 Connecting for Better Health (C4BH) convening.

After publishing this piece, I caught up on a few meetings I had missed, including the June 11 Connecting for Better Health (C4BH) convening. I expected to hear more about artificial intelligence. I came away with a clearer picture of what California is actually building.

Speaker after speaker returned to the same conclusion. The technology is not the hard part.

The hard part is building trust between organizations. It is agreeing on governance before procurement. It is deciding who owns consent, who maintains audit trails, who reviews adverse decisions, who shares data responsibly, and who stays accountable when a member falls through the cracks. The technology is moving quickly. The relationships take longer. The most encouraging lesson of the day was that once governance and partnerships are in place, the technical work moves surprisingly fast.

Two moments made that concrete.

The first was a discussion of H.R. 1 implementation. The clearest takeaway was not about any algorithm. It was that the infrastructure for sharing data already exists. What is missing is coordination and the patient work of peer learning across counties, plans, and community organizations.

The second was a set of responsible-AI best practice guides built by a cross-sector working group. They read less like a research paper and more like a procurement checklist: no fully automated denials, a human in the loop on every adverse decision, no defaulting to denial when data is missing or stale, and real accessibility for people who cannot or will not use a digital pathway. Governance is no longer aspirational. It is becoming operational.

The conversations connected a series of efforts we usually discuss separately: responsible AI, the Data Exchange Framework (DxF), CalAIM, Transitional Rent, housing partnerships, behavioral health integration, Medicaid renewals, and H.R. 1 implementation. Different programs. Different funding streams. Different acronyms. Increasingly, they are asking the same question.

Can California operate as one coordinated ecosystem instead of a collection of disconnected programs?

That may be the most important story emerging in 2026.

For years we treated interoperability as mainly a technology challenge. The June 11 discussions suggested something different. The real challenge is governance. AI is not creating an entirely new operating model. It is exposing the limits of the one we have been working within.

Which brings me back to a principle I return to often. Technology enables. Relationships transform.

The organizations that lead over the next decade will not necessarily build the smartest AI. They will build the strongest partnerships. They will establish trusted governance, align incentives, and orchestrate across health care, housing, behavioral health, social services, and community organizations, so technology amplifies relationships instead of trying to replace them.

When I wrote this article, I argued that AI had found the front door. After catching up on these conversations, I would add one observation.

California is not just building a new front door. It is beginning to build the operating system behind it.

That is the opportunity in front of us. The next chapter of CalAIM will not be defined by better algorithms alone. It will be defined by how well we orchestrate relationships across an increasingly connected ecosystem.

The front door matters. The operating system behind it will determine whether people actually make it home.

CalAIM Field Notes · calaimfieldnotes.substack.com

Jim

Sources: California Health Care Foundation, “Medi-Cal Bold Idea: AI-Powered Advocates for Medi-Cal” (Bold Ideas for Medi-Cal, April 2026). David Brailer, “Why AI Will Accelerate Health Care Inflation,” Health Affairs Forefront (June 11, 2026). CMS Innovation Center, ACCESS Model (2026); Second Opinion, “If the ACCESS Model works, should it be expanded to Medicaid?” (June 10, 2026). CMS Office of Health Technology and Products, Federal Register notice (June 11, 2026; effective June 9). CMS Eligibility Made Easy (Emmy), Medicaid community-engagement support. California Department of Technology high-risk automated decision systems report (June 2026), via CalMatters. Future of Medi-Cal Commission published meeting materials (2026); recommendations due early 2027. State AI-legislation figures via Pluribus News / State Affairs and Manatt’s 2026 tracking.

CalAIM Field Notes · calaimfieldnotes.substack.com

Jim

CalAIM Field Notes is a working notebook for the possibilists inside CalAIM. If a colleague on a county, plan, or community team would find this useful, send it their way.

On June 1, the Centers for Medicare & Medicaid Services (CMS) issued its interim final rule implementing Medicaid work (community engagement) requirements, with public comments open through July 31, 2026. The headlines will reach for the largest number they can find: the count of people who might lose coverage. That number matters. But it points us at the wrong question.

The risk in front of California isn’t that millions are suddenly ineligible for Medi-Cal. Most of the people we’re about to worry about still qualify. The risk is that proving it ends their coverage. Starting in January, the new adult group renews every six months instead of once a year, with work-status verification layered on top. The barrier isn’t who you are. It’s the paperwork that proves who you are. We have turned a routine renewal into a high-wire act, and now people cross it twice a year, not once.

This week I sat in three rooms across the state: one on the Central Coast, one in the Central Valley, one in the North Bay under a different health plan entirely. In each, I was with the counties, plans, hospitals, and community organizations doing the work of keeping people covered. Different geographies, different caseloads, one finding so consistent it stopped being an anecdote and became the story.

Most people who lose Medi-Cal lose it to incomplete paperwork, not to income.

We have seen this before: the state’s own unwinding data showed two in three enrollees who lost Medi-Cal were dropped for procedural reasons, not because they stopped qualifying. That is the shape of the loss the new projections assume. When the estimates count millions, they are counting the renewal, not the eligibility line. The friction isn't a side effect of the coverage loss — it is how the coverage loss happens.

In one Central Coast county, roughly 50,000 residents sit in the new adult group. Historically about one in four are discontinued at renewal because they didn’t return the packet. Not because they earned too much. Because a form didn’t come back. That is coverage loss in the thousands, on administration alone.

In one Central Valley county, staff project around 40,000 members will face work requirements, and the auto-renewal data tells the same story. Renewals clear automatically for about 60% of English-speaking caseloads and only 30 to 40% of Spanish-speaking ones. That gap isn’t about eligibility. It’s about documentation and trust: different phone numbers, different verification hurdles, the same underlying eligibility.

The income test confirms the point. Eighty hours a month at the federal minimum wage works out to about $580. At California’s $16.90 minimum wage, that clears in roughly 34 hours, well under half the 80-hour activity threshold. The people most at risk next year overwhelmingly meet the standard. They just may not prove it in time.

The rule’s one genuinely new move makes the same point in federal regulation. It narrows the medical frailty exclusion so a diagnosis no longer protects you on its own. Now you have to document that your condition “significantly impairs” your ability to work, often with a provider’s signature. People who plainly qualify become people who have to prove it, with a form many can’t easily get. The barrier isn’t the condition. It’s the paperwork that proves the condition.

Here is the possibilist turn, and the part I most want you to sit with: if the problem is administrative, the solution is operational, and we already run it.

The net already exists in the field. If redetermination is the high wire, the net is already strung beneath it. We don’t have to weave it. We have to widen it.

Walk through what already works. Ex parte renewal verifies eligibility against data the system already holds, before anyone lifts a finger, and keeps members covered with no packet at all. Pre-filled renewal packets ask members to confirm rather than reconstruct. Community organizations can hold BenefitsCal accounts and complete applications alongside the people they serve. One plan has put operational reserves behind retention grants that fund both awareness and hands-on help with warm handoffs. In one county, an eligibility specialist works bedside in the emergency department several days a week, turning a hospital visit into a finished application. County staff sit co-located at the plan, at public health, at the health centers, meeting people where they already are.

None of that is theoretical. It is live today in the counties I sat in this week. It is not yet at the scale we’ll need when the wire goes up in January.

There's one more thing worth naming, because it's the objection a careful reader will raise: income renewals are one thing, but verifying eighty hours of work is new. True — and the rule itself answers it. States are directed to confirm work status ex parte first, drawing on payroll, claims, and encounter data before asking a member for a single document. For many people, the eighty-hour question can be answered by data the system already holds. The infrastructure we built to confirm income is the infrastructure the work requirement calls for — same plumbing, one more field.

That’s the work, and it’s doable inside the time we have.

Start with the contact problem, because nothing else works without it. The biggest barrier to reaching members isn’t apathy. It’s disconnected phones: government-issued cells lapse month to month, voicemail goes unset, numbers churn. Before we optimize the message, we have to be able to deliver it. Re-contact workflows, address-of-record practices for unhoused members, and multiple channels (text, portal, in-person) are the unglamorous foundation everything else rests on.

Then get the renewal calendar into trusted hands. Counties can share 60-day renewal-date lists. The bottleneck is account authorization, not willingness. Resolve who can hold a BenefitsCal account, and clear it now, so community organizations can reach members in the 60 days that matter most.

Standardize the message and route it through the people members already trust. Complete the packet on time. Report changes when they happen, not at renewal. Use the portal. That message lands when it comes from a community organization a family already knows. Counties themselves name a trust deficit, and name community partners as the messengers who close it. Technology enables. Relationships transform. Ex parte and BenefitsCal are the technology. The trusted messenger is the relationship that turns a pre-filled packet into a signed renewal.

And use the comment window. Comments on the interim final rule are open through July 31, enough runway to organize a real response. A comment that names administrative burden as the real risk, and ex parte expansion, six-month renewal supports, and language access as the fixes, is grounded in what the field is seeing. We aren’t speculating. We’re reporting.

The larger opportunity is to make the net a pattern instead of a county accident. Shared Enhanced Care Management (ECM) and Community Supports (CS) language across counties and plans, so members stop getting lost in program names. A shared eligibility-checker so no one has to guess. Monthly office hours through December to keep community partners current. And the conveners who hold it together, the neutral tables where counties, plans, and community organizations coordinate, moving toward community ownership with structured backing so the work outlasts any single contract. What works in a few counties becomes a method a dozen others can lift.

This is orchestration over digitization. The wire is real, but it’s a coordination problem, and coordination is something California already knows how to do. The pieces are on the table: the data automation, the trusted messengers, the funding models, the counties willing to share their renewal calendars. The work between now and January is to connect them at scale, on purpose, before people fall rather than after.

If there is one move to make before the fall renewals begin, make it this: clear BenefitsCal account authorization for the community organizations your members already trust. It's the lowest-cost, highest-leverage step on the board — it turns the sixty days before a renewal into sixty days of help from someone a family already knows. The bottleneck was never willingness. It was permission. Grant it now, and the rest of the net has something to hang from.

We already know who slips. We already know what catches them. The work now is making the difference between a fall and a landing: getting the net strung wide before the wire goes up. Energy and persistence are what close that gap. Let’s go widen the net.

Jim

CalAIM Field Notes is a working notebook for the possibilists inside CalAIM. If a colleague on a county, plan, or community team would find this useful, send it their way.

Cover: Enrollment-to-Retention (Hickman Strategies). Photo: Deyan Georgiev / Shutterstock (editorial use) Infographic: HT4M (Health Tech 4 Medicaid).

CalAIM Field Notes · Special Edition SE 09 · June 2, 2026

On June 1, CMS released the interim final rule implementing Medicaid community engagement requirements under H.R. 1. Comments are open through July 31. States must implement by January 1, 2027, with an additional year, through January 1, 2028, to stand up certain verification processes.

The headlines will be about compliance, reporting, exemptions, and administrative burden. The Congressional Budget Office projects this is the single largest coverage-loss provision in H.R. 1, as many as 5.2 million people nationwide.

Those numbers matter.

The work requirement did not create a new population. It revealed one California has already been serving through Medi-Cal, without yet identifying it, documenting it, or connecting it to the disability-linked benefits and coverage stability many already qualify for.

What the Rule Actually Does

Applicable adults in the Medicaid expansion group must show 80 hours a month of qualifying activity: work, community service, a recognized work or training program, at least half-time education, or a combination. They can also qualify through monthly income equal to 80 hours at the federal minimum wage, which the rule sets at $580.

Around that requirement, CMS built two protective structures. Nine categories of specified excluded individuals are never subject to the requirement. A set of mandatory and short-term exceptions covers people who are.

The nine exclusions include former foster youth under 26, American Indians, parents and caregivers of children 13 and under or of disabled individuals, totally disabled veterans, people who are pregnant or postpartum, people already meeting TANF work rules or subject to SNAP work rules, people in addiction treatment, and people in public institutions. One category should command CalAIM’s attention: individuals who are medically frail or have special medical needs.

CMS built that exclusion around five categories: blindness or disability under Social Security standards, substance use disorder, a disabling mental disorder, a physical or developmental disability that impairs activities of daily living, and a serious or complex medical condition.

The qualifying test is function, not diagnosis. A condition counts only when it significantly impairs a person’s ability to meet the requirement. Someone able to complete 80 hours a month is not medically frail under this rule, even with a qualifying diagnosis.

Can the person work? Can they sustain daily activities? How severe is the impairment, and how long will it last?

Those are disability-determination questions. For one of the five categories, the rule goes past resemblance and adopts the Social Security standard directly, citing Social Security Act section 1614, the definition that governs SSI.

The rule’s separate caregiver exclusion takes a different path, using the broader Americans with Disabilities Act standard instead of the SSI standard. That difference works in our favor. The exemptable population is wider than the SSI-eligible population, which is exactly why it rewards a closer look.

The Question Underneath the Headlines

So the rule asks the state to find people whose functional limitations keep them from working. Social Security asks a close cousin of that question to decide who qualifies for SSI and SSDI.

Which raises the real question. How many Californians enrolled today through MAGI Medi-Cal would meet the functional bar this rule describes, and may already qualify for SSI or SSDI, yet have never been identified or supported through a disability determination?

That population is unlikely to be small.

For more than a decade, California’s coverage strategy worked. ACA expansion and MAGI Medi-Cal simplified eligibility, widened access, and drove the uninsured rate to historic lows. The system optimized around enrollment, and that was the right goal.

Optimization has a shape. As MAGI became the faster path to coverage, the workflows that once moved people into disability-linked benefits drew less investment. We built a system very good at helping people get covered, and less consistent at recognizing who, within that coverage, might qualify for steadier support.

Across the state, hospitals, health centers, county behavioral health, ECM teams, Community Supports providers, and health plans already serve people with significant functional impairment every day. Many experience homelessness, serious mental illness, or substance use disorders. Many manage several chronic conditions and cycle through emergency departments.

The population already exists. What we have not built is the connective infrastructure to treat disability eligibility as a routine part of care coordination.

Two Platforms Already in the Field

Enhanced Care Management can become one of the most important identification platforms in the state. ECM teams already work with members whose lives show what a disability determination measures: functional limitation, housing instability, behavioral health complexity, barriers to steady work. We have treated ECM mainly as care coordination. It can also do identification and navigation.

The workflow is not far off. ECM engagement leads to functional screening, then an SSI or SSDI referral, then determination support, then coverage stabilization and long-term care coordination. That is a new connection between systems we already run, not a new benefit.

Identification is half the work. The longer half is retention, keeping people connected as renewals, exemption decisions, and eligibility reviews arrive. The rule reinforces this. States must verify compliance at application and renewal, alongside the six-month eligibility checks H.R. 1 now requires, and reach members through notices and ex parte data matches. The questions members will face are mostly not clinical. Are you still eligible? Do you qualify for an exemption? Did you finish your renewal? Did your contact information change?

Those are navigation questions, and navigation runs on trusted relationships.

This is where Community Health Workers become one of the most valuable assets in Medi-Cal. CHWs already help members understand benefits, complete paperwork, answer notices, update contact information, keep appointments, and reconnect after a gap. They often understand a member’s daily reality better than any formal system.

If ECM is the identification platform, CHWs are the engagement platform, the human infrastructure that keeps people connected to Medi-Cal, SSI, SSDI, SNAP, housing, and other county supports.

Where the Money and the Authority Already Sit

A fair question follows. Who pays for this, and under what authority? The honest answer is that most of the capacity is already funded and already authorized. Enhanced Care Management, Community Supports, and Community Health Worker services are covered Medi-Cal benefits, paid through managed care today. The screening and navigation this rule invites ride largely on delivery vehicles already in the capitation base, not on a new appropriation.

The rule itself opens the door. CMS describes managed care plans conducting outreach and education and sharing data on medical status, the exact functions identification and engagement require. For genuinely new activity, such as work-plan referrals, the rule notes that many of these cannot be built into capitation rate development, though they can be recognized in the medical loss ratio as incurred claims. There is an accounting path.

One boundary gives the design its discipline. A plan may not be the entity that determines a member’s compliance where it holds a financial relationship with a plan responsible for covered services. Read the right way, that is a design principle, not a limit. Plans and their community partners identify and engage. The state and the Social Security Administration determine. Each partner has a clean lane.

From Enrollment to Retention

For more than a decade, California’s transformation centered on enrollment infrastructure: make coverage easier to get. That work was necessary, and it succeeded.

The next phase asks for a different build, one that helps people stay connected to coverage, benefits, and services over time. The pieces are already in the field:

• Redeterminations: prevent procedural disenrollment, keep contact information current, support timely renewals.

• Medical frailty and exemption screening: identify members with significant functional impairment and document their status.

• SSI and SSDI navigation: screen for disability eligibility, support applications, open the door to steadier coverage and income.

• Enhanced Care Management: find high-need members early and serve as the entry point for disability navigation.

• Community Health Workers: hold the trusted relationships that carry members through administrative requirements.

• Community Supports: address the housing, food, and behavioral health barriers that interrupt stability.

These are not six programs. They are parts of one retention infrastructure.

The stakes run past a work-requirement exemption. SSI eligibility can mean income support, steadier Medi-Cal, lower renewal risk, and for some, a later path to Medicare. For many Californians with significant disabilities, it is a foothold toward stability across several parts of life at once.

An Orchestration Problem

This is why I read the work-requirements debate as a stand-in for a larger one.

California does not mainly have a coverage problem. It has an orchestration problem. Orchestration over digitization.™

The state has built remarkable assets: coverage programs, care management, behavioral health, housing, Community Supports, Community Health Workers, legal aid, disability programs. What is often missing is the connective tissue between them. The next phase of Medicaid transformation depends less on new programs and more on stronger pathways among the ones already here.

Whether the medical frailty exemption is drawn too broadly or too narrowly will be argued through July 31. The strategic question sits in front of every plan, county, provider, and CBO right now. Can California rebuild disability navigation and coverage retention at a scale that matches the population already in front of us?

What We Build Next

The first chapter of California’s Medicaid transformation was about helping people get in the door. The next chapter is about making sure they do not fall back out.

We have the assets. We have the workforce. In CalAIM, we have the policy architecture. What we build next is the orchestration layer that turns nine exclusion categories and a renewal calendar into a coordinated system of identification, navigation, and retention, one that connects a member to every benefit, service, and coverage pathway they already qualify for.

That is partnership work. It belongs to plans, counties, providers, CBOs, and the disability and legal-aid communities together, with the people we serve at the center.

The population already exists. The work is to build infrastructure worthy of it.

Technology Enables · Relationships Transform.™

The interim final rule is open for public comment through 5:00 p.m. ET on July 31, 2026. The choices each state makes, how it screens for medical frailty, how it runs ex parte verification, how it designs renewals to limit procedural disenrollment, will decide whether this rule becomes a coverage cliff or a coordination catalyst. California's voice in that record matters.

A health system partner pinged us this week with the question coming up across operator calls: Is ECM getting cut?

The headline read is understandable. The policy read is the opposite.

Enhanced Care Management is being refined, not retired. The May Revision released May 14 proposes changes to eligibility, service definitions, utilization management, and payment for ECM and Community Supports effective January 1, 2027. The fiscal line is real. The Department of Health Care Services projects a General Fund reduction of $41.4 million in 2026-27 for ECM, growing to $99.2 million ongoing. Community Supports tightens in parallel: a $26.9 million General Fund reduction in 2026-27, $51.0 million ongoing. The design intent behind the dollars is fidelity to the high-touch, person-centered model the benefit was originally built for.

What is actually changing inside ECM

The Department is driving fidelity to the original service model. Three areas worth holding clearly:

• Service intensity. Refined utilization management criteria and payment adjustments target low-touch delivery (predominantly remote, generalized, or non-individualized care management). Capitated rates will adjust for low-intensity utilization, and risk corridors tighten. Plans we work with are translating this into operator-level floors well above current averages.

• Tighter graduation and duration. Eligibility and duration criteria are sharpening to end extended periods of low-intensity service.

• Reinforced non-duplication. The standing ECM exclusions for Behavioral Health Targeted Case Management, High-Fidelity Wraparound, and 1915(i) waiver care management remain the floor. Plans are sharpening enforcement so members are not enrolled in concurrent care management that the program was not designed to support.

A Medi-Cal plan recently walked operators through what this means on the ground. The Department’s intended footprint for ECM has always been a focused share of membership. Current enrollment in this plan’s region sits well above that share, with average encounters per member tracking below the underlying outcomes evidence for the benefit. The orientation is consistent across the plan briefings we have read: high acuity, high touch, in person where possible, with documented results.

We read it the way they read it. This is a protection of the program, not a contraction of opportunity. Back to basics, in the most operationally useful sense of the phrase.

What is also moving on January 1, 2027

Two changes worth holding alongside the ECM refinement.

• Community Supports tightens in parallel. Refined referral pathways, eligibility criteria, service definitions, and utilization management criteria for selected services. We will go deeper in Field Notes No. 19.

• A separate federally-driven change moves approximately 2 million state-funded immigrant Californians from managed care to fee-for-service on the same date. Because ECM and Community Supports are managed care benefits, this population loses access to both. The Department projects a $583.8 million reduction ($471.6 million General Fund) in 2026-27 and $1.5 billion ($1.2 billion General Fund) ongoing. For safety-net providers and hospitals serving large undocumented populations, this is the largest operational change in the budget. It is distinct from the refinement, and worth naming as such.

The operator work between now and January

For CBOs and providers reading this, the summer work is articulation, not access defense.

• Audit current encounter cadence against high-touch standards. Where the data shows broad but shallow, the work between now and 1/1/27 is rebuilding to the model the benefit was designed for.

• Reconcile overlap with Behavioral Health Targeted Case Management, High-Fidelity Wraparound, or 1915(i) waiver care management. The standing exclusion is being enforced. Organizations that wait until January will reconcile in arrears.

The Readiness Architecture we use with partners (Trust before Capabilities, Capabilities before Infrastructure, Infrastructure before Financing) was diagnostic when we built it. The 1/1/27 refinement makes it predictive. Counties that hardwired the lower layers absorb tighter eligibility frames gracefully. Organizations starting from Financing read refinement as cut.

Scale to Partnership

Counties that orchestrate well across plans, providers, and CBOs will find the January 1 transition does what good policy should: protect the program from drift. Technology enables. Relationships transform. Orchestration over digitization.

The constraint is not a limitation. It is the condition under which the next generation of CalAIM capacity emerges. The window between now and 1/1/27 is when the orchestration work pays off.

If you are reading the refinement from inside your county (plan, provider, CBO, or county orchestrator), we would value hearing your read. Reply and tell us what you are seeing on the ground.

Jim

A one-page field brief titled The May Revise and CalAIM: What Changes on January 1, 2027 is available for partners. Reply to this email and we will send it.

Primary sources: Department of Health Care Services 2026-27 May Revision Highlights, May 14, 2026 · DHCS Refinements and Efficiencies for ECM and Community Supports Fact Sheet, May 2026 · 2026-27 May Revision, Office of Governor Gavin Newsom, May 14, 2026 · First Look: Understanding the Governor’s 2026-27 May Revision, California Budget and Policy Center · Analysis of Governor Newsom’s May Revision, Western Center on Law and Poverty · 2026-27 Budget: Medi-Cal Analysis, Legislative Analyst’s Office.

Field Notes is a weekly dispatch from the work of orchestrating CalAIM. Published Thursdays on Substack.

CalAIM Field Notes · No. 17 · Thursday, May 7, 2026

Late April and early May produced three California convenings worth holding side by side. Different conveners. Different audiences. Different agendas. One shared through-line.

On Wednesday, April 22, the California Budget & Policy Center held its annual Policy Insights conference at the SAFE Credit Union Convention Center in Sacramento. Several hundred people — community advocates, policymakers, philanthropy, public-sector leaders — working through what California’s fiscal architecture has to do under federal pressure.

On Thursday, April 23, one block away at the Kimpton Sawyer, the California Quality Collaborative and Purchaser Business Group on Health convened Catalyst for Health 2026. Health plans, purchasers, delivery systems, county leaders. ECM and Community Supports threaded through nearly every session.

On Thursday, May 1, at the UC Berkeley School of Public Health, The SCAN Foundation convened Bridging Capital & Care — a working session on aligning California’s institutional capital with its community-care infrastructure.

Three rooms. Nine days. One gap, named in three vocabularies — and none of them fully assembled the answer.

The gap each room named

Capital signals are in place. CalPERS adopted its Total Portfolio Approach on November 17, 2025, with implementation effective July 1, 2026 — the first U.S. public pension fund to do so. DHCS APL 25-004 requires every Managed Care Plan with positive net income to file a Community Reinvestment Plan deploying 5 to 7.5 percent of net income into upstream community health investments, with first plan filings due in Q3 2026. These are scheduled events with regulatory force.

Care infrastructure is in operation. Enhanced Care Management is in its third operational year. Community Supports has 14 approved service categories, with adoption now spanning every county in the state. The CBO ecosystem delivering both has been built through PATH CPI collaboratives, county Behavioral Health Services Act planning, and four years of cross-sector relationship work most of which never appears on a contract.

Orchestration is the gap each room named in different language. Policy Insights named it as fiscal architecture under federal threat. Catalyst for Health named it as the purchaser-and-delivery alignment problem. Bridging Capital & Care named it as the capital-to-care translation question. One gap. Three vocabularies.

The question I named at Berkeley was the simplest version: California built the policy. Who builds the infrastructure? The response across the rooms: Who Pays

?

What I brought into the Berkeley room

The framework I shared in the lightning round comes from four years of PATH CPI collaboratives, ECM rollouts, Community Supports launches, and the slow-motion redetermination crisis. It names four layers of CBO readiness. Three are operational and visible. The fourth is foundational and almost always treated as background.

• Layer 1 · Capabilities and maturity. Workforce, training, supervision, the practice of doing the work.

• Layer 2 · Infrastructure. Data systems, billing, contracting, the operational scaffolding that lets capabilities scale.

• Layer 3 · Financing. Rate adequacy, working capital, sustainable margin: the conditions under which Layers 1 and 2 can be built and held.

• Layer 0 · Trust and relational capital. The pre-condition. The asset CBOs already hold

The four layers form a loop. No layer advances in isolation, and Layer 0 is where the orchestration question gets its real answer.

The matrix below — the slide I had on screen at Berkeley — shows what each layer looks like across four maturity stages. The Established column is the Managed Care Plan contracting floor. Most CBOs sit Developing-to-Established. The gap is not failure; it is underfunded potential

The layer the system keeps trying to buy

Managed Care Plans can purchase capabilities — they can hire staff. They can acquire infrastructure — they can license software. They can structure financing — they can advance payment. They cannot purchase 30 years of presence in a community. They cannot buy the credibility a peer support specialist carries because they have lived through the same system the member is now navigating. They cannot buy the network density a CBO has built handling SSI/SSDI redeterminations for unhoused clients since the 1990s.

That asymmetry is the structural argument. It is also why the financing question lands harder than it looks.

This is the line I drew in No. 08 last December — that 2025 was the year trust became infrastructure. Five months on, the question has sharpened.

Trust became infrastructure. The system has not yet learned to finance it that way.

That is the line I keep returning to across all three convenings. Capital signals are real. Care infrastructure is real. The translation layer between them — the orchestration that makes Layer 0 deliverable at scale — is the part each room circled and none fully assembled. Policy Insights asked how the fiscal architecture survives federal pressure. Catalyst asked how purchasers and delivery systems align around outcomes. Bridging Capital & Care asked how institutional capital reaches the community-care layer where the work actually happens.

All three are versions of the same question: how does the system pay for the trust it depends on?

What this means for the next two quarters

Three things are now in motion at once. Each one names the orchestration layer as the missing piece.

The Q3 2026 MCP Community Reinvestment cycle. First plans land this summer under DHCS APL 25-004. Whether backbone coordination infrastructure is named in those plans — by name, with a budget — is the question that decides whether Layer 0 gets financed or absorbed.

CalPERS Total Portfolio Approach implementation, July 1, 2026. California’s largest pool of public capital begins operating under a unified framework that explicitly considers the long-term system-level conditions in which its beneficiaries live and retire. The mechanism is real. Whether community-care infrastructure becomes a recognized portfolio object is the open design question.

The H.R. 1 floor. What Field Notes No. 15 named as the floor — the federal coverage threat that compresses every state response — is the pressure environment in which California’s orchestration architecture has to hold. CHCF and Health Management Associates published The New Uninsured: State Policy Options for Californians Losing Medi-Cal Coverage in April, citing DHCS estimates that up to 1.8 million Californians could lose Medi-Cal coverage in the coming years and modeling a state-funded bridge program at $3.1 to $4.6 billion annually. That bridge program is the coverage layer; Layer 0 is what delivers it, and the two will not be financed by the same line item. Trust infrastructure is what keeps members connected to coverage when the system underneath them shifts. It is the most durable asset the state has, and the least financed.

These three threads converge in the same place. The convergence is the moment.

A note on what is not in this issue

This issue is the spine, not the body.

• The convening synthesis itself. Each room produced its own intelligence and its own next steps. Sarita Mohanty and the SCAN Foundation team are producing the Bridging Capital & Care synthesis on their own timeline, and the credit for that working session belongs there. CBPC and CQC/PBGH each have their own continuing work streams. This piece holds the through-line, not the room-by-room report.

• Named participants and direct quotes. The convenings were working sessions. The voices in those rooms belong to the people in them.

• The full architecture of the framework. The four-layer model, the maturity matrix, the County Orchestrator two-pillar architecture, the dual-capital framing, the H.R. 1 application — these belong to a forthcoming series of pieces, academic and field-facing, sequenced to publish across 2026 in or before the Q3 capital convergence window. What I rolled out at Berkeley was the spine of that series in three minutes. What I have shared here is the spine in a longer form. The body lands in the pieces themselves. Field Notes subscribers will see each one announced here first.

  ---

  Two questions for the next conversation

  If you are sitting at a Q3 reinvestment planning table — at a Managed Care Plan, a county, a CBO consortium, or as a philanthropic partner — two questions are worth holding open:

  1. Is backbone coordination infrastructure named in your Community Reinvestment plan, with a line item — or is it absorbed into program budgets?

  2. Who at your table holds the trust that the rest of the architecture depends on, and what is your plan to finance that role across the H.R. 1 transition?

  The second question is the harder one. It is also the one each room kept returning to.

  ---

  In the weeks ahead: the Q3 2026 reinvestment cycle in operational detail — what backbone language is showing up in the plans we can see, and what to ask for in the ones we can shape.

  If you missed No. 08 last December — 2025: The Year Trust Became Infrastructure — that’s where this line came from. Read it here.

  If this piece named something you are working through, replies to this issue come straight to me. The conversation is the point.

  See you in the field.

  — Jim

  Jim Hickman · Hickman Strategies LLC · Oakland, California

  Technology enables. Relationships transform. Orchestration over digitization.

Oracle Park, April 15, 2026. The right venue for the right conversation.

The Room Where It Happens

Gateway is a curated two-day convening where healthcare executives meet the innovators building the next generation of solutions, and both sides come prepared to work. Relationships form. Deals start. This year a documentary crew was there to capture it, because showing what happens in this ecosystem is faster than describing it.

The energy this week was different. LPs and founders had flown in specifically for the convening. The Suncoast Ventures AGM ran alongside Gateway, adding a capital layer to the clinical conversations already in the room. Dinners, side sessions, skee ball.

These are the problems of our generation: health equity, access, the structural failures built into how we finance and deliver care for the people who need it most. The people in that room are building the solutions.

Mark Gannott · Mark Rangell · Robert Merold · Chima Osuka · Simran Nath · Craig Hauben · Jim Hickman · Tamy Mahaney · Genevieve LeMarchal

When Steve Miff took the stage on April 14, the room already understood what was at stake. What he gave them was the data.

The Map Upside Down

Steve Miff opened with a map of Dallas County. Fourteen years of population health modeling. 34.9 million predictions across 6.2 million patients. Zip codes color-coded “low vulnerability.” Safe-looking. Stable.

Then he zoomed in.

Inside those same zip codes: more than 100,000 people living in census block groups flagged as high-vulnerability. People who would never have been identified, never reached, never counted under the zip code frame.*

We have been drawing the map upside down. Steve Miff, PCCI, Gateway II 2026

---

The Resolution Problem

CalAIM’s core bet: wrap the right services around the right people at the right time, and bend the curve on cost, health, and equity. ECM, Community Supports, In Lieu of Services, the 1115 waiver. All of it depends on finding those people first.

Zip codes cannot do that. They are administrative boundaries that flatten neighborhood texture. The block where three generations of the same family have lived sits two blocks from the block where nobody has a primary care relationship and the nearest pharmacy is a 40-minute bus ride. A zip code calls both the same.

PCCI’s Community Vulnerability Compass plots 26 indicators across four thematic domains at census block group level, then proxy-tags those determinants to individual patient records via current address. That last part matters: when self-reported intake data is a year old and covers less than a third of a population, address-based geocoding is not a supplement. It is the more accurate record. PCCI published the validation methodology in JAMIA against 160,000 Z-codes.

PCCI's Community Vulnerability Compass at census block group resolution. The "Block Group" tab reveals vulnerability pockets invisible at zip code level. Credit: Parkland Center for Clinical Innovation (PCCI). Used with permission. *100,000+ figure: from session notes; verify against JAMIA publication before publishing.

Want to see this in action? PCCI is hosting a free live demo of the Community Vulnerability Compass on April 28, 12–1pm PDT. Register and share across your networks.

California's data infrastructure is already in motion. Public entities like CalMHSA, health information organizations like Manifest Medex, and private networks like PointClickCare are moving significant data volumes across their systems today. PointClickCare — which holds approximately 80 percent of SNF data in California — has built exactly the kind of longitudinal, facility-level data layer the state's care coordination ambitions depend on. The question is whether the state-level infrastructure being built around these networks operates at the same resolution — or whether it flattens the data back to zip code before anyone acts on it.

What “It Works” Actually Looks Like

The Accountable Health Communities program ran across 17 health systems and more than 100 communities in North Texas. CHW navigation was measured against 724 matched digital twin controls. Twelve months of post-navigation data:

CMS accepted the digital twin methodology as legitimate outcome measurement. The results were published in the New England Journal of Medicine. These are not modeled projections.

The pediatric asthma program ran seven years, deployed across health plans, health systems, and FQHCs, with sustained impact from 2016 through 2023. No app. No Alexa integration. Opt-out SMS, because that is what worked in a mostly Medicaid population. 91.9% of participants said they would recommend the program to friends and family. Technology enables where relationships already exist.

For anyone making the case for CalAIM infrastructure investment, these are the numbers that close arguments.

The Direction We’ve Had Backwards

Miff put it plainly: the right model runs community-to-healthcare, not healthcare-to-community.

“The right model runs community-to-healthcare, not healthcare-to-community.” Steve Miff, PCCI, Gateway II 2026

Everything we’ve built assumes the health system is the hub. Referrals flow outward. Technology gets designed for clinicians, then handed to CBOs. Closed-loop data gets measured from the discharge summary backward.

PCCI learned this by running AHC. CHWs embedded in individual institutions get pulled in competing directions and cannot scale. What worked was centralized management: daily targets, learning huddles, performance scorecards, staffing flexed dynamically to known caseload capacity. People accountable to a model, not to an institution’s competing priorities.

Keith Kosel and Steve Miff documented exactly this in Building Connected Communities of Care (HIMSS/CRC Press), which has been on my desk since it published. Gateway II confirmed the framework holds at scale, with CMS-accepted methodology. California’s ECM lead entities are being asked to do what this book describes. The technology they’re being handed still runs through health plan portals and EHR workflows that most CBOs cannot easily access. Build for the community organization first, and the connection back to providers follows.

AHC Lessons Learned, slide 13 from Steve Miff’s Gateway II presentation. Building Connected Communities of Care (Kosel & Miff, HIMSS/CRC Press): the operational framework behind PCCI’s results. Required reading for every ECM lead entity in California. Credit: PCCI.

Orchestration Is Not Optional

PCCI runs 19 AI models in full production. Its radiology incidental findings system runs five large language models simultaneously: the majority must agree before anything gets flagged. Against 406,000 annual scans at Parkland, it identifies approximately 23,000 patients, a 6.6x increase over manual review, at 98% accuracy versus 94% for human review. It won the 2024 John M. Eisenberg Patient Safety and Quality Award.

The trust framework behind all of it has four components: prediction transparency (what is driving a score, not just the score itself), performance transparency (continuous monitoring for model drift), security (on-premise deployment to keep data within health system firewalls), and compliance (regulatory requirements now requiring monthly review). PCCI built a monitoring system that flags drift statistically so the team investigates only when warranted.

No single plan or health system should have to build this infrastructure alone. The open question is who governs the coordinating layer, and how California’s Data Exchange Framework enforces shared definitions before agent deployment compounds the inconsistencies already in the data. Orchestration over digitization: shared standards have to precede scaled deployment. That sequencing is not academic. It determines whether any of this reaches the people it was built for.

The Equity Fault Line

In Q&A, a rural clinic techinal assistance provider made the case against. Her staff earn $40-45K a year. They are not hiring engineers. They are afraid of AI. And her clinic delivers better culturally competent care than most of the well-resourced systems racing to deploy these tools.

The concern is legitimate: AI infrastructure gets adopted fastest by organizations with the most existing capacity. The gap widens. Safety net providers fall further behind the very systems that are supposed to be learning from them.

The response is not to slow deployment. It is to design for leverage from the start. One care manager handling 300 patients who can now handle 1,000 with AI-assisted outreach is not a replacement. That is the point. PCCI’s asthma program dropped the app and the Alexa integration and built around opt-out SMS, because that is what worked in a Medicaid population. Technology enables where the relationship already exists.

California’s CalAIM implementation gets judged on whether it reached the people it was built for. That requires data at census block resolution, tools designed for CBOs rather than health plan portals, and orchestration that amplifies the care manager already in the community doing the work. The frameworks are in place. The legislation exists. The infrastructure is being built. Whether we build it at the right resolution, in the right direction, for the right people first is still being decided.

Jim Hickman is the founder of Hickman Strategies LLC. He works at the intersection of healthcare policy, innovation, and system transformation — advising health plans, entrepreneurs, and community organizations on CalAIM implementation, complex care, and scaling solutions in Medicaid and other public-sector systems. He is a senior advisor to the Camden Coalition’s National Center for Complex Health and Social Needs, a Venture Partner at Suncoast Ventures, and an advisor to Health Inno New Mexico. His organizing belief: technology enables, but relationships transform.

CalAIM Field Notes is a biweekly dispatch on what’s working — and what’s stuck — in California’s Medicaid transformation.

Disclosure: PointClickCare is a current client of Hickman Strategies LLC.

In No. 13, I wrote about a number that stayed with me: up to 50% of the 2.8 million Californians in MAGI Medi-Cal expansion could lose coverage not from ineligibility, but from paperwork burden. A paperwork problem. Paperwork problems have solutions.

That argument was right.

The California Budget & Policy Center estimates up to 3.4 million Californians could lose Medi-Cal under H.R. 1. Health Access puts the total at 5 million when federal cuts combine with the state’s own 2025-26 budget. The UC Berkeley Labor Center published county-level projections in March showing where those losses fall. They are not evenly distributed. They fall hardest on the communities Medi-Cal expansion was designed to reach.

This piece asks the question No. 13 didn’t.

After we build the best possible administrative response, after every trusted messenger is deployed and every yellow envelope handled, how many people still lose coverage? Not because the system failed them. Because the policy was designed to remove them.

Coverage loss is the designed federal policy outcome.

Not for everyone. Not even for most people at risk. But for a meaningful portion of the 2.8 million, this is not a documentation failure or a navigation gap. It is what H.R. 1 Section 71119 was written to produce. Naming that clearly is not defeat. It’s what we owe the field.

What MAGI is and why it matters here

Before the Affordable Care Act, Medicaid required a categorical reason to qualify. You had to be elderly, blind, disabled, pregnant, or the parent of a dependent child. Disability required a formal clinical determination.

The Affordable Care Act (ACA) created a new category: Modified Adjusted Gross Income (MAGI) expansion. MAGI Medi-Cal has one test: income. Earn at or below 138% of the federal poverty level, roughly $20,120 per year for an individual in 2026, and you qualify. No disability determination. No asset test. The application became simpler and faster than anything before it.

Millions enrolled who had never had coverage. Among them: millions of people with serious disabilities who chose the MAGI pathway because the old disability process was prohibitively hard and the new one was not. They qualified for disability coverage. They just didn’t go through the disability door.

The 2.8 million adults in California’s MAGI new adult group, aid code M1, ages 19 to 64, have no disability determination on record. The system knows their income. Beginning January 1, 2027, income alone is no longer enough.

MAGI’s simplification was its genius. Under H.R. 1, it became the point of maximum vulnerability.

The four paths

Every MAGI-enrolled Californian facing work requirements will end up in one of four places. Three are navigable. One is the floor.

Path A: has a non-disability exemption, needs documentation

Pregnant women, parents of children under 14, people in substance use treatment, survivors of domestic violence, students in workforce programs. The exemption exists. The coverage is protected, but only if it’s found, recorded, and resubmitted every six months.

Cathryn Couch’s California Community-Based Organization (CBO) Alliance is the right backbone for this work: 700 organizations in all 58 counties, already in members’ lives 40 to 50 times a week. Family Resource Centers hold the direct member relationships. F5 county commissions, funded through Prop 10 in all 58 counties, are the investment and coordination infrastructure that connects resources to delivery organizations. These are distinct kinds of organizations. F5 commissions fund and coordinate; Family Resource Centers deliver. Both are needed. Neither substitutes for the other.

Full Circle Health Network and the California Accountable Communities for Health Initiative (CACHI)-led Medi-Cal Redetermination Coalition are building the documentation workflows and provider training that make Path A survivable at scale.

Path B: has a qualifying disability, enrolled under MAGI, transition possible

With a trained navigator, they could move to an H.R. 1-exempt disability-linked category before enforcement. Partners in Care Foundation, built by June Simmons over decades, shows what fully resourced Supplemental Security Income and Social Security Disability Insurance (SSI/SSDI) navigation looks like: dedicated navigators, legal aid for denials and appeals, health plan integration, a financial model built to last. That model exists. The SOAR (SSI/SSDI Outreach, Access, and Recovery) TA Center that trained navigators nationally closed in August 2025. Path B is a rebuild and replace project. It has to happen before January 2027.

How large is Path B? The disability-qualifying MAGI subset is estimated at 250,000 to 450,000 Californians, a working proxy range because no public figure exists. SOAR-assisted SSI/SSDI approval rates run 65%. Unassisted: 38%. That 27-point gap is the value of what was lost.

Path C: can meet work requirement, documentation is the barrier

The person working 80 hours a month cleaning houses, doing day labor, working gig platforms, whose hours are real but undocumentable to CalHEERS. Or the person who documents one six-month cycle, misses the next, and loses coverage not because they stopped working but because the redetermination clock caught them in a gap.

The Full Circle and CACHI coalition is building the infrastructure for Path C. The binding constraint is the county eligibility worker: the only person who can complete enrollment in CalSAWS. That workforce takes about a year to train fully, turns over at high rates, and now processes redeterminations twice a year. The County Welfare Directors Association estimates California needs more than 2,000 additional eligibility workers to manage the H.R. 1 workload. The Governor’s January 2026 budget proposal does not fund those positions. The administration has indicated the question may be revisited at the May Revision.

Path D: the floor

No exemption applies. The disability is real but doesn’t meet the Social Security Administration (SSA)’s standard: unable to engage in any substantial gainful activity for 12 consecutive months. The work hours are real but undocumentable. The caregiving is genuine but the person being cared for is an adult, and H.R. 1 protects parents of children under 14, not children caring for aging parents.

For Path D, coverage loss is the designed federal policy outcome. Not a documentation failure. The designed outcome.

What a county training deck tells us

In January 2026, a county health system and its managed care partner published a 50-slide redetermination training for providers and community partners. It is well constructed, operationally thorough, and reflects exactly what responsible organizations do in a fast-moving policy environment: they prepared their provider network with the best available guidance.

Here is the challenge they were navigating. When that training was published, federal H.R. 1 implementation rules had not been issued. The U.S. Department of Health and Human Services (HHS) interim final rule arrives June 1, 2026, five months after the training was written. The California Department of Health Care Services (DHCS) H.R. 1 Implementation Plan was published January 29, just days before. The Georgetown Center for Children and Families confirmed as of February 1 that the Centers for Medicare & Medicaid Services (CMS) had issued no system modification guidance. The policy framework governing SSI/SSDI transitions had not been written.

You cannot build SSI/SSDI transition guidance into a training when the policy governing those transitions has not been issued. That is not an organizational gap. That is the policy cascade arriving at the delivery level.

Federal guidance comes first. State guidance follows. County agencies implement what they have been given. This training reflects exactly where California is in that sequence: DHCS has built strong MAGI infrastructure. SSI/SSDI transition infrastructure is not yet in the build. That is not a failure of the county. It is the status of the build.

Across all 50 slides: zero mentions of SSI or SSDI pathways. Zero references to the five disability-linked Medi-Cal categories that are H.R. 1 exempt. No guidance on MAGI-to-disability transitions. The disability determination clock is documented at 90 days versus the standard 45-day MAGI clock, correctly documented, without flagging it as a coverage risk. The rules that would define it as one had not been written.

What the training shows is that county infrastructure is MAGI-oriented because Medi-Cal policy has been MAGI-oriented. When federal guidance lands and DHCS issues its navigation framework, the organizations that produced this training will update it. The infrastructure exists. The relationships are in place. The capacity to move is there.

But the counties that respond fastest do not wait for the final guidance packet. They are the ones who started the internal conversation before it arrived. Who asked what their eligibility workers would need to know. Who mapped their CBO navigator capacity before the referral volume came. Who called the table together before the call to action.

June 1 is not the start of that conversation. It is the deadline for having started it.

If you work with or in a county agency, think ahead with them now. Not about what the federal rules will say. We will all read those together in June. About what it would take to be ready when they arrive. Who needs to be in the room before the guidance lands. What your eligibility workers already know about SSI/SSDI pathways. Whether your CBO partners have navigator capacity for Path B members. Whether your provider training can be updated inside 30 days if it needs to be.

The organizations that built that 50-slide training can build the next one. The question is whether the table is set to do it quickly when the time comes.

The gap the training reveals is not organizational failure. The guidance that would close it has not yet been issued.

The choke points nobody is solving

Paths A, B, and C share structural bottlenecks the field is not yet organized around.

The CalHEERS routing gap

Every Medi-Cal application, including from members with serious disabilities, enters CalHEERS first and is processed through MAGI income logic. Getting to a disability pathway requires a county eligibility worker to initiate a manual CalSAWS override. There is no automated disability routing flag. DHCS is rebuilding CalHEERS for H.R. 1 compliance and the build cycle closes January 2027. Adding a routing flag costs less now than after the build closes. Source: Disability Rights California and CalHEERS RFI No. 73050004 (May 2024).

The eligibility worker ceiling

The county Health and Human Services Agency (HHSA) is the only entity legally able to complete enrollment in CalSAWS. CBOs, federally qualified health centers (FQHCs), Family Resource Centers, F5 county commissions, and Enhanced Care Management (ECM) partners can assist, counsel, and prepare. They cannot enroll. Every navigation pipeline terminates at a county eligibility worker desk, a desk that is already 2,000 workers short statewide.

The six-month documentation trap

Path A exemptions are not a one-time protection. They must be re-documented every six months. One missed cycle equals coverage loss. For people whose circumstances are volatile, moving, changing providers, leaving a shelter, the cycle is not a procedure. It is a trap.

The documentation system is not finished

Federal implementation guidance arrives June 1, 2026, seven months before enforcement. The work verification infrastructure is not yet operational. Some counties are already seeing MAGI members return to paper-based enrollment, a signal that digital-first assumptions in the build are running ahead of the actual population.

The four bottlenecks above run across all MAGI paths. Path B carries two more that are its own.

The two-clock collision

The MAGI redetermination clock runs 45 days. SSA initial disability determinations average 6 to 8 months nationally, and longer in high-volume states. There is no bridge coverage for the gap between them. A Path B member who loses MAGI coverage while their disability application is still pending has no coverage until the determination lands. No provisional protection exists for applications in process. The two clocks do not connect.

The legal aid pipeline

Sixty-two percent of unassisted SSI/SSDI initial applications are denied. An appeal requires legal representation. Most applicants cannot access it. California has no funded legal aid pipeline connected to the CalHEERS or CalSAWS ecosystem. SOAR-trained navigators were the closest thing to that infrastructure. The SOAR TA Center closed in August 2025. What remains is a 62% denial rate and no systematic path to contest it.

Who is in Path D

The sub-threshold disability: fibromyalgia, episodic depression, long COVID, chronic pain that impairs work on some days and not others. Real limitation. Doesn’t meet the SSA’s 12-month durability standard. The condition is disabling. It doesn’t qualify.

The adult caregiver: the daughter who left work to care for her mother with advancing dementia. H.R. 1 protects parents of children under 14. It does not protect adult children caring for aging parents, spouses caring for disabled partners, or siblings caring for siblings. Caregiving of adults is not a qualifying work activity. It is invisible to the statute.

The informal worker: cleaning houses, picking crops, doing day labor. Working, often exhaustingly, in the cash economy. The hours are real. CalHEERS cannot verify them. Coverage loss comes not from failure to work but from inability to document work in a form the system accepts.

The person cycling through employment: seasonal, part-time, interrupted by illness or caregiving. Works enough months to qualify. Misses one six-month window. Loses coverage. Re-enrolls. Misses the next one.

None of them are failing. The policy was designed this way.

San Diego and Imperial Counties: Path D made real

The clearest example of what Path D looks like is already unfolding.

A March 4, 2026 briefing circulated among health system partners in San Diego and Imperial Counties named the plain reality: when people lose Medi-Cal and have nowhere else to go, they end up in the hospital. Neither San Diego nor Imperial County has a county hospital.

California Welfare and Institutions Code Section 17000 establishes that counties must serve as the provider of last resort for residents who do not qualify for Medi-Cal. The County Medical Services Program in Imperial County and County Medical Services in San Diego are the legal infrastructure for that obligation. Both have been largely unfunded for years. CMS currently has three people enrolled.

The San Diego County Board of Supervisors voted 4-1 in early March to begin overhauling CMS, directing a comprehensive review through its Ad Hoc Subcommittee on Social Safety Net and Behavioral Health Systems Transformation. One hundred and fifty thousand San Diegans have already left the Medi-Cal rolls in recent years without leaving San Diego. H.R. 1 coverage losses arrive on top of that.

From the March 4, 2026 briefing:

“You know all too well that when an uninsured person needs care, it’s your hospitals where they end up.”

This is not a planning scenario. This is Path D arriving early in a region that is watching it come.

The Orchestrator County

The counties that respond best to H.R. 1 will not be the ones that process redeterminations most efficiently. They will be the ones that coordinate the full ecosystem: the CBO Alliance, Family Resource Centers, F5 county commissions, FQHCs, health plans, ECM providers, and the navigator organizations working Path B, into a single legible response system with the county at the center.

The Orchestrator County model rests on one structural fact: county government is the only entity that can complete enrollment in CalSAWS. Everything else flows from that. Counties that exercise that authority actively multiply the impact of every organization they coordinate. That means convening the full provider network, integrating F5 commissions and Family Resource Centers into outreach workflows, coordinating Path B referral pipelines with ECM providers, and making the board-level case for eligibility worker investment before the coverage losses materialize, not in response to them.

The San Diego Board response is a model: a board-level commitment, an ad hoc subcommittee, an explicit accountability structure, and an honest acknowledgment that the legal obligation has been structurally unfunded for years and is about to matter enormously.

What we can still build

The possibilist’s argument is not that everything is possible. It is that the possible is larger than we are currently building toward, and that naming the limits honestly is how we stay inside them.

For Paths A, B, and C, perhaps 800,000 to 1.2 million people, the window is open. CBO Alliance coordination needs to be funded before October 2026 notices go out, not in response to January 2027 coverage losses. A CalHEERS disability routing flag added during the current build cycle would multiply every navigator in the system. County eligibility worker capacity needs board-level budget protection now. ECM providers need to build Path B navigator capacity as a funded function, not absorb it informally into unreimbursed community supports.

For Path D: document who lands there. Track what happens, the deferred care, the emergency visits, the Section 17000 programs being rebuilt from three enrolled members. Build the evidence base that makes the policy argument to Sacramento and to Congress. Make sure the safety net knows Path D is coming.

The floor is not the ceiling. The floor is where we start building.

What’s the first specific, measurable commitment?

In No. 13, the organizing question was: do we have a specific, measurable plan to close our share of the administrative gap before Q3?

This piece adds a second: do we have a specific, measurable plan to track what happens to the people the administrative response cannot reach?

Both questions need answers before October. Because the people in Path D, the people the policy was designed to leave behind, deserve to be counted. Counting is how we prove, a year from now, that this policy had a floor, and that we saw it, and that we said so.

The work that begins January 2, 2027 has to be organized before January 1.

Coverage loss is the designed outcome. What gets built from this floor is ours to design.

* The 250,000-450,000 disability-qualifying MAGI subset is a working proxy estimate. No public figure exists. DHCS claims data analysis would establish a citable number. The 1.4 million paperwork-failure figure is DHCS’s own projection from the February 5, 2026 webinar.

Sources

– DHCS H.R. 1 Implementation Plan Overview Webinar, February 5, 2026.

– H.R. 1 §71119 (One Big Beautiful Bill Act), signed July 2025.

– California Budget & Policy Center (2025): up to 3.4 million Californians could lose Medi-Cal.

– Health Access, “The State of Health Care Coverage” (2025): up to 5 million losing coverage.

– UC Berkeley Labor Center, county-level projections (March 2026).

– County Welfare Directors Association of California (Chico News & Review, March 23, 2026): 2,000+ additional eligibility workers needed; not funded in Governor’s proposed budget.

– County-level provider redetermination training (January 2026). Reviewed through provider network channels. Source protected.

– Homeless Leadership Coalition / SOAR TA Center closure, August 18, 2025.

– SSA FY2024 Agency Financial Report: 1.18 million initial claims pending nationally.

– SSA FY2024 Workload Data: 38% initial approval rate; 65% SOAR-assisted.

– SSA OIG Audit, July 2025: 222-day California initial decision average.

– Disability Rights California; CalHEERS RFI No. 73050004, May 2024.

– DHCS HR-1 Implementation Plan, January 29, 2026.

– San Diego / Imperial Counties biweekly briefing, March 4, 2026.

– California Welfare and Institutions Code §17000.

– LAO, “Considering Medi-Cal in the Midst of a Changing Fiscal and Policy Landscape,” October 2025.

The Shift Nobody Fully Planned For

We have been building California’s Medicaid transformation on top of a data infrastructure that was never designed for what we are now asking it to do.

CMS requires plans to correct encounter data errors within 15 days. The accuracy threshold is 99.5% of records. Plans that exceed a 2% denial rate face penalties and corrective action.

These are not aspirational targets. They are the operational floor every California health plan must sustain today—across Medicare Advantage, Covered California, and Medi-Cal. For MA, encounter data drives risk adjustment. For Covered California, marketplace rate development. For Medi-Cal, managed care rate setting, program accountability, and the success of Medi-Cal Connect.

The failure modes are patterned and predictable, once you can see them. IEHP’s data leadership put it simply at CAHP: if you cannot see a problem, you cannot fix it.

No single benefit makes this gap more visible than healthcare transportation.

Why Transportation Is the Proof Point

A transportation trip requires real-time coordination across at least six systems: eligibility verification, prior authorization, vendor scheduling, member interaction, encounter submission, and quality reporting.

Every data failure produces an immediate, visible consequence: the trip happens or it doesn’t.

That binary outcome makes transportation the clearest diagnostic tool we have for data infrastructure health. Kinetik’s presentation at the CAHP Seminar in Burbank laid out the failure pattern precisely: eligibility checked at scheduling, not at service. Authorization data that never reaches the scheduling system. Encounter data submitted days late with no cross-reference to what was authorized. One eligibility gap cascades all the way to understated quality metrics.

Not a transportation problem. An infrastructure pattern that repeats across every complex, multi-vendor benefit health plans administer.

The Compounding Cost

Kinetik’s analysis points to over $100 billion in improper payments across Medicare and Medicaid annually. Over $48 billion in Medicaid overpayments per year trace to eligibility gaps, duplicate billing, and incomplete encounter data. 60 percent of Medicaid audit findings repeat year over year—because disconnected systems weaken oversight at a structural level.

Note: The $100B+ and $48B+ figures reference OIG/CMS reporting. Treat as order-of-magnitude indicators.

What strikes me is not the size of these numbers—it’s their persistence. We keep producing the same audit findings because we keep tolerating the same infrastructure fragmentation. This was not an oversight. It was a structural choice compounded over decades. CalAIM is now forcing a reckoning.

“CalAIM didn’t create a data problem—it made the one we already had impossible to ignore. Transportation just makes the failure visible because every gap has a face: a member who didn’t get picked up, a trip that got billed and denied, a county that never knew the referral failed. The data layer nobody built is now the operational precondition for everything California is asking plans to do.”

— Mohammad Hossain, Chief Business Officer, Kinetik Healthcare Solutions

Kinetik’s thesis—that interoperability is not a technical upgrade but the operational precondition for CalAIM compliance—deserves to travel well beyond the transportation conversation. When all stakeholders operate from a shared data layer, compliance becomes demonstrable, coordination becomes real-time, and outcomes become measurable.

What Closed-Loop Infrastructure Actually Requires

Kinetik describes the solution as a closed-loop data layer: shared infrastructure through which eligibility, authorization, scheduling, member interaction, encounter submission, and reporting all flow in real time—continuously validated, continuously auditable.

Building this layer means surfacing failures before they become member harm, audit findings, or contract performance failures.

Four capabilities this unlocks: real-time eligibility verification at the moment of service; upstream fraud prevention before claims are paid; cross-sector visibility (plans, counties, vendors, and providers operating from the same data); and regulatory confidence (demonstrable oversight that meets the new DHCS standard, not reactive audit response).

Plans that build this now gain a compounding advantage. Plans that defer accumulate structural exposure.

The Second Proof Point: Care Transitions

The afternoon care transitions session showed us what the same failure looks like in a clinical context, with a face.

580

visits in a single year — one patient, cycling across multiple emergency departments with no shared care plan, no system-level view, no coordinated strategy.

An infrastructure failure, not a care failure. And it is happening at scale.

Rufina Wu, Director of Care Coordination at CPMC, Sutter Health, described what real-time data sharing changed at the plan-hospital boundary. CPMC delegates care for approximately 45,000 lives to San Francisco Health Plan. Due to late or missing admission notifications, only about 50 percent repatriation was being achieved, at a cost of approximately $10 million paid annually to outside hospitals for care that could have been delivered in-network.

The human impact: members discharged from SNFs without timely notification to care managers. The seven-day post-discharge PCP visit target routinely missed because no one knew the discharge had happened.

The LA Care care management team described a cultural shift as much as a technical one. Staff who had stopped trusting admission alerts started acting on them. Outreach that used to anchor on discharge date started during the inpatient stay. Plans showed up to conversations with SNFs carrying clinical history, authorizations, and social risk picture, and the relationship shifted from adversarial to collaborative.

Real-time data improved more than coordination. It changed what the hospital-health plan relationship could be.

One case stays with me.

A 75-year-old man with congestive heart failure, COPD, and alcohol use disorder had been cycling among San Francisco and East Bay emergency departments for months. Repeated detox. Returned to the street each time.

The data surfaced his pattern. A multidisciplinary care plan came together across the hospital, health plan, and community partners. They arranged placement in a care home.

He has had no emergency department visits in several years since.

Building the data layer earlier could have prevented this. Building it now makes the next version of that story possible.

What the Research Actually Shows

Three years ago, sitting around a PATH CPI table in Merced with ECM providers, community health workers, county behavioral health staff, and managed care plan representatives, the data infrastructure problem was visible in every meeting.

A lead care manager described trying to reach a member admitted to a hospital three weeks earlier. She found out through a family member who called the clinic. Not through a notification. Not through a system. Through a phone call from a worried relative.

That story is not unusual.

From May 2023 to June 2025, Manatt Health, with support from CHCF and UCSF, engaged 58 California counties to measure what practitioners already knew. The February 2026 report, Beyond Clinical Care: Closing California’s Data Exchange Gap, is the most rigorous documentation of CalAIM’s real-world data infrastructure conditions available.

The numbers:

16% county BH agencies receive hospital alerts

7% county PH agencies receive hospital alerts

26% county BH agencies receive electronic referrals

50% county BH agencies exchange data with few or no partners

The Merced care manager is the norm, not the exception.

Tracing how this happened matters. The Manatt report traces it to 2009. The HITECH Act invested $36 billion in health information exchange infrastructure, nearly all of it flowing to hospitals and clinical providers. Behavioral health, public health, and social services were explicitly excluded. Hospital EHR adoption went from 10% to 96% by 2024. Cross-sector data sharing has not advanced at all in the same period.

California’s DxF created the governance architecture. But the framework is only as functional as the organizations who can participate. A behavioral health agency without a basic electronic documentation system cannot onboard with a QHIO. The rules of the road exist. The vehicles are not there.

“71% of public health respondents cite the lack of even basic electronic systems. 70% of social service respondents do not have funds to purchase and configure electronic systems. 77% of behavioral health respondents report that unclear privacy laws impede data sharing.”

— Manatt Health / CHCF / UCSF, Beyond Clinical Care, February 2026

CalAIM did not create the data infrastructure problem. It inherited it. The difference is that CalAIM’s coordination goals (ECM, Community Supports, closed-loop referrals, HRSN documentation, D-SNP integration) all depend on real-time data exchange that was never built for the sectors that deliver those services.

What Changes When HCAI Takes the Wheel

HCAI is assuming responsibility for the Data Exchange Framework, adding DxF governance to a portfolio that already includes the Healthcare Payments Database (all-payer claims data), hospitalization and cost/quality transparency datasets, affordability oversight through the Office of Health Care Affordability, and California’s rural health transformation program, the third-largest conditional federal award in the country.

A single agency now holds: the Healthcare Payments Database (all-payer claims data), hospitalization and cost/quality transparency datasets, affordability oversight through the Office of Health Care Affordability, California’s rural health transformation program (the third-largest conditional federal award in the country) and DxF governance.

No California agency has held this combination before. The agency that can see utilization, outcomes, and spending across payers is now setting expectations for how data flows operationally.

SB 660, enacted in October 2024, required physician organizations and EMS agencies to participate and created a public signatory list, an accountability mechanism that didn’t exist before. The window to build a New Jersey-style Section 1115 Promoting Interoperability Program into California’s 1115 waiver renewal is open. Whether it gets used depends on whether DHCS and MCPs treat this as a program design problem, not an IT procurement problem.

Medi-Cal Connect illustrates what integrated authority is designed to enable. Sarah Lopez, MD, MBA, Medi-Cal Connect Product Lead at DHCS, described the platform at the morning CAHP session: it aggregates comprehensive claims, WIC, and immunization registry data for all 14 million Medi-Cal members, producing a Longitudinal Member Record, quality dashboards, and the Risk Stratification and Segmentation Tool (RSST), a statewide standardized risk-tiering system designed to identify members most at risk of adverse outcomes. All MCPs are actively being onboarded.

The Manatt data surfaces a sharp irony.

The tool that identifies who needs ECM is being built into every managed care plan. The agencies that need to act on those risk tiers are still the ones where 84% don’t receive electronic alerts. The data layer that identifies who needs care is being built. The layer that closes the loop on whether they received it still isn’t.

PointClickCare^*, SACVALLEY MEDSHARE, and Manifest MedEx, California data exchange intermediaries operating within the DxF, are the connective tissue that can bridge that gap. The closed-loop referral mandate is in effect. The infrastructure to meet it is still catching up.

JUNE 29, 2026 CalAIM Annual Monitoring Report Due

The CalAIM Annual Monitoring Report is due June 29, 2026. Required documentation: closed-loop referral system status, ECM enrollment and outcome data, HRSN data exchange evidence, D-SNP alignment milestones.

For every MCP whose ECM provider network is operating where 74% of behavioral health referrals still travel by phone or fax, the June 29 report will reflect that. For every plan whose D-SNP members are cycling through hospitals without real-time notification to care managers, the monitoring data will show it.

The report is not a compliance exercise. It is a mirror. And the mirror is going to show California’s new Medicaid infrastructure sitting on top of a data exchange foundation that was not built to support it.

“The urgency is rising. To meet its goals under CalAIM and manage anticipated H.R. 1-related fiscal constraints, California needs scalable solutions to maximize efficiency. Cross-sector data sharing is an essential step.”

— Manatt Health / CHCF / UCSF, Beyond Clinical Care, February 2026

The organizations that close this gap now will have a measurable advantage in ECM monitoring, D-SNP Stars performance, and HRSN compliance. June 29 is the first public evidence of who built the infrastructure and who did not.

Demonstrating Value, and Who It’s For

High-risk individuals experiencing homelessness cost California’s health system approximately $100,000 per person per year. When housed, that drops to approximately $40,000. The data infrastructure that makes that story visible and enables the intervention is the same investment.

Someone is on the other side of every data gap.

The member who didn’t get picked up. The patient with 580 visits and no care plan. The 75-year-old man returned to the street. The child who didn’t get vaccinated because the system that should have reached them couldn’t see them.

California is raising the bar on what demonstrating value means. The plans and partners that can tell this story, with data and in real time, will set the standard.

The Pattern Generalizes

CalAIM built an expansive vision of what Medi-Cal can do for California’s most complex members, and that vision requires infrastructure capable of holding it.

Thirty years working the intersection of healthcare policy and on-the-ground delivery reinforces one pattern: lasting change follows a sequence.

Analog before digital. Get the human systems right before layering technology on top. Workflow, consent, staff buy-in, MCP relationships. These come first. Technology on top of broken human systems doesn’t just fail. It fails faster.

Technology enables, relationships transform. Once the foundation is sound, technology amplifies it. The CPMC and LA Care panel on March 25 wasn’t a technology case study. It showed what real-time data made possible between people who finally had something to work with.

Orchestration over digitization. Connecting the whole ecosystem so the pieces move together matters more than digitizing each touchpoint separately. Medi-Cal Connect, the DxF, the RSST, the QHIO network, the closed-loop referral mandate — these aren’t separate tools. They are the architecture of an orchestrated system still under construction.

Every element we examined on March 25 fits somewhere in that arc. The data layer nobody built is the next frontier, and understanding this first separates the organizations that will define operational accountability from those still waiting to be defined by it. The plans, counties, and technology partners that move now are the ones who will set the standard.liest will be the ones that define what operational accountability actually looks like in practice.

We are at the beginning of this work, not the end. The field is watching.

Primary source: Kinetik Healthcare Solutions, CAHP Seminar Series 2026, Burbank CA, March 25, 2026. Care transitions panel: Rufina Wu (CPMC, Sutter Health), Amanda Asmus and Thomas Baird (LA Care Health Plan), moderated by Lia Quilty (PointClickCare). Additional CAHP sources: IEHP data & technology leadership; Sarah Lopez, MD, MBA, Medi-Cal Connect Product Lead, DHCS; Michael Valle, HCAI Deputy Director; Christine McKinney, Blue Shield CA; Jeff Rideout, IHA. Medi-Cal Connect / RSST: Sarah Lopez, DHCS, CAHP Seminar Series 2026, March 25, 2026; and DHCS Medi-Cal Connect presentation, March 2025. CMS encounter data thresholds: federal CMS guidance. Care transition figures attributed to panel speakers, not published slides.

Research source: Manatt Health / CHCF / UCSF. “Beyond Clinical Care: Closing California’s Data Exchange Gap.” February 2026. County statistics from the 2025 California County Data Exchange Survey. Full report: chcf.org.

*Client disclosure: PointClickCare is a consulting client of Hickman Strategies LLC. This relationship does not affect the independence of CalAIM Field Notes analysis.

On February 5, DHCS held a webinar on its H.R. 1 Implementation Plan Overview. It was detailed, serious, and member-centered. They are doing the work.

Something about it just stayed with me.

Not the policy mechanics. Not the timeline. One number.

Approximately 2.8 million new adult group members will likely require manual verification under the new work and community engagement requirements. DHCS’s own projection: up to 50% of those members could lose Medi-Cal coverage due to paperwork burden — not because they’re ineligible, but because the administrative system couldn’t reach them in time.

That’s roughly 1.4 million people.

Monday’s CCI convening in Oakland got me thinking about this number again. Seventeen TA organizations in one room, comparing notes on what’s actually working in CalAIM implementation. The room wasn’t defeated. It was diagnostic. And underneath every finding was the same unresolved question: who closes the gap when the system can’t reach the people it’s supposed to serve?

Then I had a conversation with someone I had been meaning to catch up with for a long time — Cathryn Couch.

Cathryn Couch, founder of Ceres Community Project and leader of California’s CBO Coalition, puts the available asset plainly: 700 community-based organizations, all 58 counties, serving every population of focus, every Community Support. Coalition members are already seeing Medi-Cal members 40 to 50 times a week — as trusted messengers, already in the room, already holding the relationship. They are largely not funded or included in most plan redetermination strategies.

Here’s the question nobody wants to answer directly: what are the incentives for running an aggressive redetermination campaign — and what are the disincentives?

California’s largest plans have committed tens of millions of dollars to redetermination outreach. The early campaigns lean heavily on awareness — billboards, social media, mass outreach. Those tools reach people who are already reachable. The 50% gap lives in the people they don’t reach. That’s not a criticism of the plans. It’s a design constraint that requires a different tool.

That tool exists. Not the generic mailer. Not the billboard. The trusted CBO worker who already sees 40 members a week. The food-is-medicine provider who has been in the room for four years. The three questions every community health worker should be asking every Medi-Cal member right now:

Did you get your yellow envelope? Do you have it with you? Do you need help filling it out?

Medicaid managed care was not designed with redetermination campaigns in mind. When enrollment drops, capitation revenue drops — but so does utilization. The system doesn't automatically reward the plan that works hardest to keep members enrolled. That's not a flaw in anyone's values. It's a gap in the architecture. And it's exactly the kind of gap that aligned incentives, backbone infrastructure, and accountable performance floors can close — if we build them before January 2027.

The work requirement is federal law. We implement it. The 50% administrative failure rate is not inevitable — it’s an organizing challenge. And it has a deadline: January 1, 2027, with the first coverage losses hitting by June.

January 1, 2027 is nine months away. Federal guidance arrives in June. That leaves roughly 90 days between guidance and implementation.

Here’s the question every MCP, CBO, and county needs to answer before Q3: Do we have a specific, measurable plan to close our share of that 50% gap?

If the answer is not yet — that’s the organizing challenge. And the window to build the infrastructure that addresses it is open right now.

No state agency closes that gap alone. Trusted messengers do. Coordinated outreach does. Backbone infrastructure — the kind the 1115 waiver is trying to fund — does.

This is what the CCI convening surfaced last Monday in Oakland: trust is not a soft outcome. It is infrastructure. The counties with 3–7% ECM utilization gaps aren’t failing on policy design. They’re at different points on the trust curve. Eighty percent of what works is consistently showing up.

The CBO Coalition’s 700 organizations are already showing up. The organizing challenge — specific, measurable, urgent — is building the policy and payment architecture that lets them close the gap before Q3.

There are many groups like Cathryn’s organized around critical health challenges across this state. The infrastructure exists. The relationships exist. The funding logic exists.

What’s the first funded pilot that proves the model?

We can control what we build while the window is open.

Full analysis of what this means for the 1115 waiver, MCPs, and county implementation continues in No. 14, publishing April 2.

———

Source: DHCS H.R. 1 Implementation Plan Overview Webinar, February 5, 2026. Figures represent DHCS’s own projections.