← Archive

Everyone's Watching the Startups. The Risk Sits Somewhere Else.

The durable risk-holders in pediatric value-based care are not the venture-backed names that get the press. They are hospital-built plans and ACOs that have been quietly bearing pediatric Medicaid risk for thirty years. CMS's ASPIRE model is a sorting mechanism, not a rising tide.

Pediatric Health Dispatch | Deep Dispatch | May 21, 2026

The Risk Has Been Sitting in Plain Sight Since 1994

Last week we walked through the hospital-to-home transition gap: the technology-dependent child who is medically ready to leave but has nowhere structured to go. The piece ended on a question we didn't answer. When that child finally goes home, who actually holds the financial risk for whether they thrive or bounce back through the emergency department? This week we answer it, and the answer is not the one the funding headlines would lead you to expect.

Pediatric value-based care has a press problem. The companies that get written up are the venture-backed ones: Imagine Pediatrics and its $67M Series B, Bluebird Kids Health's $31.5M Series A, Nest Health layering in midwifery on top of a $22.5M round. These are real companies doing real work, and we cover them closely. But they are not where most pediatric risk actually lives. The entities quietly bearing the financial weight of children's Medicaid populations have been doing it for decades, and almost none of them are startups.

Texas Children's Health Plan, founded in 1996 as the nation's first HMO built for children, covers more than 600,000 residents across 50-plus counties through STAR, CHIP, and STAR Kids. Cook Children's Health Plan carries roughly 120,000 members in North Texas, and during a 2025 state procurement dispute its own materials revealed that nearly 125,000 of those members were at risk of losing their plan overnight, a useful reminder of where the leverage really sits. Partners For Kids, created by Nationwide Children's Hospital in 1994, is the oldest and largest pediatric ACO in the country, accountable for somewhere between 325,000 and 470,000 children across 47 Ohio counties depending on which Nationwide page you read. A 2015 Pediatrics analysis found Partners For Kids held cost growth to $2.40 per member per month per year, against $16.15 in Ohio Medicaid fee-for-service and $6.47 in Medicaid managed care, while holding or improving quality. That is a twenty-year head start on a problem the venture market discovered in 2022.

Then there is the contracting layer. Seattle Children's Care Network, a clinically integrated network of 17 primary care practices across 23 locations, has generated over $12.5M in medical and pharmacy savings in commercial value-based contracts and holds agreements with 80% of the Medicaid MCOs in its area. It did this without becoming an insurer. That distinction matters more than it sounds, and we'll come back to it.

The strategic point is simple and underappreciated: pediatric Medicaid lives have been inside risk-bearing machinery for a long time. MACPAC data shows the vast majority of children in Medicaid have sat in some form of managed care for over a decade. What's new in 2026 is not that someone invented pediatric risk. It's that the risk is finally being unbundled, repriced, and contested, and a new layer of operators is trying to insert itself between the plans and the families. The question is whether they can.

Four Layers, and the Margin Question No One Asks Out Loud

The market keeps debating whether pediatric VBC "works," which is the wrong question asked at the wrong altitude. The useful question is structural: in a four-layer stack, which layer holds the risk, which captures the margin, and which actually controls where the medically complex child goes next?

The stack now sorts cleanly into four tiers. At the top are the risk-bearing pediatric plans, the Texas Children's and Cook Children's chassis that take full capitation at the plan level. Below them sit the hospital-led ACO and CIN infrastructure, Partners For Kids, Seattle Children's Care Network, CHOP's Compass Care, which take delegated or shared accountability without owning the insurance license. Then come the venture-backed care-delivery models, Imagine Pediatrics, Nest Health, Bluebird Kids Health, which contract into the plans above them and take risk on a defined high-cost cohort. And finally the MSO and roll-up aggregators, US Pediatric Partners (75+ offices, 400+ clinicians, built by Webster Equity Partners) and Pediatric Associates Family of Companies (260+ locations, 1.5M+ patients across seven states), which assemble practice scale and bolt value-based positioning on afterward.

Here is the part the press releases skip. The startups in layer three do not hold the patient relationship the way the plans in layer one do, and they do not own the practice footprint the way the aggregators in layer four do. They sit in the middle, dependent on a Medicaid MCO above them for the contract and on a network of providers around them for the care. Their margin exists only to the degree they can demonstrably bend avoidable utilization on a population the plan can't manage itself. That is a genuinely defensible position when the cohort is right. Imagine Pediatrics reports $65M in 2024 plan savings, 70,000+ children served, and 8,350 additional Safe Days at Home, all built around children with special health care needs, the 5% of Medicaid kids who drive roughly half of pediatric spending. That is the cohort where the math works.

It is a precarious position when the cohort is wrong. The vault is full of companies wearing the VBC label that, on current public materials, still monetize like fee-for-service access businesses. Zarminali Pediatrics raised $110M and looks more like a multispecialty super-group than a delegated-risk operator; Summer Health is moving into Medicaid channels with no visible capitated economics. The label is cheap. The risk-bearing chassis underneath it is not, which is why the most credible new entrant, Bluebird, is run by Chris Johnson, who built Landmark Health's at-home model and sold it to Optum for roughly $3.5 billion. He is not improvising. He is porting a proven adult playbook into pediatrics, and even he started with six clinics in Florida.

ASPIRE Is a Sorting Hat, Not a Subsidy

On March 24, CMS handed this entire stack a catalyst it has never had: a federal pediatric care model. The ASPIRE model commits up to $125M to as many as five states over an eight-year grant period to build whole-child Medicaid and CHIP programs for children with complex medical and behavioral needs. Its design requirements read like an Imagine Pediatrics product spec: a single care coordinator per family, 24/7 clinical access with full context, and joint management of physical, behavioral, and social needs. After thirty years of CMS building accountable-care pathways for seniors and almost nothing equivalent for children, this is the first real federal scaffolding for pediatric VBC. The May 15 EPSDT Coverage Guide reinforces the floor underneath it, narrowing states' room to cap or delay children's behavioral-health and telehealth services. The policy tailwind is genuine.

The temptation is to read this as a rising tide that lifts every boat with "value-based" in its pitch deck. That reading is wrong, and the evidence arrived three weeks ago. An April 23 study in JAMA Network Open found that Medicaid ACO implementation was not associated with broad improvements in behavioral-health access for children, and was associated with higher reported unmet mental-health need. Structure alone does nothing. Putting kids in an ACO box and expecting outcomes to follow is the pediatric equivalent of buying a gym membership and waiting to get fit. What works is operator capability tightly matched to a high-cost subgroup, the exact thing that distinguishes a Partners For Kids or an Imagine Pediatrics from a practice group that recently learned to say "value-based."

So ASPIRE functions as a sorting mechanism. The states that win will need a real operator to run the model, and they will reach for entities that already combine 24/7 access, care coordination, behavioral integration, and Medicaid plan partnerships. That is a short list, and it is concentrated in layers one through three of the stack. The aggregators in layer four, the PE-built practice platforms, are exposed here: scale without a delivered risk model is not what a state Medicaid agency writing an ASPIRE application needs. Expect the winning consortia to pair a hospital-led network or a provider-sponsored plan with a complex-care operator, not a roll-up.

Who wins, then? The hospital-built plans and ACOs win by default: they already hold the risk and the relationships, and now have a federal model to formalize them. The best pure-play operators, Imagine above all, win by becoming the delegated engine inside those state programs, a better business than trying to displace the plans. The aggregators are exposed, because their thesis assumes practice scale converts to risk-bearing leverage, and ASPIRE plus the JAMA evidence both suggest scale is necessary but nowhere near sufficient. Most exposed of all are the access companies still calling themselves VBC. When states start writing real risk into contracts, the gap between a company that holds risk and a company that holds appointment slots stops being a marketing nuance and becomes a balance-sheet fact.

---

---

Pediatric Health Dispatch publishes every Tuesday (curated roundup) and Thursday (deep-dive analysis). Subscribe at pedshealthdispatch.com.