Shut Down by August 16: Provo Canyon, the Fourteen-Day Loophole, and the Machine That Warehouses America's Foster Children
Investigation by Project Milk Carton
Shut Down by August 16: Provo Canyon, the Fourteen-Day Loophole, and the Machine That Warehouses America's Foster Children
A Project Milk Carton investigation — child welfare transparency, public data, and the per-diem economy built on children who cannot leave.
By PMC Intelligence Desk | July 19, 2026
PROLOGUE: THE LICENSE ON THE WALL
On July 17, 2026, the State of Utah revoked the license of Provo Canyon School's main campus and ordered every service on the property to end by August 16. The revocation notice, issued by the Utah Department of Health and Human Services, did not speak in the abstract language of regulatory noncompliance. It cited "cruel and unnecessary" pain-compliance techniques used on a child. It cited discipline designed to "frighten or humiliate." It cited the deprivation of water, rest, and access to a toilet. It cited a failure to protect children from abuse — and, perhaps most tellingly, it cited the facility for blocking the children inside from reporting what was happening to them.
"No child should be hurt in a program that is meant to protect them," said Shannon Black, the state's licensing director.
Eleven days earlier, on July 6, Utah had already revoked the license of Provo Canyon's Springville campus — the girls' program — citing inadequate staff ratios, unnecessary restraint, background-check failures, and the systematic under-reporting of incidents, with the children ordered out by August 6 and the operator barred from reapplying for five years. The July 17 order finished the job: the entire institution, shut. Universal Health Services, the hospital giant that has owned Provo Canyon School since 2000, said it was reviewing its appeal options within the fifteen-day window the state allows.
Paris Hilton — who spent eleven months inside Provo Canyon as a teenager and spent the last six years telling Congress and the country what happened there — posted the same day: "Provo Canyon School is officially shut down."
Here is the question this investigation exists to answer, and it is not what happened at Provo Canyon. That is now a matter of state record. The question is: what kind of system kept a facility like this full, funded, and federally reimbursed for decades — and what happens to the thousands of facilities like it that no state has gotten around to inspecting on the right day?
The answer runs through a single subsection of the Social Security Act — 42 U.S.C. § 672(k), the fourteen-day rule and the loophole drilled through it — down into a per-diem business model a U.S. Senate committee has called "endemic[ally]" abusive, through a private-equity ownership map that has extracted hundreds of millions of dollars from taxpayer-funded child placements, past a paid-for accreditation certificate that substitutes for public oversight in nearly half the states, and into a second, quieter federal revenue stream that Project Milk Carton traced through its own grant-ledger database — a stream no outside reporting has connected to this industry.
Everything in this article is drawn from primary sources: the U.S. Code, the Senate Finance Committee's June 2024 report Warehouses of Neglect, the Government Accountability Office's March 2026 survey of all fifty states (GAO-26-107592), Department of Justice settlement records, state licensing actions, Bureau of Justice Statistics survey data, and Project Milk Carton's own federal grant ledgers, cross-validated against USASpending.gov. Every figure is attributed. Where a widely repeated number could not be traced to a primary source, we did not use it — and we tell you which ones at the end.
One rule governs the money sections of this piece, and we state it up front: a funding record is a funding-structure fact, not an allegation. Where we document that an operator receives federal money, that is a statement about how public dollars flow — nothing more. Where we describe abuse findings, license revocations, or survey data, those findings belong to the public agencies, courts, and committees that made them, and we cite each one.
The investigation proceeds in seven movements. First, the law — the fourteen-day rule Congress wrote in 2018 and the certificate that switches it off. Second, the model — what the Senate found when it spent two years inside the four largest operators. Third, the owners — the private-equity chain of custody over the beds. Fourth, the certificate — the accreditation economy that defeats two layers of oversight at once. Fifth, the ground truth — what all fifty states told the GAO in 2026 about where the children actually are. Sixth, the second stream — Project Milk Carton's own money findings. Seventh, the missing — the bridge between the understaffed congregate facility and the child who walks out the door and is never properly counted, which connects this investigation to our § 671(a)(35) reporting-failure series.
PART I — THE FOURTEEN-DAY RULE AND THE HOLE IN THE MIDDLE OF IT
What Congress actually wrote
In February 2018, Congress passed the Family First Prevention Services Act as part of Public Law 115-123. Its central bet was one the child-welfare research community had urged for decades: children do better in families than in institutions, so the federal government should stop paying for institutions by default.
The mechanism was Title IV-E — the multi-billion-dollar federal foster-care reimbursement stream that flows to states for eligible placements. Family First amended 42 U.S.C. § 672(k) to impose a simple clock: federal IV-E reimbursement for a child placed in a non-family setting stops after fourteen days. After two weeks, if the child is still in a group setting, the state pays alone.
Unless.
Section 672(k) carries its own exceptions, and the largest of them — § 672(k)(4) — is the Qualified Residential Treatment Program, the QRTP. If the congregate facility qualifies as a QRTP, the fourteen-day clock never runs. Federal money continues indefinitely. Two narrower exemptions in § 672(k)(2) also switch the clock off: settings for pregnant and parenting youth, and — remember this one, because it returns in Part V with a vengeance —settings serving youth who are victims of, or at risk of, sex trafficking.
So, the entire architecture of Family First's congregate-care reform compresses into a single question: what does it take to become a QRTP?
The gate is made of paper
To qualify, a facility must be accredited by one of three private organizations — the Joint Commission, the Council on Accreditation, or CARF — and the placement must clear a set of procedural safeguards codified at 42 U.S.C. § 675a(c): within thirty days, a "qualified individual" must assess whether the child actually needs a residential level of care; within sixty days, a court must approve the placement; and every twelve months (every six months for a child under thirteen), the head of the state agency must personally sign off on continued placement.
Investigation data visualization
Read those safeguards again and notice what every one of them is: a document. An assessment form. A court order. A signature. Not one of them tests whether the child got better, whether the facility's restraint numbers are rising, whether children keep going missing from the campus, or whether the "treatment" being billed to Medicaid and Title IV-E is happening at all. There is no federal audit of placement necessity. There is no federal outcome standard a QRTP must meet to keep its status. The gate that controls indefinite federal reimbursement for institutionalizing a foster child is a paperwork gate — and the accreditation certificate at the center of it, as Part IV will show, is purchased by the facility from a private vendor that answers to no public body.
It is worth walking through the three § 675a(c) safeguards one at a time, because each fail in its own instructive way. The thirty-day assessment asks a "qualified individual" whether the child needs residential care — but it is a one-time judgment made at the front door, before the facility's conduct has had a chance to reveal itself, and nothing in the statute requires anyone to revisit it when the restraint logs start filling up. The sixty-day court approval puts a judge's signature on the placement — but a dependency judge sees the assessment paperwork, not the facility's staffing ratios or its police-call rate, and the Senate record in Part II is a catalogue of facilities whose internal documents said one thing while their court-facing paperwork said another. And the twelve-month agency-head sign-off — six months for a child under thirteen — asks the head of a state agency responsible for thousands of children to personally certify the continued necessity of each institutional placement, a certification that in practice can only ever be a signature atop a stack of assurances assembled by the layers below. Three checkpoints; not one of them touches the child, visits the building, or reads an outcome.
Congress built a fourteen-day fence around congregate care and then handed the industry the key. What the industry did with that key is the subject of the rest of this article.
PART II — THE MODEL: WHAT THE SENATE FOUND INSIDE
"Endemic to the operating model"
On June 12, 2024, the Senate Finance Committee released Warehouses of Neglect: How Taxpayers Are Funding Systemic Abuse in Youth Residential Treatment Facilities— the product of a two-year investigation into the four largest operators of residential treatment facilities for children: Universal Health Services, Acadia Healthcare, Devereux Advanced Behavioral Health, and Vivant Behavioral Healthcare, the successor to Sequel Youth & Family Services.
The committee's core finding was not that abuse occurs in these facilities. It was that abuse is structural: harm to children, the report concluded, is not isolated but" endemic to the RTF operating model"— a model in which revenue is maximized by filling beds at a per-child, per-day rate, and profit is maximized by cutting the largest cost line, which is staff.
That is the entire economic engine of this industry in one sentence, and every horror in the committee's 100-plus pages is downstream of it. Consider what the per-diem model looks like from the inside:
The rate. Facilities charge $275 to more than $800 per child per day, per NBC News's investigation of Sequel — figures the Senate's own document trail corroborates. At Sequel's Northern Illinois Academy, the charge ran$450 to $850 a day, and internal planning documents obtained by APM Reports showed the facility mapping a path to grow its profit roughly six-fold, from $460,000 to $3 million.
The revenue. Sequel — a chain built substantially on foster children and juvenile-justice placements — generated$210 to $230 million in revenue and $30 to $32 million in operating profit in 2015, according to a video presentation by co-founder Jay Ripley obtained by the Senate committee. Ripley told his investor audience the demand was like "drinking from a firehose" — and stated the model's arithmetic plainly: "you can make money in this business if you control staffing." In that year, Sequel spent62 percent of its revenue on staffing— meaning the path to the $30 million sat directly on top of the payroll line.
The scale. At the top of the industry the numbers are of a different order entirely. UHS's behavioral health division booked$6.2 billion in net revenue in 2023— roughly 43 percent of the entire company — with Medicaid supplying 39 percent of it. At Acadia, Medicaid was53.9 percent of revenue in 2023, and the company told investors it was adding "as many as 1,000 or more" beds a year. Devereux, a nonprofit, runs more than $500 million in annual revenue, about half from Medicaid.
The instruction to fill the beds. On an October 2023 earnings call, UHS's chief financial officer told investors that "broadly increasing occupancy [of our behavioral business] is the most significant opportunity we see." Occupancy — children in beds — described to Wall Street as the company's most significant opportunity, four months before the Senate report and thirty-three months before Utah shut the company's flagship youth facility.
What the model does to the children inside
The Senate report's evidentiary spine is a set of internal documents, inspection records, and Bates-stamped files from the four operators. A sample, each item cited to the report's own pages:
Piney Ridge (Acadia — Fayetteville, Arkansas). Children were "routinely simultaneously chemically restrained and secluded... in violation of federal regulation." In one thirty-day period the facility logged110 restraints and seclusions. Acadia's corrective action, after its leadership learned of the pattern: a ten-question multiple-choice test for staff. When investigators pressed executives on the numbers, the executives "objected to the premise" that 110 restraint-and-seclusion events in a month was a problem.
Cedar Ridge (UHS — Oklahoma). A staff member was sexually abusing a child on an ongoing basis. The facility's response was not to fire her but to move her to another wing— from which, the committee found, she returned to the child's window night after night.
Devereux Red Hook (New York). Staff falsified bed-check records. A child was logged as present through the night and discovered missing at 6:45 a.m. He had been struck and killed by a truck at 4:50 a.m., four and a half miles from the facility. The bed-checks that would have noticed him gone existed only on paper — the same medium, it bears noting, as every QRTP safeguard in § 675a(c).
The elopement pattern. At Acadia's Millcreek of Arkansas, at least one child eloped — walked away from the facility —in every single month of 2022. At Cedar Ridge, three teenagers walked past sleeping staff and out of the building; two of them were found in the woodsten days later. Hold that pattern in mind for Part VII, because a child who walks out of an understaffed facility becomes, in the language of a different federal statute, a child missing from foster care — and this investigation's companion piece documents what states do and do not do about those children.
The police-call rate. Salt Lake Tribune reporting cited by the committee found Provo Canyon School generated 29 police calls over four years, with a sex-crime call rate four times the average of comparable Utah facilities; Copper Hills ran 1.6 times the average. Asked about pervasive sexual abuse in residential facilities, Devereux senior vice president Leah Yaw offered the industry's own verdict: it "is an industry-wide problem."
That is the operating model the QRTP certificate protects. Now the owners.
PART III — THE OWNERS: A CHAIN OF CUSTODY WRITTEN IN DEBT
Investigation data visualization
The Sequel file: a child-warehousing company as a financial instrument
No single company illustrates the financial life cycle of this industry like Sequel Youth & Family Services, and the paper trail assembled by the Senate, APM Reports, and the Private Equity Stakeholder Project reads less like a healthcare history than a securities prospectus:
Founded with backing from Patriot Capital; sold to Levine Leichtman Capital Partners in 2010; sold to Alar is Royalty in 2013, which booked$71 million in profit — a 23 percent annual return— and took a dividend recapitalization in August 2016; sold to Altamont Capital Partners in a 2017 leveraged buyout.
Between 2017 and 2021, at least$121.5 million in debt was loaded onto the company.
When the company collapsed in 2021 — after NBC's and APM's reporting on restraint deaths and abuse, including the 2020 death of sixteen-year-old Cornelius Fredericks in a Michigan facility, made the brand untenable — its roughly$200 million debt to lender FS KKR Capital resolved with lenders expecting to recover$78 million of $199 million: thirty-nine cents on the dollar.
And then the model simply re-formed. Vivant Behavioral Healthcare, founded in 2021 by Sequel co-founder Jay Ripley, took over thirteen former Sequel facilities — staffed, the Senate committee documented through LinkedIn records on file, by former Sequel executives: a senior vice president of strategy, a chief clinical officer, a president. During the Senate's own investigation, Vivant divested nine facilities in Alabama and Florida.
For readers outside finance, one term in that chain deserves translation, because it is the heart of the extraction: the dividend recapitalization. In a dividend recap, the owner does not wait for the company to earn profits and distribute them. Instead, the company borrows money — debt that lands on the company's own balance sheet — and pays the borrowed cash out to the owners as a dividend. The owners are paid immediately; the company, and everyone depending on it, carries the repayment obligation. When the company is a chain of children's residential facilities whose revenue is public per-diems and whose largest controllable cost is staff, the repayment pressure has exactly one place to go. Alaris took its dividend recap from Sequel in August 2016; Altamont's leveraged buyout followed in 2017; at least $121.5 million in debt was added over the next four years; and the collapse settled at thirty-nine cents on the dollar. The children in the facilities were never party to any of those transactions — but the staffing levels that federal and state inspectors later documented were, in an arithmetic sense, those transactions' residue.
The same extraction pattern repeats across the ownership map:
The Private Equity Stakeholder Project's appendix runs longer still: Webster Equity (Discovery), Pharos (Logan River), Ridgemont (Perimeter), Relay (Embark/Calo), Pine Tree (Paradigm, Intercept, Shortridge), Palladium (Health Connect America), VSS/Cimarron (Ascent), KKR (BlueSprig), Blackstone (CARD).
Follow the through-line: firms buy operators whose revenue is overwhelmingly public money — Medicaid and Title IV-E per-diems attached to children who did not choose to be there and cannot leave — load them with debt, pay themselves dividends out of the borrowing, and exit before, or through, the collapse. When the collapse comes, the lenders take thirty-nine cents on the dollar, the executives found a successor company, and the children are transferred to the next licensee. The public pays at every stage: the per-diem going in, the Medicaid claims throughout, and the cost of the harm afterward.
And the federal government's defense against all of this — the one gate between this ownership model and indefinite IV-E reimbursement — is the certificate we turn to now.
PART IV — THE CERTIFICATE: AN OVERSIGHT SYSTEM THE INDUSTRY PAYS FOR
The price list
To hold QRTP status — to switch off the fourteen-day clock — a facility needs accreditation from the Joint Commission, the Council on Accreditation, or CARF. These are private organizations. The facility being inspected is the paying customer.
The fees are public. The Joint Commission's behavioral-health accreditation starts at $1,990 per year plus an on-site survey fee starting around$3,430, per the Commission's own pricing documents; industry all-in estimates run $13,000–$25,000 for initial accreditation and $6,000–$12,000 a year thereafter. CARF charges a $995 application fee plus $1,670–$1,840 per surveyor, per day, under its published 2025–27 fee schedules. (The Council on Accreditation's exact fee amounts we could not verify from an accessible primary document, and so we do not state them.)
For an industry charging $275 to $800 per child per day, accreditation is not an obstacle. It is a rounding error — a few days of one bed's revenue, purchased from a vendor whose income depends on facilities seeking its certificate.
What the certificate unlocks
The same purchased certificate defeats two independent layers of oversight simultaneously:
First, the federal layer. Accreditation is the load-bearing wall of QRTP status under § 672(k)(4) — it is what converts a fourteen-day federal reimbursement limit into an unlimited one. Joint Commission accreditation additionally carries CMS "deemed status," meaning it substitutes for public Medicare/Medicaid certification surveys.
Second, the state layer. Per the findings attached to the BRIDGES for Kids Act introduced in December 2025, in twenty-one states an accreditation certificate lets a residential facility bypass some or all of the state's own licensing inspection steps. In those states, the private certificate is not a supplement to public inspection — it is a replacement for it.
Does the certificate mean anything?
Investigation data visualization
The Senate committee answered this question with the accreditors' own paperwork. Among the Bates-stamped documents in the committee's files are the Joint Commission's own Final Accreditation Reports documenting restraint-related failures at UHS facilities — Palmetto Summerville in South Carolina, Hill Crest in Alabama, Copper Hills in Utah (where restraint orders were unauthenticated in half the records reviewed), Pavilion in Illinois, North Star DeBarr in Alaska. Every one of those facilities remained accredited and operating. The committee's Recommendation 10 addressed the accreditors directly, suggesting they "should consider withholding accreditation" from facilities with documented abuse — a sentence remarkable mostly for the fact that a U.S. Senate committee felt the need to write it down.
And Provo Canyon School itself — the facility whose license Utah just revoked for "cruel and unnecessary" treatment of children — operated for decades as a licensed, accredited, federally reimbursed facility. The certificate did not fail to detect what Utah found in 2026. The certificate was never designed to detect it. It measures the presence of policies, not the treatment of children — paper, again, all the way down.
PART V — THE GROUND TRUTH: WHAT ALL FIFTY STATES TOLD THE GAO
If the QRTP framework worked as designed, six years of implementation should show congregate placements falling, qualified facilities rising, and the fourteen-day rule doing its job. In March 2026, the Government Accountability Office published the first comprehensive answer — GAO-26-107592, a survey answered by 49 of 50 state child-welfare agencies and all 50 juvenile-justice agencies. It is the closest thing that exists to the system's own confession.
The placements did not fall. Twenty-six of forty-nine states reported the share of their foster youth in congregate care had increased or stayed flat since October 2021. Twenty-six said the same about the average number of days children spend there.
The qualified facilities did not materialize. Only 34 of 49 states have any QRTPs at all — roughly fifteen states have zero. Of the 34 that have them, ten report that only 1 to 25 percent of their congregate facilities qualify, and two states could not say what share of their own facilities qualified. Thirty states called the QRTP requirements challenging to meet; the hardest, they said, is the six-month aftercare obligation — the one requirement aimed at what happens to the child after the institution.
The children went to hotels. Forty-three states have used out-of-state congregate placements since October 2021. Thirty-five states report that non-QRTP providers refuse children with significant behavioral needs, autism, or intellectual disabilities — the children the system most needs placements for are the ones the market declines. And forty-two states have housed foster children in hotels, offices, or emergency rooms; eighteen states increased that practice. All forty-nine responding states report placement-capacity problems. Exactly four say their supply of therapeutic foster homes is sufficient.
The states quietly ate the cost — which is the loophole's fiscal signature. Twenty-nine of forty-nine states saw their federal IV-E congregate reimbursements fall after the fourteen-day rule took effect. But the children were still in the facilities — so thirty-three states raised state, county, or local money to keep paying for the same placements. One state reported the federal share of its congregate-care funding fell from17.9 percent to 1.6 percent between 2018 and 2024; another, from 22.3 percent to 3.3 percent. Understand what that means: the fourteen-day rule, in large parts of the country, did not move children out of institutions. It moved the bill— from the federal government to state and local taxpayers — while the per-diem economy underneath kept collecting.
And then there is the trafficking exemption. Recall § 672(k)(2)'s exemption for settings serving victims of sex trafficking — written so that a trafficked child's specialized placement would never lose funding on day fifteen. The GAO found states reading that clause in flatly divergent ways. Among eight follow-up states, three claim IV-E reimbursement past fourteen days for any youth placed in a designated facility, trafficking victim or not. At least one state has designated every congregate facility it operates as a trafficking-serving setting — and claims federal money past the fourteen-day limit statewide. An exemption written for trafficked children, converted into a master key for the entire fourteen-day rule.
The GAO's first recommendation asked the Administration for Children and Families to clarify the exemption. HHS did not concur. The agency that runs Title IV-E declined the recommendation to close the loophole's widest-open door. As of the GAO's publication, no state has faced a penalty for any of it — and there is no federal audit of QRTP placement necessity for a penalty to arise from.
Step back and consider what the cost-shift numbers actually prove, because they are the closest thing this story has to a controlled experiment. If the fourteen-day rule had worked as designed, federal congregate spending would fall because placements fell— the two curves moving together. Instead, the states' own answers show the federal curve collapsing — 17.9 percent to 1.6 percent in one state, 22.3 to 3.3 in another — while placements held flat or rose in twenty-six of forty-nine states and thirty-three legislatures back-filled the difference with local money. Federal spending fell; institutionalization did not. That divergence is the loophole's fiscal fingerprint: the rule succeeded in moving dollars and failed at moving children, which means the per-diem economy's grip on placement decisions is stronger than the federal incentive designed to break it. A policy that shifts costs without shifting outcomes is not a reform. It is a relabeling — and forty-nine state budget offices have now documented it in a federal survey, in their own words.
The shipping lanes
The capacity collapse has a geography. West Virginiaspends$62 million a year placing foster children in out-of-state facilities — a figure from February 2026 reporting — while its governor seeks $6 million to bring some of them home; the imbalance between those two numbers is its own indictment. Texas sent one in three children placed in 2025 outside their home region. California shipped 676-plus children out of state over five years before halting subsidies for out-of-state treatment. Oregon pulled every foster child out of out-of-state for-profit facilities in 2020 after the Sequel scandals — and 2025 legislation would quietly permit the practice again. The national baseline, per AFCARS data surfaced by ProPublica: 1,716 out-of-state treatment placements as of 2018, before the post-Family First capacity squeeze drove states back to the interstate market.
Utah's revocation of Provo Canyon is real enforcement — as are Ohio's move against Mohican Young Star Academy (improper restraints, escapes) and Utah's February 2026 termination of Crimson Heights after a staff-minor sexual relationship. But note what enforcement looks like in this system: a state revokes one license, and the children scatter into a national market that forty-two states are already backstopping with hotel rooms and office floors. The machine does not stop. It re-routes.
PART VI — THE SECOND STREAM: WHAT OUR OWN LEDGERS SHOW
This section reports Project Milk Carton's original data findings, queried July 19, 2026, from our database of federal grant records (TAGGS and USA Spending-derived tables, cross-validated). The rule stated in the prologue applies with full force here: these are funding-structure facts — documentation of how public money flows — and are not allegations of wrongdoing by any operator named.
The same industry, a second federal spigot
The congregate-care industry under Senate investigation for warehousing foster children has a second line of business that almost no coverage connects: operating shelters for the Office of Refugee Resettlement's Unaccompanied Children Program— the federal system that houses migrant children who arrive without a parent. Same buildings-and-beds business, different federal agency, different funding code (ALN 93.676), and a population even further from public view than foster youth.
Investigation data visualization
Two names from this investigation's cast appear in our ledgers as major recipients:
Rite of Passage, Inc.— headquartered in Minden, Nevada — holds at least $177,667,396in ORR Unaccompanied Children awards across five grants (FY2017–FY2025), per our deduplicated tally of TAGGS obligation records. We validated the method award-by-award: our deduplicated total for award 90ZU0350 comes to $35,219,349, against USA Spending. gov's independently published total of $35,219,346.89 — a match to rounding, on an award whose federal description reads, verbatim, "RESIDENTIAL SHELTER FOR UNACCOMPANIED ALIEN CHILDREN." Within our tracked data, Rite of Passage is the sole Nevada recipient of Unaccompanied Children money — one hundred percent of the state's tracked UC stream.
Rite of Passage's operational record, from public primary sources: in the Bureau of Justice Statistics' 2008–09 National Survey of Youth in Custody, its Silver State Academy registered17.7 percent of youth reporting sexual victimization — all of it staff sexual misconduct, against a 10.3 percent national average for staff misconduct, roughly 1.7 times the national figure. (Plaintiff-side framing of this statistic as "three times the national average" overstates what the primary tables show, and we decline to repeat it; the facility does not appear in the high-rate tables of the 2012 or 2018 survey rounds.) Colorado moved to revoke the license of the company's Ridge View campus in June 2021 after finding roughly 100 potential violations; Nevada pulled teenagers from Silver State Academy after a March 2015 riot; and the company sued Nevada over the closure of its Red Rock Academy. Facts of record, from the agencies and courts that made them.
The Devereux Foundation— the Pennsylvania nonprofit from the Senate's four-operator investigation, the operator of the Red Hook facility where a child died 4.5 miles from his falsified bed-check — holds at least$126,260,490in ORR Unaccompanied Children awards across four grants (FY2019–FY2025), by the same deduplicated method; an earlier USA Spending snapshot showing $95.7 million obligated on four awards is consistent with our figure as a floor.
A transparency note our own SKEPTIC review requires us to publish: raw sums over our TAGGS tables for these entities run higher (~$351M and ~$251M respectively) but contain exact-duplicate ingest rows; the figures above are the deduplicated, citable floors, and we corrected our own earlier working dossier to match. TAGGS records obligations — money committed — not audited final expenditures.
The philanthropy that isn't, and the money you cannot see at all
Two more findings from our tables complete the picture.
First: Devereux's visible "private philanthropy" is mostly its own money moving between its own pockets. Of roughly $18.1 million in foundation grants to Devereux entities in our Form 990 Schedule I data,$15.35 million came from The Helena Devereux Foundation— an affiliated foundation granting to its own operating organization. External charitable support is a rounding error against a half-billion-dollar Medicaid-fed budget.
Second — and this is the structural finding that should outlast every individual name in this article: Sequel, Vivant, Acadia, and UHS appear nowhere in the federal grant ledgers at all. Not because they receive no public money — Part II documented billions — but because per-diem congregate revenue flows through Medicaid claims and state contracts, which no public grant-transparency ledger captures. The industry's primary revenue stream is structurally invisible to every federal grant database that exists. The ORR awards above are visible only because that particular program moves money as grants. The foster-care per-diem economy — the $6.2 billion behavioral division, the 53.9-percent-Medicaid revenue base, the $275-to-$800 daily rates — leaves no entry in any ledger a citizen can search. When we say this industry is unaudited, we mean it literally: the money cannot even be seen, let alone examined.
PART VII — THE DOOR: WHERE THE CONGREGATE MODEL MEETS THE MISSING
This is the fourth investigation in our series, and it connects directly to the third. Our § 671(a)(35) investigation documented the nationwide failure to report children missing from foster care to law enforcement and NCMEC within the 24 hours federal law commands. Here is where the two systems meet: the understaffed congregate facility is a machine for producing missing children.
The federal numbers, from our companion investigation's primary sources:4,831 children — about one percent of the national foster-care population — were on missing or runaway status on a single day, September 30, 2020, per ACF data cited by the HHS Office of Inspector General. Roughly 30 percent of foster youth aged twelve and older have run from care at least once. And the pediatric and OIG literature is direct about where running happens youth in foster homes, especially with relatives, are less likely to run than those placed in residential care.
Now put the Senate's findings beside that: a facility where at least one child eloped every month of 2022. Three teenagers walking past sleeping staff, two recovered from the woods ten days later. A boy dead on a highway at 4:50 a.m. while his bed-checks read "present." Understaffing is the elopement mechanism — and Part II established that understaffing is not a failure of the per-diem model but its profit strategy." You can make money in this business if you control staffing."
What happens after the child walks out is the subject of our companion piece, but two figures belong here. NCMEC reports that17 percent of children reported missing from care fall into NCMEC's own classification category "likely child sex trafficking victims" — that phrase is the organization's official label, quoted, not our inference. And the HHS OIG found that in five audited states, there was no evidence that many foster youth who returned from missing episodes were ever screened for trafficking; in Missouri alone, of 978 children who went missing from care in 2019, nearly half were never reported to law enforcement or NCMEC at all.
And recall, from our companion investigation, what the law demands at the moment the child clears the fence line: 42 U.S.C. § 671(a)(35) requires the state agency to report a missing foster child to law enforcement and NCMEC "immediately, and in no case later than 24 hours." The Red Hook case shows what that clock means in hours a person can feel. The bed-checks were falsified; the child was logged present all night; he was discovered missing at 6:45 a.m. and had been dead since 4:50 a.m. The reporting statute never even got the chance to fail — the facility's own falsified paperwork consumed the window in which reporting could have mattered. A reporting deadline is only as good as the staffing that notices the empty bed, and the staffing is set by the per-diem margin. That is the whole bridge between these two investigations, compressed into one night in New York.
One caveat we owe you, per our verification rules: no public NCMEC breakdown exists of missing-from-care episodes by placement type — group facility versus foster home. The national link between congregate placement and elevated running is established in the OIG and pediatric literature; a facility-type census of missing episodes is exactly the data a functioning federal audit regime would produce, and does not.
So follow one hypothetical child through the whole machine this series has mapped: placed in congregate care under a QRTP certificate no one audits; held past fourteen days on a paperwork exemption; watched by a staff thinned to protect a per-diem margin; walks out an unwatched door like the children at Millcreek and Cedar Ridge; is reported late or never under § 671(a)(35), a statute with a zero-dollar enforcement history; and if she returns, may never be screened for what happened to her while she was gone. At every single stage, the failure is documented by a federal or state body — and at every single stage, the money keeps flowing.
PART VIII — WASHINGTON: A STUDY, A DEAD BILL, AND A REFERRAL GONE QUIET
Investigation data visualization
What has the federal government done with all of this?
A study.The Stop Institutional Child Abuse Act — Public Law 118-194, the bill Paris Hilton championed to passage in December 2024 — commissions work from the National Academies of Sciences, Engineering, and Medicine, with a first report expected around late 2026. It creates no licensing standard, no audit, no penalty. It is a study — a necessary one, and only that.
A dead letter, so far.The BRIDGES for Kids Act, introduced by Senator Ron Wyden in December 2025, is the one bill that aims at the machinery this article has described: a public national dashboard of facilities and their records, two-day federal complaint investigations, and — directly at Part IV — closure of the twenty-one-state loophole that lets a paid accreditation substitute for state licensing inspection. It has not passed.
A bill running the other direction.H.R. 8095 would strip the Medicaid "IMD exclusion" — the sixteen-bed cap that is currently the last federal limit on institutional scale — for foster children in QRTPs. Whatever its sponsors intend, its mechanical effect would be to make the largest institutions fully Medicaid-billable for the one population least able to walk away. For the operators of Part II, whose CFOs describe occupancy as their most significant opportunity, it is difficult to design a more welcome amendment.
A referral, twenty-one months old.On October 9, 2024, Senator Wyden referred UHS, Acadia, Devereux, and Vivant to the Department of Justice, urging investigation under a "worthless services" False Claims Act theory — the argument that billing Medicaid and IV-E for "treatment" delivered under the conditions his committee documented is billing for something worthless. As of this writing,twenty-one months later, there is no public disposition.
The False Claims Act record that already exists suggests the theory is not exotic. UHS paid$122 million in 2020 to resolve eighteen qui tam cases alleging, among other things, medically unnecessary admissions and improper holds in its behavioral facilities — the whistleblowers shared $15,862,457. Acadia paid $19.85 million in September 2024 ($16.6 million federal, the rest to Florida, Georgia, Michigan, and Nevada) over admissions and length-of-stay practices, and in May 2026 a federal court in Tennessee gave final approval to a$179 million securities settlement over Acadia's admissions and patient-holding practices. Alabama's Medicaid program alone paid Sequel roughly$25 million between 2018 and August 2020, within more than $68 million in state contracts since 2016 — payments that continued while the state's own inspectors documented abuse.
The pattern across every settlement: the money is recovered years later, in fractions, from shareholders and insurers — and no settlement has ever turned on the question a child would ask, which is whether anyone checked that the "treatment" was real before the fourteen-day clock was switched off.
EPILOGUE: AUGUST 16
Sometime before August 16, 2026, the last child will leave Provo Canyon School. Utah's order gives the operator until that date to end all services; UHS has fifteen days to appeal. Paris Hilton, who told Congress what the inside of that building was like, called the revocation what it is — the end of one institution, twenty-six years into its ownership by a company whose behavioral division booked $6.2 billion the year the Senate report landed.
But this article has not really been about Provo Canyon, and the fourteen-day rule's authors were not really writing about any single facility. Somewhere on the day you read this, per the states' own answers to the GAO: a foster child is sleeping in a caseworker's office or a hotel room in one of forty-two states, because the placement system has no bed for her. Roughly fifteen states have not one facility meeting the federal quality bar — and in at least one state, every congregate facility has been paper-designated a trafficking-serving setting so the federal money never stops. A private certificate, purchased for less than the revenue of one bed-week, is standing in for public inspection in twenty-one states. A Senate criminal referral is entering its twenty-second month of silence. And the industry's principal revenue stream is invisible to every public ledger in existence.
The children in this system did not choose their placements, cannot refuse them, and — as Utah's own revocation order documented — are sometimes blocked from reporting what happens inside. They are the least free people in American public life, and their custody has been made a yield-bearing asset. The fourteen-day rule was Congress's attempt to sayenough; the QRTP certificate is the industry's answer; the GAO survey is the scoreboard; and the scoreboard says the certificate is winning.
Utah just proved a state can read its own inspection files and act. There are forty-nine more states, roughly fifteen of them with zero qualifying facilities and forty-two of them with children sleeping in offices — and one federal agency, asked by its own auditor merely toclarifythe loophole's widest door, that answered no.
The paper trail is public. All of it. The statute, the Senate report, the GAO survey, the settlements, the license revocations, the grant ledgers. Pull up a chair. Pull up the file.
The pending NEICE mandate is often presented as a solution. NEICE, the National Electronic Interstate Compact Enterprise, is designed to modernize ICPC transmission and reduce paper-based delays. Digital transmission matters, and full implementation is better than fax machines, mail bins, and scattered local processes. But digitizing a broken process is not the same as fixing it. If the underlying compact still lacks enforcement, penalties, staffing requirements, family remedies, and real-time accountability, then NEICE may simply move files more efficiently into the same queues. A digital bottleneck is still a bottleneck.
The ICPC crisis should force a much bigger question: why does the system make it so hard to place children with family while making it so easy to keep them in paid care? The answer is not one villain. It is design failure. It is legacy law. It is federalism without accountability. It is underfunded administration. It is risk avoidance. It is provider economics. It is the absence of a private right of action. It is decades of reform language without enforcement teeth. It is the oldest story in child welfare: the state says family matters, then builds a process that treats family as a paperwork problem.
A serious reform agenda would begin with enforceable deadlines. The 60-day home study requirement should not be a polite suggestion. If a receiving state misses the deadline without documented good cause, there should be automatic escalation, public reporting, and financial consequence. If the delay continues, relatives should have a direct administrative appeal and a fast-track court remedy. A compact that can stop a child from moving should also give the child and relative a way to force the government to act.
Second, the ICPC should include a child-centered emergency placement pathway for vetted relatives. When a relative has already passed background checks, has an existing relationship with the child, and meets basic safety criteria, the law should allow provisional placement with supervision while the full home study is completed. The system already takes risks by placing children with strangers in emergency settings. It should not pretend that relatives are uniquely dangerous simply because they live across a state line.
Third, Congress and the states must fund ICPC offices like critical infrastructure. A state with thousands of foster care cases cannot run interstate placement through one or two overburdened staff members and then act shocked when children wait. Staffing ratios should be published. Caseloads should be published. Average processing times should be published. Returned packets should be tracked. Clock restarts should be tracked. Every state should have a public ICPC dashboard showing how many children are waiting, how long they have waited, and where the file is stuck.
Fourth, the law must close the clock manipulation loophole. The 60-day period should not begin only after the receiving state accepts a completed referral while the sending state can take weeks or months to prepare it. There should be deadlines on both sides. The sending state should have a fixed period to identify relatives, prepare the referral, and transmit the packet. The receiving state should have a fixed period to accept, reject, or request specific missing information. If a packet is returned as incomplete, the reason should be documented, time-limited, and reviewable.
Fifth, ICPC delays should trigger placement cost review. If a child remains in congregate care or high-cost foster care solely because of interstate paperwork delay, the state responsible for the delay should bear financial consequence. Right now, the cost of delay is diffused across systems, while the child pays emotionally. Accountability should follow the bottleneck. If a state’s failure to process a home study causes months of unnecessary paid placement, that failure should show up in the budget.
Sixth, federal child welfare reviews should include ICPC compliance as a core performance measure. It is absurd to review state child welfare performance while allowing interstate family placement delays to remain a side issue. If a state cannot move children to safe relatives across state lines within legal timelines, that is a permanency failure. It should affect federal oversight, corrective action plans, and funding scrutiny.
Finally, families need rights, not just promises. Relatives who step forward should receive notice, status updates, reasons for delay, and access to review. Children should have counsel empowered to challenge ICPC delay when it harms permanency. Parents should be able to argue that the state cannot move toward termination while kinship placement is delayed by government inaction. No child should lose family because two state agencies could not move paperwork on time.
The ICPC began as a protective reform. That origin matters because it explains why the system has been so hard to challenge. Nobody wants to be accused of weakening safeguards for children. But after sixty-six years, the question is no longer whether children need protection across state lines. Of course they do. The question is whether the existing compact protects them from harm or creates a new kind of harm through delay, drift, and family separation. A child left in congregate care for six months while a safe relative waits is not being protected by paperwork. That child is being failed by it.
The phrase “fall through the cracks” is too soft for what is happening here. Children are not falling through cracks. They are being held in a pipeline with known failure points, known delays, known staffing shortages, known enforcement gaps, and known financial consequences. When a system knows where children get stuck and still does not build a release valve, the problem is not accidental. It is tolerated.
This is why the ICPC crisis deserves national attention. It is not as visible as a courtroom termination. It is not as emotionally packaged as an adoption ceremony. It is not as politically explosive as a removal caught on video. But it may be one of the most consequential hidden mechanisms in the child welfare system because it decides whether family can actually function as family when state lines are involved. In a mobile country where relatives often live in different states, a slow interstate placement system is a family-separation engine.
The children caught in this engine do not need another conference panel. They do not need another best-practices memo. They do not need another voluntary compact reform stalled for twenty years. They need enforceable timelines, funded staffing, transparent dashboards, provisional kinship pathways, legal remedies, and consequences for states that ignore deadlines. They need a system that treats family placement as urgent, not optional.
The ICPC was built to stop children from being sent across state lines without protection. But in 2026, the greater danger is that children are being kept from family without accountability. A compact designed to protect children has become a maze that delays them. A law built for paper files is governing digital-age families. A process created to ensure safety is too often producing separation.
And the most damning fact remains the simplest: when states miss the deadline, nothing happens to the states. Something happens only to the child.
If you enjoyed this work and feel encouraged by independent investigations like this, please consider becoming a paid subscriber to The Constitutional Republic on Substack or donate at
Project Milk Carton Inc. Your gift keeps our site alive, spotlighting high-risk children!
Your support helps us continue:
investigative reporting
public records research
educational transparency projects
community outreach
and future investigations like this series
Most importantly, please share this article.
Independent journalism survives when ordinary people help spread information powerful institutions would prefer remain ignored, misunderstood, or buried beneath complexity.
Because silence protects systems.
Informed communities protect children.
ALL paid subscriptions are tax deductible Project Milk Carton | 501(c)(3) | EIN: 33-1323547
SOURCES AND VERIFICATION NOTES
Primary documents:42 U.S.C. §§ 672(k), 675a(c) (Family First Prevention Services Act, P.L. 115-123); Senate Finance Committee,Warehouses of Neglect(June 12, 2024) — page-level citations throughout Parts II–IV; GAO-26-107592 (March 3, 2026) — all fifty-state survey figures in Part V; DOJ press releases (UHS 2020 FCA settlement; Acadia 2024); Robbins Geller release on the Acadia $179M securities settlement (final approval May 2026); Utah DHHS license revocations of July 6 and July 17, 2026 (via ABC4, Deseret News, Utah News Dispatch, CBS News, The Imprint); NBC News investigation of Sequel (canonical NBC URL now returns an error; identical text verified via Yahoo syndication — we cite NBC and note both); APM Reports (May 21, 2021) on Sequel's finances and collapse; Private Equity Stakeholder Project,The Kids Are Not Alright; BJS National Survey of Youth in Custody (2008–09); HHS OIG missing-from-care audits and ACF data (Part VII, shared with our § 671(a)(35) investigation); AFCARS via ProPublica; WV Watch (Feb 2026); Houston Public Media (Apr 2026).
Project Milk Carton original data (Part VI):TAGGS-derived federal grant obligation tables and Form 990 Schedule I tables in our research database, queried 2026-07-19; deduplicated totals cross-validated against USASpending.gov award pages (validation match documented in text). TAGGS figures are obligations, not audited expenditures; stated totals are floors.
Figures we declined to use:the "three times the national average" characterization of Silver State Academy's BJS numbers (primary tables support ≈1.5–1.7×, and we say so); the Council on Accreditation's fee amounts (no accessible primary document); any list naming the GAO's zero-QRTP states (GAO does not name them; circulating lists trace to a 2021 secondary source); aggregate values of Rite of Passage's state contracts (unverified); and any NCMEC breakdown of missing episodes by placement type (none exists — which is itself a finding).
Project Milk Carton is a 501(c)(3) organization producing child welfare transparency research from public data and primary documents. Where our database findings concern funding flows, they document funding structure and are not allegations of unlawful conduct by any recipient.
Project Milk Carton
501(c)(3) Nonprofit • projectmilkcarton.org • t.me/ProjectMilkCarton