OSINT · Publicly Available Sources INVESTIGATION LEDGER

The 30-Day Notice Nobody Checks: How HHS Built a Federal Kinship Mandate With No Data, No Audit, and No Penalty

August 15, 2026 OPUS · Claude Opus Project Milk Carton

The 30-Day Notice Nobody Checks: How HHS Built a Federal Kinship Mandate With No Data, No Audit, and No Penalty

I have a solid evidence base. Writing the report now.

The 30-Day Notice Nobody Checks: How HHS Built a Federal Kinship Mandate With No Data, No Audit, and No Penalty

Since October 2008, federal law has required every state child welfare agency to find and notify a removed child's grandparents and adult relatives within 30 days — and to tell them, in writing, that the child is in state custody and that they have the option to take that child in. It is one of the least ambiguous commands in Title IV-E of the Social Security Act. It is also, functionally, unenforced. The federal government does not collect a single data element measuring whether the notice was sent. The only Title IV-E review that carries a financial disallowance does not examine notice at all. States self-certify compliance in annual narrative reports that no one samples or verifies. And because Congress wrote the mandate as a duty owed by the state rather than a right held by a family, the Supreme Court's § 1983 jurisprudence has largely foreclosed grandparents from suing over it. The result is a federal requirement with three enforcement doors — audit, data, litigation — and all three are welded shut.

What the Statute Actually Commands

The Fostering Connections to Success and Increasing Adoptions Act of 2008 (P.L. 110-351, originally H.R. 6893) added paragraph (29) to Section 471(a) of the Social Security Act, codified at 42 U.S.C. § 671(a)(29). (Both citations describe the same provision; SSA § 471 is 42 U.S.C. § 671 — a point that trips up even search engines.)

The text is specific. Within 30 days after a child is removed from the custody of the parent, the state "shall exercise due diligence to identify and provide notice to" three categories of people: all adult grandparents; all parents of a sibling of the child where that parent has legal custody of the sibling; and other adult relatives of the child, including any other adult relatives suggested by the parents — subject to exceptions due to family or domestic violence.

The notice is not a courtesy call. The statute prescribes its contents: it must specify that the child has been or is being removed; explain the options the relative has under federal, state, and local law to participate in the care and placement of the child, including any options that may be lost by failing to respond; describe the requirements to become a licensed foster family home and the services and supports available to children placed in such a home; and describe how the relative could enter a kinship guardianship assistance arrangement.

That third and fourth element matter enormously, because they are the difference between a grandmother becoming a paid, supported, licensed caregiver with Medicaid access and casework services, and a grandmother taking a child under the table with nothing. The statute anticipated that distinction. The implementation did not.

States, on paper, went further than Congress required. Missouri's Children's Division manual directs that grandparents be sought and contacted within three hours of removal. Connecticut requires "best efforts" to identify and locate all grandparents within 15 days. Los Angeles County DCFS maintains a dedicated Due Diligence procedure (0300-306.75); Georgia codified Diligent Search at PAMMS 19.20. The policy language is excellent almost everywhere. The question this investigation asks is whether anyone in the federal government has ever checked whether the policy is followed. The answer is essentially no.

The Audit That Looks Everywhere Except Here

The federal government does audit Title IV-E. It audits it aggressively, with real money at stake — and it audits the wrong thing.

Under 45 C.F.R. § 1356.71, ACF conducts Title IV-E Foster Care Eligibility Reviews. The mechanics are precise: a primary review samples 80 cases. Substantial compliance means no more than eight error cases in an initial primary review, dropping to four or fewer in every subsequent review. If a state fails, ACF assesses a disallowance covering the entire amount and period of maintenance payments and associated administrative costs for the error cases — and the state must execute a program improvement plan before a secondary review, where substantial compliance requires a case-ineligibility or dollar error rate of 10 percent or less.

That is a serious audit with teeth. But what it examines is whether the child was eligible and whether the provider was properly licensed and safety-checked: the judicial determinations of "contrary to the welfare" and "reasonable efforts," the AFDC-relatedness test, the criminal background checks. Relative notice under § 671(a)(29) is not an eligibility element. It cannot generate an error case. It cannot generate a disallowance. It is structurally invisible to the only federal review that costs a state money.

The Child and Family Services Reviews (CFSRs), now in Round 4, are the other federal oversight instrument. They measure seven outcomes and seven systemic factors through case file review and stakeholder interviews. Relative placement surfaces indirectly — in permanency items on placement stability and preservation of connections — but there is no CFSR item that asks whether the 30-day notice was issued to identified grandparents. And CFSR consequences are slow and rare: a state found out of substantial conformity enters a program improvement plan, and only a failed PIP triggers withholding of IV-B and IV-E funds.

Then there is the Annual Progress and Services Report. Under 45 C.F.R. § 1357.16(b), every state must submit an APSR with the CFS-101 forms to its ACF Regional Office by June 30 each year as a condition of receiving Title IV-B, CAPTA State Grant, and Chafee/ETV funding. The APSR is a narrative self-report. States describe their own progress toward their own goals. There is no sampling protocol, no independent verification, no error-rate threshold, no disallowance. A state can assert that it exercises due diligence to notify relatives, and that assertion becomes the federal record.

HHS's own Office of Inspector General has demonstrated it can dig into Title IV-E case files when it wants to. OIG has repeatedly audited IV-E administrative cost claims — Virginia on behalf of Fairfax County and Arlington County (both 2007), Ohio's claims for delinquent youth (2008), Georgia (2010), Los Angeles County (2011). Every one of those audits was about dollars improperly claimed. None was about whether a grandparent was ever told the child existed in the system.

The Data Element That Doesn't Exist

The most damning fact in this investigation is a comparison inside a single federal database.

The Adoption and Foster Care Analysis and Reporting System (AFCARS) is the mandatory federal data collection every Title IV-E agency must submit. The 2020 AFCARS final rule stripped a great deal out of the 2016 rule — eliminating more than 90 percent of the 60-plus Indian Child Welfare Act data elements, nearly all data on LGBTQ children, youth and parents, most health care tracking, and some education elements. Litigation followed; the Protect ICWA Campaign backed a suit challenging the withdrawal.

But look at what survived the cut. Under the 2020 rule, agencies must still report whether the child is an Indian child as defined by ICWA, tribal membership, which tribe — and whether the tribe was sent legal notice.

So the federal government maintains, right now, a national data element tracking whether tribal notice was given. It maintains no equivalent element tracking whether the § 671(a)(29) relative notice was given. HHS can produce a count of ICWA notices. It cannot produce a count of grandparent notices, because it has never asked for one. Twenty-eight years of AFCARS revisions, eighteen years after Fostering Connections, and the compliance rate with a mandatory federal statute is not merely poor — it is unknown and unknowable at the federal level.

That is not an oversight failure. It is an architectural choice. You cannot penalize what you do not count, and you do not have to count what you would rather not have to penalize.

The Courthouse Door

The third possible enforcement path is private litigation, and the Supreme Court has spent thirty years narrowing it.

In Suter v. Artist M., 503 U.S. 347 (1992), the Court held that § 671(a)(15) — the "reasonable efforts" requirement, sitting in the same subsection as the notice mandate — conferred no enforceable private right on its beneficiaries and created no implied cause of action. Congress pushed back with the so-called "Suter fix," 42 U.S.C. § 1320a-2, providing that a Social Security Act provision is not unenforceable merely because it appears in a section specifying the contents of a state plan. But the fix was deliberately narrow: it expressly declined to alter the Suter holding as to § 671(a)(15), and it disclaimed any intent to expand the grounds for private enforcement beyond overturning Suter's specific reasoning.

Then the Court tightened further. Blessing v. Freestone, 520 U.S. 329 (1997) held that Title IV-D gives individuals no federal right to compel a state agency's substantial compliance, and that a § 1983 plaintiff must assert violation of a federal right, not merely of federal law. Gonzaga University v. Doe, 536 U.S. 273 (2002) completed the doctrine: Congress must speak in "clear and unambiguous" rights-creating terms.

Now read § 671(a)(29) against that standard. It is phrased as an obligation on the sovereign — "the State shall exercise due diligence" — with no rights-creating language, no reference to individual entitlement, no "no person shall be denied." It is a funding condition addressed to the state, embedded in a state plan section, aggregate in focus. It matches, almost point for point, the profile of provisions federal courts decline to enforce under § 1983.

The practical consequence: a grandmother who learns three years late that her grandson was removed, warehoused with strangers, and adopted out has, in most circuits, no federal cause of action for the notice she never received.

Where families have won, they have had to win sideways. In November 2021, the Legal Aid Society and Dechert LLP filed a class action against New York State and New York City over practices that deny children foster placement with relatives — specifically a statutory bar disqualifying kin who have any of nearly 300 mandatory excluding crimes, regardless of how old or minor the conviction, and even where city officials themselves believe the relative can provide a safe home. A federal district court dismissed. The Second Circuit revived it, finding that "the plaintiffs have suffered a real-world harm" and reinstating claims for eleven remaining plaintiffs. The pleaded harms are the whole indictment: some children "grew up with strangers," exposed "to risks of psychological and emotional harms"; others landed in informal relative arrangements that deprived them of "medical and social services provided to children in foster care."

That case had to be framed as a substantive due process claim about family integrity — not as a notice claim — precisely because the notice statute is not privately enforceable.

State courts occasionally catch it. In an Iowa Court of Appeals decision (No. 15-2022, Feb. 24, 2016), the court found the department had breached its duty to notify relatives of pending proceedings, overlooked licensing violations in the foster parent's in-home daycare, and ignored a court order authorizing consideration of relative placements. That grandparent got a hearing. She got it because of a state judge's discretion, not because of any federal mechanism.

The Money: Why the Incentive Runs Against the Grandmother

Federal foster care is a large, administratively top-heavy program. In FY2023, total Title IV-E spending was $9.5 billion — $5.1 billion reimbursing states for foster care, $4.3 billion for adoption and guardianship assistance, and $172 million for services.

HHS's own policy shop documented the structural problem. The ASPE brief "Administrative Costs Drive Foster Care Claims and Are Increasing Per Child as Caseloads Decline: Trends in Title IV-E Foster Care Expenditures FY2019–FY2023" found that administrative costs — not payments to caregivers — have been the largest driver of Title IV-E foster care expenditures, and that per-child administrative cost is rising even as the number of children in care falls. The system's dollars increasingly flow to the machinery of placement rather than to the people raising children.

Now compare what the two kinds of caregivers receive. Policy Matters Ohio documented the gap in stark terms: an approved kinship caregiver in Ohio receiving an Ohio Works First TANF child-only grant gets $302 per month for one child, or $412 for two siblings. A licensed foster parent in Scioto County receives a minimum of $836 per month and a maximum of $4,258 per month for a single child.

That spread — up to fourteen times — is the entire economics of this story. Nationally, roughly 13 states extend full foster care maintenance payments to unlicensed kinship caregivers, and about 20 more have simplified relative approval; in the rest, unlicensed kin typically receive only TANF child-only grants, not foster care rates. AFCARS data indicate that in FY2024 about 30 percent of children in foster care were placed with relatives; broader counts including fictive kin put the 2025 figure near 39 percent (roughly 128,000 children) — of whom about 44 percent (some 56,665 children) were with unlicensed caregivers, meaning the lower payment tier.

The counterparty is a large and well-capitalized private placement sector. National Youth Advocate Program, Inc. appears in the PMC CivicOps federal awards database with roughly $1.79 billion in cumulative federal awards traced — including single obligations of $121.5 million (FY2023) and $71.99 million under CFDA 93.676, and $113.9 million and $95.6 million under CFDA 93.566. Those particular dollars are Unaccompanied Children and Refugee & Entrant Assistance funds, not domestic Title IV-E — but they illustrate the scale of the contracting infrastructure that stands ready to receive placed children. KVC Health Systems, Inc. (EIN 26-2516589, Olathe, Kansas), the parent entity of one of the country's larger privatized child welfare operations, reported $16.25 million in revenue and $18.55 million in expenses for FY2023 at the holding-company level alone, with aggregated officer, director and trustee compensation of $1,373,854 in its most recent reported year, plus a $1,995,551 HHS award in FY2024 under CFDA 93.493. Kansas appellate courts have entertained direct challenges to "the reasonableness of the efforts made by the agencies involved," including argument about "ongoing problems with DCF/KVC having the resources" (In re Interests of M.S., Kan. Ct. App. No. 119797, June 21, 2019).

Nobody in this chain has a financial reason to find a grandmother in 30 days. The relative is unpaid or underpaid, generates no per diem, requires no case management contract, and — critically — is the cheapest outcome for the federal treasury and the least remunerative for everyone in the placement business. The notice requirement asks a system to spend labor hunting for the option that pays it least.

What the Audits Do Find When Someone Looks

The rare instances where an auditor examined relative placement practice suggest the notice compliance rate would be dismal if anyone measured it.

The California State Auditor, reviewing Los Angeles County DCFS, found that when placing children with relatives the department failed to conduct required pre-placement criminal background checks in 77 percent of 30 cases reviewed and failed to conduct pre-placement home inspections 27 percent of the time. Drilling into 22 relative placements specifically: in-home inspections were completed before placement in only 16, and required pre-placement background checks were documented in only 5. The same audit found DCFS completed just 72 percent of safety assessments and 76 percent of risk assessments on time, and never completed 10 percent of safety assessments and 8 percent of risk assessments at all.

The Michigan Office of the Auditor General found investigators failed to document required Central Registry clearances for 262 individuals across 112 investigations (72 percent), with documentation too thin to complete a CPS history review for another 59 individuals (39 percent).

These are documentation failures in areas that are audited. Relative notice is documented in the same case files by the same overloaded caseworkers — and audited by no one.

And when the story does surface, it surfaces through journalism. KING5's investigative unit in Washington reported the case of Doug and AnneMarie Stuth of Enumclaw, grandparents of an infant girl, in which state social workers pushed for adoption by the foster mother on bonding grounds despite state law making relatives the first choice. A high-ranking DSHS official wrote in an internal email: "If we don't (place the child) with a relative there will be a lot of explaining to do." The state ombudsman's office, KING5 reported, investigates dozens of relative custody complaints every year.

Note the escape hatch the statute itself provides. The domestic violence exception is real and necessary — but in practice it becomes documented discretion. Typical state policy reads: "A supervisor shall be consulted and must approve any decision to not notify a grandparent or relative, and the specific safety concern and justification for not notifying shall be documented in the electronic case record." That justification is written into a file that no federal reviewer will ever read, and reported in aggregate to no one.

Who Is Supposed to Be Watching

The Children's Bureau within HHS/ACF administers Title IV-E, runs the CFSRs and eligibility reviews, and receives every APSR. It writes the AFCARS data elements. It could add a notice element tomorrow. It has not.

GAO has circled the kinship problem for a quarter century — from Foster Care: Kinship Care Quality and Permanency Issues (HEHS-99-32) through GAO-20-434 (HHS Could Enhance Support for Grandparents and Other Relative Caregivers), GAO-23-105624 (HHS Is Taking Steps to Help States Support Relative Caregivers with Evidence-Based Programs), and GAO-26-107658 (Relatives Raising Children: Federal Support Helps, but Challenges Persist, December 2025). GAO's core finding is a scandal in itself: roughly 2.7 million American children are being raised by relative caregivers, and fewer than 1 percent of them are in a formal foster care placement. Those families are more likely than the general population to be in poverty; grandparent-headed households are more likely to include a disability and to be out of the labor force. That 99 percent is the informal-care sink — kin raising children with no maintenance payment, no Medicaid pathway through foster care, no casework, no respite. Many of them are precisely the relatives who should have received a § 671(a)(29) notice explaining how to become a licensed, supported placement.

Congress has legislated around the edges. Senators Bob Casey (D-PA) and Todd Young (R-IN) introduced the bipartisan Informing Grandfamilies Act, which would require states to communicate directly with grandparents applying for or receiving TANF cash assistance and inform them of SNAP, LIHEAP and Medicaid, and to staff up expertise to guide kinship families through benefits. The Senate Special Committee on Aging under Casey and Ranking Member Tim Scott advanced the broader Grandfamilies Act. These are good bills. None of them touches § 671(a)(29). Congress's instinct has been to add a new notice duty rather than enforce the one already on the books — which is itself a tell.

ACF's one real structural reform is stalling. The September 28, 2023 final rule (implemented through PI-23-10) finally allowed Title IV-E agencies to adopt separate, simplified licensing or approval standards for relative and kinship foster homes — the single change that would let a notified grandmother actually get paid at the foster care rate. As of February 2026, only 24 eligible jurisdictions (19 states and five Tribes) had submitted plan amendments — meaning nearly two-thirds have not adopted separate kin standards. Eighteen state plans were approved as of January 2026, including Colorado, Illinois, Iowa, Kansas, Kentucky, Michigan, Nevada, Oklahoma, Oregon, Texas, Washington and Wisconsin. Separately, as of January 2026, fewer than 20 percent of jurisdictions (11 states and one territory) had a formally approved plan to operate a kinship navigator program. HHS made the fix optional, and most states opted out.

Why It Matters, and the Fix

The research consensus is not in dispute: children placed with kin experience fewer placement moves, are more likely to stay with siblings and remain in their school and community, and show better behavioral and mental health outcomes than children placed with unrelated foster parents. HHS itself calls kinship care "often the recommended option." The notice requirement is the on-ramp to that outcome. When it fails, the failure is invisible and permanent — a child spends a childhood with strangers, and a grandmother finds out too late that there was a form she was supposed to receive.

Four changes would close this, in descending order of leverage:

1. Make the court do it. State courts already must make "contrary to the welfare" and "reasonable efforts" judicial determinations as a condition of Title IV-E claimability. Add a third: a judicial finding at the first permanency-relevant hearing that the agency issued § 671(a)(29) notice, identifying who was noticed and who was excepted and why. This converts an unaudited clerical duty into a finding that gates federal money — and it uses machinery that already exists in every juvenile court in the country.

2. Count it. Add a relative-notice data element to AFCARS, structurally identical to the existing ICWA tribal-notice element. This costs essentially nothing, requires no legislation, and makes national compliance measurable for the first time since 2008.

3. Audit it. Add a notice item to the CFSR case review instrument, and — the version with teeth — make notice an element in the Title IV-E Foster Care Eligibility Review, so a missing notice becomes an error case subject to disallowance under 45 C.F.R. § 1356.71.

4. Make the notice worth answering. A notice telling a grandmother she may take a child with $302 a month and no support, while a stranger down the road receives $836 to $4,258, is a notice engineered to be declined. Congress should require — not permit — Title IV-E agencies to adopt kin-specific licensing standards under the 2023 rule and to extend full foster care maintenance payments to approved kin.

And if Congress wants families to be able to enforce this themselves, it must amend § 671(a)(29) in the rights-creating language Gonzaga demands. Right now the statute reads like a memo to a bureaucracy. Rewritten as a right held by the child and the relative, it would read like a law.

Until then, the 30-day notice remains what it has been for eighteen years: a mandatory federal requirement whose compliance rate is not merely low, but unmeasured — by design.

Sources:

A note on two figures I deliberately hedged: the FY2024 "30 percent placed with relatives" and the 2025 "39 percent relative/kin" counts come from different definitions (blood relatives vs. relatives plus fictive kin), so I presented them as parallel measures rather than as year-over-year growth. And the National Youth Advocate Program award totals are Unaccompanied Children and Refugee program dollars (CFDA 93.676 and 93.566), not Title IV-E — I labeled them as sector-scale context rather than foster care spending, since conflating them would be the kind of error a hostile reader would use to discredit the whole piece.