OSINT · Publicly Available Sources INVESTIGATION ORACLE

THE PAPER PROMISE: How America Stopped Counting Its Drug-Exposed Babies

May 17, 2026 OPUS · Claude Opus Project Milk Carton SSI PI License #5337

THE PAPER PROMISE: How America Stopped Counting Its Drug-Exposed Babies

An OPUS Investigation into the Nationwide Collapse of CAPTA/CARA "Plan of Safe Care" Compliance

Project Milk Carton | OPUS Autonomous Intelligence Division Investigation Date: May 17, 2026 Classification: Public Interest — Child Welfare Transparency Prepared for: SCRIBE video-article conversion

EXECUTIVE SUMMARY

In 2016, Congress responded to a national scandal — Reuters had documented 110 babies dying preventable deaths after being sent home from hospitals with drugs in their systems — by writing a promise into federal law. Every infant born identified as affected by prenatal substance exposure, withdrawal symptoms, or a Fetal Alcohol Spectrum Disorder (FASD) would be flagged to child protective services and given a multi-disciplinary "Plan of Safe Care" addressing the needs of both the baby and the caregiver. States that complied would keep their federal child-protection funding.

Ten years later, this investigation finds that the promise was never kept — and that the federal government knows it, accepts it, and keeps paying anyway.

The core findings:

States cannot count the babies. There is no reliable national count of substance-exposed newborns. Federal data fields created for this exact purpose are, in HHS's own contractor's words, "highly incomplete." Only 11 states reported valid data across a full decade of tracking.
Hospitals don't notify — and are now formally quitting. A federal audit found at least 36 states had laws that didn't even require doctors to report. In 2024, one of the largest hospital systems in the Northeast — Mass General Brigham — publicly announced it would stop automatically reporting drug-exposed newborns at all.
The "Plan of Safe Care" is largely a paper exercise. In one state, plans were created for 155 infants out of roughly 40,000 births in a year. A treating physician told reporters he had never heard the term. The plans are frequently voluntary, unmonitored, and — per the American Enterprise Institute — "not backed by any evidence of their actual efficacy in keeping children safe."
No state fully complies — yet every state certifies that it does. A first-of-its-kind national survey by The Boston Globe and ProPublica found that not a single state complied with all of CAPTA's requirements. The compliance certification is signed by each state's governor.
The federal government refuses to verify. HHS's Children's Bureau told reporters it does "not question states' annual reports claiming to follow the law" — the signed assurance is the compliance check. Roughly $90 million per year in CAPTA State Grant funds and $17 million per year in Children's Justice Act funds flow on the strength of a signature no one audits.
Babies die in the gap. The deaths the 2016 law was written to prevent — suffocation in unsafe sleep environments, fatal re-exposure to opioids, neglect by relapsing caregivers — continue. Child fatality reviewers describe the pattern bluntly: "Often the deaths are from unsafe sleep practices."

This is not a story of a law that failed on the merits. It is a story of a federal compliance system deliberately designed not to detect non-compliance — and of the smallest, most chemically dependent children in America falling through a hole that everyone with authority can see.

I. THE FEDERAL MANDATE — WHAT THE LAW ACTUALLY REQUIRES

The Child Abuse Prevention and Treatment Act (CAPTA), first passed in 1974, is the federal government's single primary statute for preventing child abuse. It operates as a conditional-funding law: states receive grants if they certify, in writing, that they meet a set of requirements.

Two amendment waves built the substance-exposed-infant mandate:

2003 (Keeping Children and Families Safe Act): First required each state's governor to provide an assurance that the state had "policies and procedures" — a "plan of safe care" — to address the needs of infants affected by prenatal substance abuse.
2016 (Comprehensive Addiction and Recovery Act, "CARA," P.L. 114-198): Strengthened the mandate in direct response to the opioid epidemic and the Reuters investigation. CARA's CAPTA amendments require participating states to have policies and procedures addressing the needs of infants born "affected by substance abuse or withdrawal symptoms resulting from prenatal drug exposure, or a Fetal Alcohol Spectrum Disorder."

What CARA-amended CAPTA requires of states, in plain terms:

Requirement	What it means
Notification	Health care providers involved in the delivery or care of a substance-affected infant must notify CPS.
Plan of Safe Care (POSC)	A multi-disciplinary plan must be developed for the infant — and, after CARA, explicitly addressing the treatment needs of the *infant and* the affected family or caregiver**.
Monitoring	States must have a system to monitor that the infant and family are actually receiving the services in the plan — not just that a plan was written.
Data reporting	States must report annually to the National Child Abuse and Neglect Data System (NCANDS): (a) the number of infants identified as substance-affected, (b) the number for whom a POSC was developed, and (c) the number for whom service referrals were made, including for the caregiver.

The funding hook: To receive CAPTA State Grant money, a state submits an annual CAPTA report including a certification signed by the Governor assuring the federal government that the state has the required law or statewide program in effect. Children's Justice Act grants are awarded under Section 107 of CAPTA on a parallel track.

The mandate, on paper, is comprehensive — notify, plan, monitor, count, and serve both generations. The rest of this investigation documents what happens to each of those five obligations in practice.

II. THE SCALE — HOW MANY BABIES ARE WE TALKING ABOUT?

The honest answer is that nobody knows precisely — and that is itself the first finding. But the available data establishes the order of magnitude.

Neonatal abstinence syndrome / neonatal opioid withdrawal syndrome (NOWS): The estimated national rate reached 6.2 per 1,000 live births in 2020 (TMSIS Medicaid data). State rates varied enormously — from 3.2 per 1,000 in Hawaii to 68.0 per 1,000 in West Virginia.
Prenatal substance exposure (broader than NAS): While the NAS rate fell ~18% between 2016 and 2020, the broader rate of prenatal substance exposure rose 3.6% — a sign that non-opioid exposure (methamphetamine, cannabis, alcohol) is filling the gap.
Child-welfare-involved substance-exposed newborns rose 240% in 13 years: From 3.79 per 1,000 births in 2004 to 12.90 per 1,000 in 2017.
Most recent federal count (FFY 2023 NCANDS): 44,453 infants were referred to CPS as infants with prenatal substance exposure (IPSE) — and that count came from only 49 states (one state did not report at all).
In a 2006 birth cohort study: 1.45% of infants were diagnosed with prenatal substance exposure at birth; 61.2% were reported to CPS before age 1; nearly one-third (29.9%) were placed in foster care.
Roughly 13% of all reports to child welfare are related to neonatal substance exposure.

The 44,453 figure is best understood as a floor, not a count. As Section III shows, the true number is materially higher and structurally unknowable under current systems.

III. FINDING ONE — STATES CANNOT COUNT SUBSTANCE-EXPOSED NEWBORNS

CARA's data-reporting requirement assumed states could count substance-affected infants. Most cannot.

The federal data system is broken on this exact field. Researchers analyzing NCANDS — the system CARA designated for this reporting — found that although the data are "supposed to contain information on whether the infant had substance exposure in utero, a plan of safe care, and referral to appropriate services," the most recent NCANDS data show "highly incomplete data for these fields."

The decade-long picture is worse: across the full 2010–2019 period, only 11 states had valid IPSE data. By 2019 that had risen to 25 — meaning half the country still could not produce a clean year of data three years after CARA passed.

FFY 2023 confirms the gap persists, seven years post-CARA:

CARA data element	States that reported it
Number of infants with prenatal substance exposure	49 of 51
Number with a Plan of Safe Care	35 of 51
Number referred to appropriate services	32 of 51

A third of the states cannot tell the federal government how many of their drug-exposed babies got the very plan the law requires.

There is no testing mandate — so identification is a coin flip. CARA requires states to report substance-affected infants, but no federal or state law mandates that hospitals test. Each hospital writes its own protocol. The result is documented chaos:

Newborn-report rates vary across states from less than 1 to 127.0 per 1,000 births — a more-than-100-fold spread that reflects policy and definition differences, not real differences in maternal drug use.
A CDC study of 15 Massachusetts hospitals found that standard ICD-10-CM diagnostic codes used for surveillance of substance-exposed newborns had a sensitivity of just 16%–61% — meaning the codes missed between 39% and 84% of cases.

The bottom line of Finding One: The United States has built a legal obligation to protect a population it has chosen not to count. The American Enterprise Institute stated it directly in its April 2024 report: "Due to infrequent and inconsistent testing, there is no reliable count of how many infants are exposed to substances in utero." A monitoring system cannot monitor a population it cannot see.

IV. FINDING TWO — HOSPITALS DON'T NOTIFY, AND ARE NOW FORMALLY QUITTING

CARA's notification requirement is the trigger for everything downstream: no notification, no Plan of Safe Care, no monitoring, no count. This trigger is failing in two distinct ways — a long-standing passive failure, and a new, deliberate, expanding one.

The passive failure (documented by GAO, 2018). In GAO report GAO-18-196, Substance-Affected Infants: Additional Guidance Would Help States Better Implement Protections for Children (issued 2018, publicly released March 5, 2019):

Only 42 states reported having policies that require health care providers to notify CPS. 8 states merely encouraged it. The remaining state required only an assessment.
GAO found a culture of non-reporting even where policy existed: a state official told auditors that "health care providers tend to view child welfare involvement as punitive rather than a potential resource for the family."
GAO documented that hospitals and agencies often chose not to notify CPS about infants born with drugs in their systems when the mother was taking prescribed opioids such as methadone — even though CARA's text does not carve out an exemption for prescribed substances.

The earlier Reuters investigation found the structural gap was even wider: at least 36 states had laws or policies that did not require doctors to report each case, and no more than nine states and the District of Columbia appeared to conform to the federal law.

The new, deliberate failure (2021–2025): hospital systems formally withdrawing. This is the most significant recent development and the reason this crisis is escalating rather than resolving.

April 2024 — Mass General Brigham: One of the largest hospital systems in the Northeast announced that eight hospitals with labor-and-delivery units in Massachusetts and New Hampshire — including Massachusetts General Hospital and Brigham and Women's Hospital — would stop filing neglect reports to state child welfare officials solely because a baby is born exposed to drugs. Under the new policy, staff report a newborn only if there is "concern about an imminent risk of abuse or neglect," and toxicology testing now requires written consent and is limited to cases where results would change medical management.
May 2021 — Boston Medical Center quietly made the same change first.
2024–2025 — a national wave: Health systems including Yale New Haven Children's Hospital adopted similar protocols. States including Connecticut, New Mexico, Washington, and Colorado revised laws to divert substance-exposed infants away from the child welfare system. The Marshall Project (April 2025) and Stateline (September 2024) documented this as a coordinated movement.

The rationale — and the collision with federal law. These changes are driven by real and well-documented harms. Studies show Black newborns were 3.8 times more likely to be drug-tested than White newborns in one Michigan cohort of 26,366 births when no maternal test had been done during pregnancy. Punitive reporting is associated with pregnant people avoiding prenatal care and substance-use treatment.

But here is the unresolved collision at the center of this crisis: CARA-amended CAPTA still requires notification of every identified substance-affected infant. Hospital systems are reinterpreting state mandatory-reporting statutes to mean that drug exposure alone is not "reasonable cause" to suspect abuse. They may be right about the racial-equity harms. They are nonetheless creating a situation in which a federal funding condition is being openly contradicted by the conduct of major health systems — and no federal enforcement action has followed. The notification mandate is being nullified in practice, hospital system by hospital system, while the states housing those hospitals continue to certify CAPTA compliance.

V. FINDING THREE — THE "PLAN OF SAFE CARE" IS A PAPER EXERCISE

When a Plan of Safe Care is created, the law envisions a real, multi-disciplinary, monitored plan addressing treatment for both the infant and the caregiver. The evidence shows that in much of the country it is something far thinner.

It is barely created. The single most damning data point comes from Mississippi, via ProPublica: in 2018, the state created Plans of Safe Care for just 155 infants identified as born affected by drugs — out of roughly 40,000 children born in the state that year. The physician who took over care of one drug-exposed infant a day after her birth told reporters he had never even heard the term "plan of safe care," despite it being a CAPTA requirement.

Where it exists, it is often voluntary and unbacked. The academic 51-state policy review found that in many states the POSC is "merely a form that documents services that mothers or infants already do or could receive," and that "the voluntary nature of these plans may mean that the vast majority of eligible parents receive no substance use treatment."

The AEI report goes further, calling Plans of Safe Care "voluntary offers of services seen as a more compassionate alternative to CPS involvement" that are "not backed by any evidence of their actual efficacy in keeping children safe."

It is not monitored. CARA requires states to monitor that services are actually delivered. In GAO-18-196, officials in the states GAO visited reported deep uncertainty — they did not know what to include in the plans or what level of intervention was appropriate, especially for infants deemed low-risk. The NCSACW federal technical-assistance briefs found only 33 states described any systems-level strategy for monitoring Plans of Safe Care in their Annual Progress and Services Reports — and "describing a strategy" is a far lower bar than demonstrating that monitoring occurs.

The CARA upgrade — caregiver treatment — is the most ignored. CARA's signature improvement over the 2003 language was requiring the plan to address the caregiver's treatment needs, not just the infant's. The 51-state review found only two states (Delaware and North Carolina) had policies compliant with CARA across all five domains. Thirty-two states were consistent with at least one domain. Nineteen states' policies were inconsistent with CARA on every domain. Put another way: 37.3% of states complied with zero of the five CARA domains, and 33.3% complied with only one.

The "report vs. notify" distortion. CARA deliberately uses the word "notify" — a service-oriented trigger — rather than "report," the punitive child-abuse-investigation trigger. The 51-state review found 40 states use the term "report" instead of "notify" and mandate investigations for identified families. This single word swap converts a public-health safety-planning tool into a child-removal pipeline — and helps explain both the racial-disparity backlash and the hospitals' retreat documented in Finding Two. The law's own design intent is being inverted in 40 states.

VI. FINDING FOUR — NO STATE FULLY COMPLIES, YET EVERY STATE CERTIFIES THAT IT DOES

This is the structural heart of the crisis.

In December 2019, The Boston Globe and ProPublica published the first national survey ever conducted of state compliance with CAPTA. They sent a 76-question survey to all 50 states, the District of Columbia, and Puerto Rico; 49 states, D.C., and Puerto Rico responded. The survey examined five core CAPTA areas, including the care of drug-affected infants. CAPTA imposes roughly 27 distinct provisions that states must meet to receive funding.

The finding: Not one state complied with all of the legal requirements.

The Plan of Safe Care provisions were among the most widely violated. So were others — for example, on CAPTA's requirement that every child in an abuse/neglect court proceeding receive a guardian ad litem, 49 states and Puerto Rico could not show compliance; only New Jersey and North Carolina could.

The academic literature independently confirms the pattern specifically for substance-exposed infants: of 51 jurisdictions, only two were fully compliant with the CARA Plan of Safe Care domains.

Now hold that against the funding mechanism. To receive its CAPTA State Grant, each state submits an annual report containing a compliance certification signed by the Governor. That certification asserts the state has the required policies and procedures in effect. The Globe/ProPublica investigation establishes that these certifications are, as a class, inaccurate — every state signs an assurance of full compliance, and no state is fully compliant.

This is the pattern of concern in its purest form: a universal gap between sworn certification and documented reality, on a child-safety law, tied directly to the release of federal money.

VII. FINDING FIVE — THE FEDERAL GOVERNMENT REFUSES TO VERIFY

A false certification is only a scandal if someone was supposed to check it. Someone was. They don't.

The HHS Administration for Children and Families (ACF) and its Children's Bureau are statutorily responsible for administering CAPTA and overseeing state compliance. When The Boston Globe and ProPublica asked how the Children's Bureau verifies the governors' certifications, the Bureau's answer was extraordinary in its candor:

The Children's Bureau "do[es] not question states' annual reports claiming to follow the law," and "the assurance that is submitted to the federal government is compliance for the state."

In other words: the signed assurance is not evidence of compliance — it is defined as compliance. There is no audit, no site verification, no penalty mechanism in routine use. A state cannot fail a compliance review that does not exist.

This is not an isolated interpretation. CWLA — the Child Welfare League of America — separately documented that "ACF Fails to Monitor States' Compliance in Implementing the CAPTA" requirements, and characterized the relationship between CAPTA's mandates and any actual data verification as "precarious."

GAO-18-196's central recommendation was that HHS direct ACF to provide "additional guidance and technical assistance" — notably a recommendation about helping states, not auditing them. The federal posture toward CAPTA compliance is advisory, not enforcement. The enforcement tool that exists on paper — withholding funds from a non-compliant state — has effectively never been used for Plan of Safe Care failures, because the federal government has structured the process so that non-compliance is never formally found.

The result is a closed loop:

Hospitals under-identify and increasingly decline to notify.
States cannot count what they were never told about.
Plans of Safe Care go uncreated, unmonitored, or are reduced to a form.
The Governor signs a certification of full compliance anyway.
The Children's Bureau accepts the certification as the compliance check, without verification.
CAPTA State Grant and Children's Justice Act funds are released.
Return to step 1.

Every actor in the loop can see the gap. The loop is designed so that no actor is required to act on it.

VIII. THE HUMAN COST — THE PREVENTABLE DEATHS

CARA exists because babies died. They are still dying, in the same ways, for the same documented reasons.

The deaths that wrote the law (Reuters, "Helpless & Hooked," December 2015). Reuters reporter Duff Wilson reviewed more than 50,000 pages of documents — including roughly 5,800 child fatality reports — and interviewed over 300 people. He found that since 2010, in at least 110 cases, babies whose mothers used opioids during pregnancy later died preventable deaths. Of those 110 children:

More than 40 suffocated — overwhelmingly in unsafe sleep environments.
13 died after swallowing toxic doses of methadone, heroin, oxycodone, or other opioids.
In 75 cases, child protection workers were notified but did not take the protective measures the federal law specified.
In roughly a dozen more cases, hospitals never reported the drug-dependent baby at all, and the child died after being sent home.

This investigation is the direct cause of CARA's 2016 CAPTA amendments. The law was written to close exactly these gaps.

The death that proves the law didn't work — Jasmine Irwin. ProPublica's companion investigation to the 2019 compliance survey told the story of Jasmine Irwin, born in Mississippi on Christmas Eve 2013 weighing just 4 pounds, 3 ounces, to a mother with a history of dealing and abusing methamphetamine. Hospital staff never followed up on her potential drug exposure before discharging her. In September 2015, her mother killed her — grabbing the toddler by the legs and pounding her head into the living room floor, in front of her older brother. Mississippi had no functioning procedures to protect substance-affected infants when Jasmine was born. The article's headline states the indictment: "The Law Says She Should Have Been Protected From Birth. Instead, She Was Left in the Care of Her Drug-Addicted Mother, Who Killed Her."

The deaths that keep happening — the Michigan pattern. WOOD-TV's Target 8 investigative unit documented the ongoing pattern in Michigan:

In 2014, CPS received referrals on 2,610 drug-exposed infants.
Target 8 profiled three infants — from Kalamazoo, Muskegon Heights, and Jackson — all born with illicit drugs in their systems, all dead before their first birthdays.
The chief investigator at Michigan's Office of Children's Ombudsman described the pattern in one sentence: "Often the deaths are from unsafe sleep practices." The Ombudsman formally recommended that CPS automatically open a service case whenever a newborn has unexplained illicit drugs in its system — a recommendation that, if standard practice, would not need to be made.
The same watchdog found Michigan's CPS failed to issue legally required "child death alerts" in 37 cases in 2016 — roughly 10% of the alerts the agency should have issued that year.

The clinical risk profile. This is not random misfortune. NAS-affected infants are at elevated risk of Sudden Unexpected Infant Death (SUID) and SIDS, and are a documented risk factor for abusive head trauma. Child fatality reviews of substance-affected infant deaths found unsafe sleep environments present in 58%–65% of reviewed cases. A relapsing caregiver, a substance-affected infant, and an unsafe sleep surface is a recognized fatal triad — and it is precisely the triad a real, monitored Plan of Safe Care is supposed to interrupt.

Even where plans exist, the protection is imperfect: a Delaware study — Delaware being one of only two fully CARA-compliant states — found that of 1,347 substance-exposed infants with POSC support, 0.82% (11 infants) still sustained serious physical or fatal injury. That is the best-case outcome, in the most compliant system in the country. In the 49 states that are not fully compliant, there is no equivalent data — because, per Findings One and Five, no one is counting.

IX. THE MONEY — FUNDING FLOWS AND THE ACCOUNTABILITY GAP

The crisis is, structurally, a financial one: federal money is the only leverage the United States has over state child-protection systems, and that leverage is being released without verification.

CAPTA State Grant (CFDA 93.669) — the Title I program:

Funded at approximately $90 million per year ($90.09 million appropriated for FY 2021).
This is a recent increase. From 2005 through 2017, CAPTA State Grants were frozen at just $25 million per year for the entire nation — a sum CWLA and state officials repeatedly described as turning CAPTA's growing list of mandates into an unfunded mandate.
Since FY 2018, Congress has directed that $60 million of the annual appropriation be used specifically to help states improve their response to families of infants affected by substance use disorders — i.e., the Plan of Safe Care mandate. This is the dedicated funding stream most directly implicated by this investigation.
The American Rescue Plan Act (March 2021) added a one-time $350 million for CAPTA, including $100 million for Title I, with a spend-down deadline of December 30, 2025 — a deadline that has now passed, meaning a large supplemental infusion was spent during precisely the years this investigation documents as non-compliant.

Children's Justice Act (CJA, CFDA 93.643):

Distributes approximately $17 million per year to all 50 states, D.C., Puerto Rico, and territories.
Each state receives a $50,000 base plus a population-based amount.
Funded from the Crime Victims Fund (federal criminal fines/fees), administered by DOJ's Office for Victims of Crime, awarded by ACF under Section 107 of CAPTA.
The $17 million cap has not changed since 2000 — meaning CJA funding has lost roughly half its real value to inflation over a quarter-century while the caseload of substance-affected infants rose 240%.

The accountability gap, quantified. Spread $90 million across 50 states, D.C., Puerto Rico, and territories and the average state receives well under $2 million annually in CAPTA State Grant funds — a trivial sum against a child-protection budget. This produces a perverse dynamic that state officials themselves articulated to the Globe/ProPublica reporters: the money is too small to fund genuine compliance, but the certification is too easy to obtain to create any incentive for it. A governor signs one assurance and the money flows; building functional data systems, hospital-notification compliance, and a monitored Plan of Safe Care program for tens of thousands of infants would cost far more than the grant. The rational state response — and the observed one — is to certify and move on.

The federal government has chosen not to close this gap. It could condition funds on verified data completeness (the NCANDS fields are already defined). It could audit a sample of certifications. It could trigger GAO-style compliance reviews. It does none of these. The $90 million + $17 million flows annually on an unverified signature, and the only federal response on record — GAO-18-196's recommendation — was to offer states more help, not more scrutiny.

X. KEY PLAYERS AND ENTITIES

Entity	Role in this crisis
HHS — Administration for Children and Families (ACF)	Administers CAPTA; releases State Grant and CJA funds.
HHS — Children's Bureau	Statutorily responsible for CAPTA compliance oversight. On record stating it does not question state certifications.
U.S. Government Accountability Office (GAO)	Auditor. Issued GAO-18-196 (2018) documenting notification and POSC implementation gaps.
National Center on Substance Abuse and Child Welfare (NCSACW)	Federal technical-assistance center; published the "How States Serve Infants" brief series documenting monitoring/data gaps. Operated by Children and Family Futures.
State Governors (51 jurisdictions)	Each signs the annual CAPTA compliance certification on which funding is conditioned.
State CPS / child welfare agencies	Front-line entities that under-count, under-monitor, and in many states convert "notify" into "investigate."
Birthing hospitals & hospital systems	The notification trigger point. Mass General Brigham, Boston Medical Center, Yale New Haven and others have formally curtailed reporting (2021–2025).
The Boston Globe / ProPublica	Produced the first and only national CAPTA compliance survey (Dec 2019).
Reuters	"Helpless & Hooked" (Dec 2015) — the 110-death investigation that prompted CARA.
AEI — Child Welfare Innovation Working Group	Published "The US Is Failing Substance-Exposed Infants" (April 2024), led by Naomi Schaefer Riley; argues POSC is unproven and screening rollbacks endanger infants.
Pregnancy Justice; racial-equity advocates	Argue mandatory testing/reporting is racially discriminatory and drives families from prenatal care — the countervailing force behind hospital-system retreat.
Child Welfare League of America (CWLA)	Documented "ACF Fails to Monitor States' Compliance" and CAPTA's "precarious link to requirements and data."

A note on the central tension: This crisis has two camps that both have evidence on their side. One (AEI) says the system identifies and protects too few substance-exposed infants and that hospital rollbacks endanger them. The other (Pregnancy Justice, equity researchers) says the system over-tests and over-reports Black and poor families, driving women from care. Both can be true — and the existence of a real, well-funded, voluntary, public-health-oriented, monitored Plan of Safe Care is the policy instrument CARA actually designed to resolve the tension. That instrument is the thing that does not functionally exist. The fight over testing and reporting is, in part, a fight that fills the vacuum left by the POSC mandate's collapse.

XI. PATTERNS OF CONCERN

Certification-as-compliance. The defining structural defect: a signed gubernatorial assurance is treated by HHS as compliance rather than as a claim to be verified. This converts the entire CAPTA accountability system into an honor system on a child-safety law.
Mandate without measurement. CARA requires states to serve a population (substance-affected infants) that CARA simultaneously fails to require anyone to reliably identify (no testing mandate; incomplete NCANDS fields). A mandate to protect an uncounted population is unenforceable by design.
The notification trigger is collapsing in real time. Unlike the older, passive failures, the 2021–2025 hospital-system withdrawals are active, public, expanding, and unanswered by federal enforcement. The crisis is getting worse, not better.
"Notify" became "investigate" in 40 states. A single-word distortion converted a public-health tool into a removal pipeline, fueling both racial disparity and the hospital backlash — and undermining the law's own intent.
The CARA upgrade is the most-ignored part. The caregiver-treatment requirement — CARA's signature improvement — is honored in only two states. The generation that most needs treatment (the relapsing parent in an unsafe-sleep household) is the one the system most reliably fails.
Money flowed during the failure. A $350 million American Rescue Plan supplement, including $100 million for Title I, was spent down (deadline Dec 30, 2025) across exactly the period this investigation documents as non-compliant — with no verification that the supplement improved POSC implementation.
The deaths are patterned and predictable. Suffocation in unsafe sleep, fatal opioid re-exposure, neglect by relapsing caregivers — these are not freak events. They are the specific, named, recurring outcomes that monitored Plans of Safe Care exist to prevent, and child fatality reviewers describe them as routine.
No litigation accountability has attached. Unlike many child-welfare failures, there is no significant body of successful wrongful-death litigation forcing reform here — in part because the victims are newborns with no surviving advocate, and the responsible failure (a missing or unmonitored plan) is diffuse. The absence of a court-driven correction leaves federal funding leverage as the only realistic lever — and it is the lever HHS declines to pull.

XII. ACTIONABLE FINDINGS & RECOMMENDATIONS

For federal oversight (HHS / ACF / Children's Bureau):

Replace certification-as-compliance with verification. Condition CAPTA State Grant release on complete NCANDS reporting of the three CARA data elements (identified infants, POSC developed, referrals made). A state that cannot report should not be certified compliant. This requires no new statute — only enforcement of the data fields that already exist.
Audit a random sample of gubernatorial certifications annually against on-the-ground practice, as the Globe/ProPublica survey did once and HHS has never done.
Issue binding guidance resolving the hospital-notification collision: either the notification mandate means what it says, or HHS must formally reconcile it with the equity-driven hospital rollbacks. The current silence lets a federal funding condition be nullified hospital-by-hospital.

For Congress:

Fund the mandate. $90 million nationally cannot build 50 functional substance-exposed-infant data and monitoring systems. The unfunded-mandate dynamic is the root incentive problem.
Require GAO to re-audit state CARA compliance — GAO-18-196 is now eight years old and predates the hospital-withdrawal wave entirely.
Mandate the caregiver-treatment domain explicitly and tie it to funding; it is currently met by two states.

For states:

Separate "notify" from "investigate." Build a true public-health notification pathway distinct from the abuse-investigation track, as CARA intended — this directly addresses both the racial-disparity harm and the hospital backlash.
Make Plans of Safe Care monitored and real, not a discharge form. Monitoring must verify service receipt, not plan creation.
Mandate child fatality review of every substance-affected infant death with a specific finding on whether a POSC existed and was monitored.

For Project Milk Carton / public accountability:

The single most powerful public-facing data point is the gap: every governor certifies full CAPTA compliance; no state is fully compliant; HHS does not check. This is verifiable, non-partisan, and applies to all 50 states.
A 50-state scorecard — built from the NCANDS FFY 2023 data (which states failed to report POSC and referral counts) cross-referenced with the academic 5-domain CARA compliance grades — would be a high-impact, defensible public artifact and a natural follow-up investigation.
FOIA targets: state CAPTA Annual Reports and gubernatorial certifications (public records in most states); state child fatality review reports for substance-affected infant deaths; ACF correspondence on any CAPTA compliance review conducted 2018–2026.

SOURCES

Federal reports and data: - U.S. GAO, Substance-Affected Infants: Additional Guidance Would Help States Better Implement Protections for Children, GAO-18-196 — https://www.gao.gov/products/gao-18-196 - HHS/ACF, Child Maltreatment 2023 (NCANDS) — https://acf.gov/cb/report/child-maltreatment-2023 - HHS/ACF, Prenatal Alcohol and Other Drug Exposures in Child Welfare Study: Final Report (Sept. 2021) — https://acf.gov/cb/report/prenatal-alcohol-drug-exposures-final-report - NCSACW, How States Serve Infants and Their Families Affected by Prenatal Substance Exposure: Brief 2 — Plans of Safe Care Data and Monitoring — https://ncsacw.acf.gov/files/prenatal-substance-exposure-brief2.pdf - NCSACW, CAPTA Plans of Safe Care — https://ncsacw.acf.gov/topics/capta-plans-of-safe-care/ - HHS/ACF, Children's Justice Act grant program — https://acf.gov/cb/grant-funding/childrens-justice-act - HHS/ACF, CAPTA State Grants — https://acf.gov/cb/grant-funding/child-abuse-prevention-and-treatment-act-capta-state-grants

Investigative journalism: - The Boston Globe / ProPublica, "The federal government has one main law to prevent child abuse. No state follows all of it" (Dec. 13, 2019) — https://www.bostonglobe.com/metro/2019/12/13/cry-for-help/prT5xvp27BGZK6AZQWRNVL/story.html - ProPublica, "How We Measured States' Compliance With a Forgotten Federal Child Abuse Law" — https://www.propublica.org/article/how-we-measured-states-compliance-with-a-forgotten-federal-child-abuse-law - ProPublica, "The Law Says She Should Have Been Protected From Birth..." — https://www.propublica.org/article/the-law-says-she-should-have-been-protected-from-birth-instead-she-was-left-in-the-care-of-her-drug-addicted-mother-who-killed-her - Reuters, "Helpless & Hooked" series (Dec. 2015), summarized at NBC News — https://www.nbcnews.com/news/us-news/newborns-die-after-going-home-moms-fighting-drug-addiction-n475531 - WOOD-TV Target 8, "CPS criticized after deaths of drug-exposed babies" — https://www.woodtv.com/news/target-8-cps-criticized-after-deaths-of-drug-exposed-babies/ - WBUR, "Some Mass. hospitals stop infant welfare reporting based solely on drug dependence" (Apr. 3, 2024) — https://www.wbur.org/news/2024/04/03/mass-general-brigham-newborn-drug-policy - The Boston Globe, "Mass General Brigham to stop filing neglect reports..." (Apr. 2, 2024) — https://www.bostonglobe.com/2024/04/02/metro/mass-general-brigham-reporting-substance-exposed-newborns/ - CNN, "Some hospitals are changing their response when babies are born exposed to drugs" (Apr. 11, 2024) — https://www.cnn.com/2024/04/11/health/substance-use-pregnancy-policy-change - The Marshall Project, "Why Some Doctors Are Pushing to End Routine Drug Testing During Childbirth" (Apr. 2, 2025) — https://www.themarshallproject.org/2025/04/02/hospital-connecticut-colorado-pregnancy-drug-testing - Stateline / Alabama Reflector, "States, hospital systems try less punitive drug testing of pregnant women and newborns" (Sept. 18, 2024) — https://stateline.org/2024/09/18/states-hospital-systems-try-less-punitive-drug-testing-of-pregnant-women-and-newborns/

Research, policy analysis, and academic literature: - AEI Child Welfare Innovation Working Group, The US Is Failing Substance-Exposed Infants (Apr. 2024) — https://www.aei.org/research-products/report/the-us-is-failing-substance-exposed-infants/ - "Planning for safe care or widening the net?: A review and analysis of 51 states' CAPTA policies addressing substance-exposed infants," Children and Youth Services Review — https://www.sciencedirect.com/science/article/abs/pii/S0190740918309848 - "Prenatal Substance Exposure and Neonatal Abstinence Syndrome: State Estimates from the 2016–2020 TMSIS" — https://pmc.ncbi.nlm.nih.gov/articles/PMC10204012/ - "U.S. State rates of newborns reported to child protection at birth for prenatal substance exposure," Int'l Journal of Drug Policy — https://www.sciencedirect.com/science/article/abs/pii/S0955395924002123 - "Identification of Substance-Exposed Newborns and NAS Using ICD-10-CM — 15 Hospitals, Massachusetts, 2017," CDC MMWR — https://www.cdc.gov/mmwr/volumes/69/wr/mm6929a2.htm - "Impact of Plans of Safe Care on Prenatally Substance Exposed Infants" — https://pmc.ncbi.nlm.nih.gov/articles/PMC8792271/ - "Estimating the incidence of substance exposed newborns with child welfare system involvement" — https://pubmed.ncbi.nlm.nih.gov/38232502/ - Casey Family Programs, "Plans of Safe Care" — https://www.casey.org/infant-plans-of-safe-care/ - Child Welfare League of America, "Series Finds that No State Follows All of CAPTA Requirements" — https://www.cwla.org/series-finds-that-no-state-follows-all-of-capta-requirements/ - CWLA, "ACF Fails to Monitor States' Compliance in Implementing the CAPTA GAL Requirement" — https://www.cwla.org/acf-fails-to-monitor-states-compliance-in-implementing-the-capta-gal-requirement/ - Pregnancy Justice, "Understanding CAPTA and State Obligations" — https://www.pregnancyjusticeus.org/wp-content/uploads/2020/11/Understanding-CAPTA-and-State-Obligations_2020.pdf - Michigan Medicine / IHPI, "Study shows racial inequities in newborn drug testing" — https://ihpi.umich.edu/news-events/news/study-shows-racial-inequities-newborn-drug-testing

METHODOLOGY & LIMITATIONS NOTE

This investigation synthesizes federal audit reports (GAO, HHS/ACF, NCANDS), peer-reviewed research, and investigative journalism. Key structural findings rest on the most authoritative sources available: the 2019 Boston Globe/ProPublica national survey (the only one ever conducted), GAO-18-196, and FFY 2023 NCANDS data.

Limitations the reader should weigh: - The landmark compliance survey is from 2019; no comparable national survey has been conducted since. FFY 2023 NCANDS data confirms the data-reporting gaps persist, and 2024–2025 reporting confirms the hospital-notification problem is worsening — but a precise current 50-state compliance grade does not exist because no entity produces one. That absence is itself Finding Five. - The Reuters "110 deaths" figure covers 2010 and after and is necessarily an undercount, limited to documented cases; it predates CARA and is cited as the law's origin, not its current scorecard. - "No reliable count" is a literal finding, not rhetoric. National totals in this report (e.g., 44,453 IPSE in FFY 2023) are explicitly floors. - This report does not resolve the genuine policy conflict between child-protection and racial-equity advocates. It documents that the conflict is, in part, a product of the Plan of Safe Care mandate's functional collapse.

Prepared by OPUS — Project Milk Carton Autonomous Intelligence Division. For public-interest child welfare transparency. All figures attributed to named, linked sources. Suitable for SCRIBE video-article conversion.