Welcome to the latest Weekly Wonk.
We’ve loved hearing from those of you in public-sector agencies about our new offering just for you. Reach out here if you want to learn more.
From the Founder’s Desk
Welcome to the latest Weekly Wonk, where we’re continuing our series with the Center for the Helping Profession for Child Welfare Workforce Development Month.
Part-one looked at the relationship between workforce conditions and safety.
For part-two, first-time Wonk contributor Tiffany Lindsey has a Deep Dive looking at how policy and institutional design generate a fundamental tension for child welfare agencies; rational risk management can create new safety risks.
For last week’s WonkCast, I had the privilege of sitting down with Sierra Burns to talk about what she’s learned from an advocacy trajectory that has taken her from South Carolina to the State of the Union.
Let’s get into it, with a Wonkatizer on Chafee and Tiffany’s Deep Dive.
Wonkatizer
The latest signals worth watching.
Whither Chafee?
Federal child welfare policy world continues to grapple with an open question this Congress; what will happen with the Chafee program?
What’s Happening
Congress established the John H. Chafee Foster Care Program for Successful Transition to Adulthood in 1999. It’s the largest federal funding source dedicated to helping youth prepare for life after foster care.
We’ve recently analyzed the latest data on services it funds for older youth.
A bill to make modest reforms to the program unanimously passed the House in May and has deep support from First Lady Melania Trump.
For anyone not in the weeds of DC policy process, you’d likely have a reasonable question; why hasn’t it moved through the Senate, then? Funding.
Why Funding Matters
The program has $143 million annually, plus an additional appropriation of around $43 million for education and training vouchers.
Funding has been essentially flat since the program’s start.
Complicating the point of eroded purchasing power by inflation is this 2025 GAO report, looking at unspent and returned funding despite unmet service needs for youth, which it attributes to factors including procurement barriers and staff turnover.
GAO also noted that ACF reported prioritizing reallocation of unspent funds to requesting jurisdictions rather than working directly with states with unspent funds to address underspending.
HHS indicated plans to support states on using unspent funds going forward, but what that led to is less clear.
Wonks have been watching for a proposal to update those mechanics in response to the GAO report, and Senator Grassley’s discussion draft gestured at addressing it. But the absence of a fleshed out proposal on that further complicates talks.
The Politics of Funding
The Senate could simply advance the House-passed bill to finish this work before the current Congress ends in January. But assuming they would misreads the incentives.
Whatever your politics or policy preferences, policy negotiation is inevitably driven by choke points. When someone needs your support, you can extract concessions.
If you’re perplexed that a bill would get slower after unanimously rolling through the House with the First Lady’s support, that’s actually why it’s likely to get slower.
When the minority party knows the majority wants something, that makes it a bargaining chip. Think like a staffer.
What to Watch
There’s already bipartisan legislation (H.R. 9485) to boost Chafee by $20 million annually, which makes that a natural opening point for talks to finish this effort.
It will be worth watching whether a compromise can cohere that increases funding, and addresses the issues GAO has identified so those funds actually reach youth.
Weekly Wonk Deep Dive
Designed Not to Know
How we perpetuate the conditions that discourage child welfare from asking the questions that can save children’s lives.
By Tiffany Lindsey, EdD, LPC/MHSP, Co-Founder and Chief Operating Officer, Center for the Helping Professions
A caseworker enters a home on a general neglect referral involving a twelve-year-old who has been missing school.
Inside, the worker finds an infant, a couch, and an exhausted single parent in recovery from a substance use disorder and taking sedating medication.
Two of the highest-yield safety questions are simple: Where does the baby sleep? Where are medications stored?
Those questions often can go unasked (Ahlers-Schmidt et al., 2021). Not because the worker does not care, but because asking can create a record of risks the agency may not be able to eliminate.
If the family insists on bedsharing, or if the worker identifies unsecured medications or firearms that remain in the home, that documentation can later become evidence that the agency “knew.”
After a fatality, lawsuit, or news story, “did the agency know” is often the first question.
That means workers and agencies face a dangerous incentive: stay close to the allegation, document what they can control, and avoid surfacing hazards they may not be able to fix.
Asking about risk can help prevent harm. But in child welfare, knowledge can also create liability.
The system has been designed, unintentionally but effectively, not to know. This paradoxically undermines safety and accountability in the name of advancing them.
The Question Unasked
Blame-oriented accountability can produce defensive, compliance-focused behavior rather than safety-focused practice (Munro, 2011).
When identifying a hazard also creates a record that can later be used to judge whether the agency did enough, even routine prevention decisions can carry institutional risk.
Consider the decision facing a worker who identifies a risk but cannot control what the family does next:
Ask about infant sleep practices → the family insists on bedsharing → documented, unresolved risk.
Provide a trigger lock → it sits unused → the record shows the agency knew of an unsecured firearm.
Provide a lockbox → the family sells it → the agency knew of unsecured medications.
Provide Narcan → the family uses it to save their baby → the agency knew drugs could be an ingestion risk.
After a tragedy, hindsight can make commonplace judgment appear negligent (Hugh & Dekker, 2009), so workers learn to minimize documented exposure to ambiguity.
As Cull (2026) posited, when accountability demands certainty the evidence can’t support, errors are more likely to get concealed rather than examined. The frontline result is that risks may go unelicited rather than engaged.
Workers stay close to the referral, assess what the allegation requires, and hesitate to surface hazards the agency cannot eliminate.
The predictable consequence is allegation-bounded assessment that training alone will not resolve.
The Resources Withheld
The same dynamic can operate at the agency level.
In my work on child death review and injury prevention with more than half of the state child welfare agencies in the country, I’ve heard multiple accounts of agencies obtaining naloxone with opioid-settlement dollars, then delaying distribution while leaders considered the legal and public consequences of providing it to families.
Others have raised concerns about unintended effects of take-home naloxone– that an antidote in the home might make riskier use feel safer, or make families slower to call 911 (Ager et al., 2024).
The available harm-reduction evidence points to the contrary. People who receive naloxone and overdose training show no increase in drug use afterward (Jones et al., 2017), and communities that distribute naloxone see overdose deaths fall (Fischer et al, 2025).
Naloxone reduces the likelihood that an overdose becomes a death. It doesn’t add drugs to the home nor create the underlying drug risk.
Yet the institutional calculation remains. The image of a child welfare agency providing naloxone to a family can be recast as evidence that the agency knowingly left a child in danger.
That scrutiny is not hypothetical. (National Partnership for Child Safety, 2025). It can unfold publicly, become politicized, and last for years (Rantz, 2026).
Under those conditions, locking away the naloxone can feel institutionally safer than distributing it.
Consider the contrast. When a police department hands a gun owner a free cable-style gun lock, such as Project ChildSafe - no one later treats the gift as admission the department is accountable for the family’s firearm storage.
Likewise, when a hospital provides parents of a newborn with a car seat, it is not presumed to control how the seat will be installed or used.
Child welfare operates under a different logic because its helping function is legally and culturally fused to its investigative function.
An effort to reduce danger can be recast as evidence that the agency knew danger was present—and should have prevented whatever happened next.
The Accident Pathways the Data Obscure
The public imagines maltreatment fatalities as intentional brutality. The data says otherwise.
In federal fiscal year 2024, states reported 1,773 child maltreatment fatalities.
Fewer than 40 percent of those fatalities involved physical abuse, while the majority involved neglect without co-occurring abuse (Children’s Bureau, 2026).
But NCANDS does not record cause or manner of death - itself a telling data gap. A death classified as neglect could involve unsafe sleep, drug ingestion, drowning, a firearm, or another mechanism entirely.
That omission prevents the field from determining how many maltreatment fatalities follow potentially preventable pathways. It also makes those pathways easier to overlook when designing child welfare practice.
Three hazards that drive a significant number of overall child fatalities illustrate the gap:
Sleep-related deaths. About 3,400 infants died suddenly and unexpectedly in 2024.
Of those, 947—approximately 28 percent— were classified as accidental suffocation and strangulation in bed (CDC, 2026).
Risk concentrates exactly where child welfare already is - with caregivers who are overwhelmed, under-resourced, and sometimes managing substance misuse (Hornstein et al, 2014).
Drug exposures. In 2023, 539 children younger than six experienced fentanyl exposures (Temple and Hendrickson, 2024).
These incidents can arise from failures of storage and containment rather than an intent to harm.
Unsecured firearms. Roughly once a day in America, a child finds a loaded, unsecured gun and unintentionally shoots themselves or someone else (Cannon et al, 2023)).
Of 1,262 unintentional firearm deaths among children, the firearm was stored loaded in three-quarters of cases - most often in a nightstand or bed (Wilson et al., 2023)
Storage choices by caregivers can reflect circumstances that complicate prevention: a recent victimization, neighborhood violence, or hypervigilance following trauma.
These national figures do not tell us how many child maltreatment fatalities resulted from each pathway.
They show that serious and sometimes fatal hazards can arise through the interaction of household conditions, caregiver capacity, product access, and safety design — not only through an intent to injure.
If child welfare data collapse those deaths into the broad category of neglect, the system cannot reliably identify which prevention strategies might have interrupted them.
Design with an Assumption of Error, not Perfect Compliance
The accident pathways described above raise a different question: How should a system respond when serious harm can result from predictable human error?
The contrast with medicine and aviation offers one framework.
Human factors engineering — the safety science used in both fields — starts from a basic premise: error is often a consequence of upstream conditions like fatigue, stress, and cognitive load, not the root cause of a failure.
Environments can steady or overextend individuals. People generally do what makes sense to them at the time, given the resources and constraints they face (Dekker, 2019).
The goal is not to eliminate individual responsibility. It is to design systems in which predictable mistakes are less likely to become catastrophic.
Safety science also ranks interventions. Engineering controls, which make a hazard physically harder to reach, are generally stronger than administrative rules. Rules, in turn, are generally stronger than warnings and education alone.
A medication lockbox, a trigger lock, naloxone on the shelf, and a portable crib are engineering controls. These interventions work because they do not depend entirely on perfect behavior.
Locked and unloaded firearm storage, for example, is associated with meaningfully lower risk of youth firearm injury and death (Grossman et al., 2005).
Current child welfare practice leans largely on the weakest rung: safety brochures, verbal education, assessment of the home environment, and safety plans that depend on a caregiver’s promise to avoid crib clutter and keep medications and firearms out of reach.
Those approaches have a role. But they depend on sustained attention and compliance from caregivers who may be exhausted, under-resourced, managing substance use, or responding to trauma.
And when a risk falls outside the original allegation or formal assessment, it may not be addressed at all.
We know which controls work better. The system is simply not built to place those protections routinely in families’ hands.
WHAT DECISION MAKERS NEED TO KNOW
A human factors approach to child welfare would begin with universal precautions. Ask every family and offer every family the same basic protections.
When the questions and resources are routine, asking signals prevention rather than suspicion.
That could include:
Safe sleep assessment, education and portable crib distribution in every case involving an infant, aligned with pediatric guidance (Moon et al., 2022).
Medication lockboxes and naloxone co-distribution in households where high-risk drugs may be present.
Lethal means counseling and free storage devices, drawing on models used in pediatric health care and the Department of Veterans Affairs.
This is not a new investigative model. It is motivational interviewing and collaborative safety planning applied to common accident pathways.
It positions the caseworker as safety consultant for accident-pathway risks, alongside their protective mandate.
Workers and families need to know that asking about a hazard is meant to reduce it—not simply document it for later use.
Harm reduction has never meant accepting unnecessary risk. It means meeting families where they are, and reducing the chance that a dangerous condition becomes a child’s death.
But whoever funds the naloxone must also protect the worker who distributes it. Otherwise, the rational institutional response remains the same: Do not ask about a risk you cannot fully control.
No new statutes are required to start, but statutory protections and legislative support will be needed to sustain.
A System Able to Know
Return to the living room. Same baby, same couch, same exhausted parent.
Under redesigned incentives, the caseworker addresses the truancy conditions involving the 12-year old but does not stop at the allegation.
The worker also asks where the baby sleeps, where medications are stored, what support the parent needs, and what contingency plans are in place for continued sobriety and safety.
The questions are routine. The resources are available. The caseworker can leave a portable crib, a medication lockbox, and naloxone—and document not only that the risks existed, but what the agency did to reduce them.
None of those interventions gives the caseworker control over what happens next. That is the point. A safety system should not require control it cannot possess.
It should make it easier to identify foreseeable hazards, provide practical protections, and respond proportionately to the risk.
Concrete help may also change the relationship between families and the agency. A worker who arrives with something useful—not only questions, findings, and requirements—has another way to build the trust on which honest safety planning depends.
A system designed not to know is a system unable to protect. Families are not the only ones who need harm reduction. Our systems do too.
Tiffany Lindsey is Co-Founder and COO and founder of the Center for the Helping Professions.
That’s it for this week.
Stay sharp, Wonks.
~ Z
Partners Making Your Weekly Wonk Possible
