Aboriginal and Torres Strait Islander children are increasingly overrepresented in out-of-home care. Protecting children means making the system answer for the warnings it misses and the failures it repeats.

Child protection asks some of the hardest questions any government can ask of a family. Is the child safe? What went wrong? What must change? The family’s answers are recorded, assessed and tested. But when the system repeatedly fails children and families, who asks those questions of it?

The Productivity Commission’s 2026 Closing the Gap report provides an urgent reason to ask. Aboriginal and Torres Strait Islander children remain dramatically overrepresented in out-of-home care, and the national trajectory is worsening. In 2025, 52.2 Indigenous children per 1,000 were in care. In 2024, they were 11 times as likely as non-Indigenous children to be in out-of-home care, up from 9.6 times in 2020.

These figures do not mean every removal was unnecessary. Some children face immediate danger and must be protected. But a worsening pattern cannot be explained only by what families did or failed to do. It also demands scrutiny of the institutions deciding their futures.

Recent contributions to P&I reveal failures at both ends of the child-protection system. Ruth Clare described children living with serious adversity who remain unseen because services and research often assume that a functioning parent will recognise the danger and seek help. Children most practised at hiding distress may therefore be least likely to receive support before crisis arrives.

Fabian Webber follows the story beyond intervention. Child-protection agencies know when a young person will leave care, just as hospitals know when a patient will be discharged and prisons when someone will be released. Yet young people still leave care without secure housing or continuing support, becoming homeless through a transition the state could see coming.

Together, these pieces expose a troubling continuity: some children are missed before intervention; others are abandoned at its end. The system activates at crisis but struggles to take responsibility across the child’s whole journey.

My research colleagues and I approached the same problem from a different angle. In a systematic review of continuous quality improvement in child protection, we examined how services might move beyond periodic evaluation and make learning part of everyday work.

The term sounds technical, but the principle is simple. A learning system regularly examines what is happening, listens to families and frontline workers, identifies where processes are failing, tests a change, studies the result and adapts again. It does not wait for tragedy, a royal commission or a damaging audit before asking whether its practices are helping children.

This is different from merely collecting more data. Child-protection systems already produce enormous quantities of information about families. Continuous improvement turns the lens back on the institution. Were warning signs recognised? Was help offered early enough? Did agencies coordinate? Were relatives and community-controlled organisations heard? What happened after the child entered care?

The aim is not to find someone else to blame. It is to create a system capable of noticing, learning and changing before repeated failure becomes another public scandal.

A practical test of whether child protection is learning begins with four questions.

First, what did the family need before the crisis? Housing insecurity, disability, mental illness, family violence, poverty and substance dependence do not automatically make parents unsafe. But without support, these pressures can intensify until risks become unmanageable. The question is not only what parents failed to do, but whether appropriate help arrived while change was still possible.

Second, what did the system do, or fail to do? Were warning signs shared across agencies? Did services coordinate rather than refer the family from one doorway to another? Were relatives, Elders and Aboriginal community-controlled organisations involved early enough? If removal became necessary, was kinship care seriously pursued?

Third, what changed after failure was identified? Reviews often produce careful recommendations, accepted in principle and gradually absorbed into administrative memory. A learning system should show which practice changed, who became responsible, what resources were provided and whether children’s lives improved.

Fourth, how will families and communities know? Findings should not disappear into internal reports or emerge only after freedom-of-information requests, investigative journalism or another tragedy. Those affected should be told what was learned, what will change and how progress will be measured.

These questions shift accountability without weakening it. Families remain responsible for their children’s safety. But departments, ministers and contracted services also become answerable for the help they offered, the conditions they created, the warnings they missed and the lessons they failed to apply.

None of this means governments should hesitate when a child faces immediate danger. Poverty, disadvantage or cultural difference cannot excuse violence, abuse or serious neglect. Nor should criticism of the system erase the courage of social workers making difficult decisions, often under intense pressure and with incomplete information.

Indeed, a learning system would support those workers better. It would provide stable staffing, reflective supervision, usable evidence, cooperation across agencies and authority to correct failing processes. Blaming an individual caseworker for systemic failure is no more useful than blaming a family for pressures created by housing shortages, inaccessible services or poverty. Protecting children and improving institutions are not competing obligations. Each depends on the other.

Child protection asks families to acknowledge harm, accept responsibility and demonstrate change. Reciprocal accountability requires the institutions judging them to do the same.

The standard should be clear: when child protection causes harm or fails to prevent it, the system must identify what went wrong, explain what it learned and show what will change. Families should be accountable for children’s safety. Governments should be equally accountable for building a system capable of learning before another child pays the price.

The views expressed in this article may or may not reflect those of Pearls and Irritations.