This main part of the review begins with the five critical gaps identified in the Poutasi report, what agencies are doing to respond to the critical gaps and our assessment of whether these gaps are closing.
Introduction
has one of the highest rates of child deaths by abuse in the OECD. Most of the children who die from abuse are aged under 5, and the largest group is aged under 1.8 This is something the Poutasi report highlighted.
The Poutasi report showed that, over the previous 30 years, there had been 33 reviews and reports about child abuse and deaths, a mix of coronial inquests, reviews by past Children’s Commissioners and reports by independent reviewers as well as by the Family Violence Death Review Committee.9
Eight of the 33 reports10 identified similar practice and systems gaps as the Poutasi report. These eight included high-profile cases of killed by those who were supposed to be caring for them. The similarity of the findings – and lack of change as a result – was noted by Dame Karen.
“I find it unacceptable that I need to once again make similar findings about how the system is – or is not – interacting. The majority of my recommendations are not new.” 11
Using data provided by Police and Oranga Tamariki, we identified that, between December 2021 and June 2025, 24 in died because of confirmed or suspected abuse, homicide, non-accidental injury or maltreatment by a person who was supposed to be caring for them. We are aware that, since June 2025, more tamariki have died in similar circumstances, and we acknowledge that these tamariki are not reflected in this review.
There is significant sensitivity in this information, and we respectfully note we have not reviewed the specifics of the lives and circumstances of these tamariki or the nature of their deaths. For example, we have not focused on the ethnicity of tamariki because Oranga Tamariki and Police have different ways of collecting ethnicity information, which makes reporting on ethnicity across these agencies problematic.12 Potentially, this is a gap for children’s agencies. A 2021 report13 found that “mortality is not evenly distributed in the population: rates are higher in Māori and Pacific children and young people than in other ethnic groups”.
| 13 were under the age of one |
Of these 24 tamariki |
14 died because of actions by a parent, all but one had day-to-day care |
| Nine were aged between one and five | Six died because of actions by a caregiver or step-parent with day-to-day care | |
| Two were aged six to eight | Four cases remain under investigation |
We have focused on these tamariki because of some of the circumstances they share with Malachi. Almost half (11) of the children were known to Oranga Tamariki before their death. Two further children aged under 1 had had siblings involved with Oranga Tamariki. Most of the alleged perpetrators (19) were known to the Police in varying ways. However, most of the children were not known to the Police.
Our 2024 review referenced a Child Matters statistic that one child dies every five weeks from abuse. This cannot be compared with the statistics in this year’s review because our focus is on where a child is killed by the person who was supposed to be caring for them. The average figure in our 2024 review was from Child Matters, and was based on any death resulting from abuse. This will include deaths caused by a wider group of people than we have focused on in this review.
Almost all 24 deaths have been, or will be, reviewed by Police and one-third reviewed by Oranga Tamariki
A death review had already been completed by one or both agencies for 14 of the tamariki. In addition to data, Police provided 10 completed Police family violence death reviews (PFVDRs) and told us a further 12 were in progress. Oranga Tamariki provided seven completed child death reviews and told us two more were in progress.
In Aotearoa, there are several different ways that non-accidental deaths of tamariki and are reviewed. Each of the six agencies we requested information from about child death reviews has a different focus and may or may not choose to undertake a review of a child death. The only agency that confirmed it does not have a formal agency review role was the Ministry of Education. However, any of these six agencies may participate in a review at the request of the National Mortality Review Committee | He Mutunga Kore,14 provide information for a coronial process or contribute to a multi-agency review. A more detailed overview of the mechanisms that may review tamariki deaths is provided in Appendix D.
Agency reviews are often conducted in isolation, do not always look at systemic issues and do not lead to change across the system
The literature and research around child death reviews often points to child deaths being hard to predict and notes reviews that focus too heavily on trying to find blame can lead to recommendations that have the wrong focus.15 Research has found recommendations of reviews often give undue weight to readily measurable aspects of practice rather than the more complex understanding of how the system did not prevent the death and what changes are needed at a system level.16
“When a child dies violently, New Zealand has closely followed other countries in adopting recommendations that emerge from the child death review process. But this has often been done uncritically, using a bureaucratic rather than a professionally focused approach, with the introduction of more protocols and the revision of procedures for social workers and allied professionals, as well as concurrent demands for greater compliance. This response assumes, incorrectly, that the often idiosyncratic circumstances surrounding a single child’s death can necessarily be generalised across other cases within the statutory child protection system, and that the specific professional responses that might have saved that particular child will necessarily be useful if applied more widely.”17
From the child death reviews provided to us by Police and Oranga Tamariki, we note the following.
- Police and Oranga Tamariki, for the most part, undertake reviews in isolation from other agencies. Their reviews focus on internal practices rather than broader systemic issues.18
- Most of the PFVDRs do not make specific recommendations and do not introduce new or systemic changes to existing Police practice – meaning that Police does not make any recommendations for itself.19 Some PFVDRs included findings and/or recommendations for Oranga Tamariki and the Ministry of Health. As reviews are undertaken in isolation from other agencies, it is unclear whether the recommendations or reviews that made recommendations for other agencies were given to those agencies to learn from. Some PFVDRs have a finding that no government agency held information that, if acted upon, could have prevented the death.20
- Some of the reviews undertaken by Oranga Tamariki appear comprehensive, make multiple findings, identify areas for improved practice and make recommendations. Some include a plan in response to the review that shows that actions were completed but provide no detail or explanation of how or whether it led to any change. Actions did not appear to align with the level of risk identified in the reviews.
- Reviews by Police and Oranga Tamariki focused on addressing symptoms contributing to harm but not the underlying causes.
Police told us that PFVDRs consistently highlight areas for improvement that align with recommendations made by the Poutasi report and that these areas are already being addressed by Police’s continuous improvement efforts. For example, Police has invested in frontline training and interagency co-ordination initiatives to support early identification and response to family harm.
Oranga Tamariki told us it makes changes to address specific findings of child death reviews. As an example, it said what has been learned has been incorporated into practice improvements and in implementing its new Practice Approach.
The second part of the review looks at the front door of the child protection system and how current reports of concern are handled. Even if the critical gaps identified by the Poutasi report are closed, does not yet have a child protection system that is fit for purpose and that is always able protect and when called upon.
We finish with a look at how child deaths are reviewed in and if and how learnings from these reviews inform improvements to policies and practices to keep safe.
8 Child Matters. (2025). New Zealand child abuse statistics. childmatters.org.nz/insights/nz-statistics/
9 In 2023, the Health Quality & Safety Commission made changes to the mortality review committees. The Family Violence Death Review Committee is now the Family Violence Death Review Subject Matter Experts, who report to the National Mortality Review Committee. The new national mortality review function is outlined in Appendix D.
10 See (pp. 29–31, 55–56) of report at footnote 1.
11 See (p. 31) of report at footnote 1.
12 Oranga Tamariki records multiple ethnicities and follows the Stats NZ ethnicity classifications and standards while Police does not. Police records one ethnicity for each individual despite guidance from Stats NZ that individuals should be able to identify with multiple ethnicities. Information about the approaches of Oranga Tamariki and Police to ethnicity data collection can be found in Ethnicity data collection by justice sector agencies: Prepared for Te Rau o te Tika – the Justice System Kaupapa Inquiry (WAI 3060) – June 2024.
13 Te Rōpū Arotake Auau Mate o te Hunga , Taiohi | Child and Youth Mortality Review Committee. (2021). 15th data report: 2015–19 Health Quality & Safety Commission (p. 4). hqsc.govt.nz/assets/Our-work/Mortality-review-committee/CYMRC/Publications-resources/CYMRC-15th-data-report2015-19_final_2.pdf
14 The Health Quality & Safety Commission has established a national mortality review function to review and report
on mortality through the collection, analysis and review of mortality data on specific classes of death with the aim
of preventing future premature death and promoting continuous quality improvement.
15 Connolly, M., & Doolan, M. (2007). Lives cut short: Child death by maltreatment. Office of the Children’s
Commissioner. manamokopuna.org.nz/publications/reports/lives-cut-short-child-death-by-maltreatment-marieconnolly-and-mike-doolan-published-for-the-office-of-the-childrens-commissioner-2007-/
16 Munro, E. (2011). The Munro review of child protection: Final report – a child-centred system. Department for
Education. gov.uk/government/publications/munro-review-of-child-protection-final-report-a-child-centred-system
17 See (p. 68) of report at footnote 16.
18 Police confirmed this is consistent with Police policy and the intent of PFVDRs. It noted that broader system issues are reviewed by the Family Violence Death Review Subject Matter Experts through the National Mortality Review Committee – Appendix D refers in more detail.
19 Police told us that, for many child death reviews, Police’s first engagement is at the point of notification of serious
injury or death. It told us that, in these cases, there are often few or no recommendations for Police action that
would have changed the outcome. It explained that PFVDRs tend to generate more recommendations in caseswhere an adult has died as a result of intimate partner violence, where Police may have had prior involvement.
20 Police guidance outlines the scope of PFVDRs in relation to other agencies as follows: “The review will also
consider the interaction of Police with external agencies. However, it must not examine the practices of those
other agencies apart from their engagement with [Police] and their known interaction with the parties involved, if relevant.”