68.8% of Children Who Died From Abuse Were Under One, and 48.5% of Those Died on the Day They Were Born — Of Those 16, an Agency Had Been Involved With One

Abuse deaths excluding murder-suicide48100%

44 cases

of which under 3 years3879.2%

of which under 1 year3368.8%

of the infants, those who died on the day of birth1648.5%

share of the 33 infants

of those 16, cases where an agency had been involved16.3%

share of those 16

The 48.5% for day-of-birth deaths is measured against the 33 infants; the 6.3% for agency involvement against those 16; all other shares against the 48. Problems noted during pregnancy and the perinatal period include abandonment in 18 cases (37.5%), unplanned pregnancy in 13 (27.1%), no antenatal check-ups in 11 (22.9%) and no notification of pregnancy in 10 (20.8%).

The 21st report from the expert committee of the Children and Families Council covers 56 cases involving 65 children, occurring or coming to light between 1 April 2023 and 31 March 2024[1]. Of these, 44 cases involving 48 children were abuse deaths excluding murder-suicide, and 12 cases involving 17 children were murder-suicides.

The age breakdown is extreme. Of the 48 deaths excluding murder-suicide, 33 were infants under one year, 68.8%, and 38 were under three, 79.2%[1]. Nearly seven in ten had not reached their first birthday.

There is a further layer inside that. Of those 33 infants, 16 died on the day they were born, 48.5%[1].

The heaviest figure in the report comes next. Among those 16 day-of-birth cases, an agency had been involved in one, 6.3%[1].

Prevention is built on the points where a child meets society

Child abuse prevention in Japan is assembled around the points where a child comes into contact with society.

A pregnancy notification and a maternal and child health handbook, a birth registration, infant health checks, nursery or kindergarten, then school. Each is an occasion for an adult to see the child. Signs of abuse are picked up there. Child guidance centres and municipal abuse response units mostly act on information arriving through that route.

A child guidance centre was involved in 31.3% of cases

How far did involvement actually reach? Of the 48 abuse deaths other than murder-suicide, a child guidance centre was involved in 15 cases (31.3%), not involved in 25 (52.1%) and unknown in 8 (16.7%)[2].

By agency, both a child guidance centre and the municipality were involved in 14 cases (29.2%), the municipality alone in 3 (6.3%) and a child guidance centre alone in 1 (2.1%)[3]. Eighteen had some involvement.

For 30 of the 48, neither was involved.

Municipalities report that they use the council well

The local coordination mechanism is the regional council for children requiring protection, where agencies meet to share information on children needing support.

Municipalities were asked how well they use it. In the local authorities where a death occurred or came to light, 30 cases (73.2%) were in authorities reporting that they use it well and 11 (26.8%) that they use it to some extent. None reported using it little or hardly at all. Two authorities had no council and four had unknown status, so the base is 41[4].

Not one authority reported not using it.

Yet 75.6% of the children who died were never discussed there

Whether the child's own case was actually taken up by that council is a separate question.

Of the abuse deaths other than murder-suicide, the case had been discussed by the council in 10 instances (24.4%) and not discussed in 31 (75.6%); in more than seven cases in ten no discussion had taken place. The base is 41[5].

The breakdown follows. Of the 10 discussed, 8 (19.5%) were taken up at a working-level meeting and 6 (14.6%) at an individual case conference. Seven (17.1%) had been treated as children requiring protection[5].

Murder-suicide cases are starker. For deaths involving murder-suicide, the council had not discussed the case in 16 instances (100.0%); in every case there had been no discussion[5].

The mechanism runs. These children were not on it. How an authority rates its own use of the council and whether a child who died was discussed there are two different things.

The pregnancy-stage picture

The report also tallies circumstances during pregnancy and around birth. Abandonment appeared in 18 cases (37.5%), unexpected or unplanned pregnancy in 13 (27.1%), no antenatal check-ups in 11 (22.9%) and no pregnancy notification in 10 (20.8%)[3]. Multiple responses were allowed and cases overlap.

185 day-zero deaths over twenty years

The long series is tabulated too. Abuse deaths other than murder-suicide among children under one came to 504 across reports 1 to 20, 33 in report 21, and 537 in total. Of those, day-zero deaths came to 185, 16 and 201; deaths in the first month came to 43, 2 and 45[6].

Day-zero and first-month cases account for 45.2% of deaths among children under one across reports 1 to 20[7]. Close to half of the children who died before their first birthday died within a month of being born.

One note. When the agency cites these figures in its FY2026 budget request, it gives the 185 day-zero deaths as 37.7%. Dividing 185 by 504 gives 36.7%. Two documents from the same agency do not agree on the percentage. The counts from the verification report are used here.

The machinery sits after birth; the deaths sit on the day of it

Set the figures together and the prevention machinery and the place where deaths occur do not line up.

The machinery is post-birth. The health handbook, infant check-ups, nursery, school. All of it works once a child has been born. A report of abuse requires that somebody has seen the child.

The deaths sit on the day of birth. 48.5% are there. A child who died the day they were born never appears at a health check, a nursery or a school. Nobody who might report has seen them.

For this group the machinery has no point of contact. 6.3% is not a statement about insufficient effort by agencies but about there being almost no occasion, by design, to make contact.

More than six in ten day-zero perpetrators are the birth mother

Who was responsible? Among the 16 day-zero cases in report 21, the main perpetrator was the birth mother in 10 (62.5%) and unknown in 6 (37.5%). Across reports 1 to 21, the birth mother was the main perpetrator in 205 cases (83.3%)[8].

The birth father accounts for 3 cases (1.2%) across reports 1 to 21[8]. This is a situation where the person who gave birth is alone.

Half of the deaths are being left after birth

How the children died is recorded. Among the 16 day-zero cases in report 21, the form of abuse causing death was neglect in 8 (50.0%), physical abuse in 7 (43.8%) and unknown in 1 (6.3%)[9].

By direct cause, being left after birth accounts for 7 (43.8%, or 58.3% excluding unknowns) and suffocation other than strangulation for 4 (25.0%, or 33.3%)[10].

The most common cause is being left alone after birth. Not an act so much as the absence of anyone present.

In 79.2% of cases there was no third-party review

There are figures on what happened afterwards as well. Third-party review of the case had not been carried out in 38 instances (79.2%), had been carried out in 5 (10.4%) and was under way in 5 (10.4%)[11].

In eight cases out of ten, nobody outside examined what happened.

Route the system through a notification and it misses those who do not file

Is there a route to identification before birth? One: the pregnancy notification. File it and a maternal and child health handbook follows, and the municipality knows the pregnancy exists. But 10 cases (20.8%) had no notification, and 11 (22.9%) no antenatal check-ups. A route built on filing misses whoever does not file.

The shape is the same as non-take-up. The provision exists and does not trigger until the person presents themselves at the entrance. With pregnancy, the reasons for not being able to present are the same reasons the situation is severe. An unexpected pregnancy, circumstances that cannot be told to anyone, no financial room. What makes someone hesitate to file is what makes them need help afterwards.

A new programme puts contact before the notification

The agency's FY2026 budget request includes a new programme to publicise pregnancy counselling, reaching women through places they visit such as pharmacies selling pregnancy tests. Rather than waiting for a notification, it tries to create contact ahead of one.

The direction is the only one available: increase the occasions for contact with this group. But publicity solves part of it and not the rest. Knowing a service exists and being able to go are settled by different conditions.

The need to count provision and reach separately is the structure covered in exclusion and non-take-up（このサイトの記事）; here, presenting at the entrance is itself the heaviest barrier.

Building a detection net from the child's side hit its limit in young carers fall to 4.1% in the final year of high school and return to 6.2% in the third year of university（このサイトの記事）. A mechanism placed in schools does not reach those who have left. Here it fails further upstream, before birth.

What has to be counted before the next decision

What is known is how many children died and which agencies were involved. Two things are missing.

The first is how many pregnant women file no notification. Comparing births with notifications would give an approximation, but no such figure is published. Without the denominator there is no measuring whether the publicity reached anyone.

The second is what happened to the people who did contact a service. The programme sets its target as an increase in the number of enquiries. An increase in enquiries and an actual connection to a choice about birth, adoption or termination are different things. Set the target on the count alone and the count is all that can be known.

Further Reading

• 『子どもの貧困 — 日本の不公平を考える』 (Child Poverty: Considering Inequality in Japan)（外部サイト、新しいタブで開きます） (Aya Abe, Iwanami Shoten). A data-based account of where disadvantage in childhood arises and how it accumulates. Useful for thinking about where a system places its points of contact.

References

Verification of Deaths and Other Cases from Child Abuse (21st Report) — Expert Committee on the Verification of Child Abuse and Related Cases, Children and Families Council (2025). Children and Families Agency

Implementation of the 2025 Amendments to the Child Welfare Act and Related Laws — Children and Families Agency, Child Abuse Prevention Division (2025). National Conference of Child Guidance Centre Directors

Verification of Deaths and Other Cases from Child Abuse — Children and Families Agency (2026). Children and Families Agency