The death of a child

We have already written seven guides out of this loss. This one deliberately covers what those seven do not: the siblings who are still here, the fathers the research keeps leaving out, what an employer actually owes you, and the three places where the received wisdom about a child's death turns out to have nothing behind it.

Two things travel with every guide we write. Where we offer a way through, it is the way that worked for us — it may not be yours, and our family is not a study. And where we give you a fact, we have read the source: author, publication and year are at the back so you can check us. Do your own reading and compare, including on us. If you check something here and we have it wrong, that is the most useful message you could send us.

The numbers, stated carefully

The American Cancer Society estimates that about 9,680 children aged 0 to 14 and about 5,660 adolescents aged 15 to 19 will be diagnosed with cancer in the United States in 2026, and that about 730 adolescents will die of it. Five-year survival now runs at about 85% for children and 88% for adolescents, against roughly 58% for children in the mid-1970s. The National Cancer Institute puts it in round terms: approximately 1,600 children and adolescents still die of cancer each year, and the death rate for under-20s fell from 5.1 per 100,000 in 1975 to 2.2 per 100,000 in 2022.

An 85% average hides the only thing that matters to the family inside it. Using NCI's figures for cases diagnosed 2014–2020: acute lymphoblastic leukemia has gone from 57% in 1975 to 92.3%. Non-Hodgkin lymphoma from 43% to 91.4%. And then diffuse intrinsic pontine glioma, where half of children survive less than a year from diagnosis and only about 10% survive two. There is no single childhood cancer survival rate. There is a list of very different diseases wearing one word.

Two numbers people will quote at you that are wrong

“Cancer is the number one cause of death in children.” It is not, in any childhood age band, in any recent year. CDC's 2024 leading-causes data, from the National Vital Statistics System: unintentional injury is first in every band. Cancer is third at ages 1–4 (278 deaths), second at ages 5–9 (400), third at ages 10–14 (446, behind injury and suicide), and fourth at ages 15–24 (1,421).

The accurate version is the one ACS and NCI actually use: cancer is the leading cause of death from disease among children and adolescents. That sentence is true, it is strong enough, and it will survive being checked. The other one will not, and a charity that gets caught on it loses the right to be believed about anything else.

“Only 4% of federal cancer research funding goes to childhood cancer.” We went to look. NCI's funding-by-cancer-type table has no pediatric or childhood cancer line in it at all — the rows are brain and CNS, breast, cervical, colorectal, head and neck, Hodgkin, leukemia, liver, lung, melanoma, myeloma, non-Hodgkin, ovarian, pancreatic, prostate, stomach and uterine. No percentage of the NCI or federal cancer budget is attributed to childhood cancer anywhere on it. The 4% figure is a derived number, not an NCI line item. The case for pediatric research funding does not need it, and using it puts the argument on ground that will not hold.

The siblings who are still here

This is the largest gap in everything else we have written, and the research says two true things at once.

Sharma and colleagues (Pediatric Blood & Cancer, 2025) looked at 4,558 adult siblings — 733 bereaved, 3,825 not — in the North American Childhood Cancer Survivor Study. In absolute terms the bereaved siblings' long-term psychosocial health was excellent: depressive symptoms in 6.5%, somatization 4.4%, anxiety 3.5%. In relative terms their risk was genuinely raised: depression RR 1.53, reduced social quality of life RR 1.35, diminished educational attainment, and a greater burden of chronic health conditions than their non-bereaved counterparts.

Both halves are the finding. Most bereaved siblings do well. Their risk is still higher than it should be. Anyone telling you only one of those is selling you something. One limit worth knowing: these were siblings of children who had already survived five years from diagnosis, so it is not the same population as a sibling of a child who died within months.

What actually seems to make the difference, and almost all of it is about information

A Swedish nationwide cohort followed 174 of 240 eligible siblings whose brother or sister was diagnosed with cancer before 17 and died before 25, surveyed two to nine years afterwards. Across six papers from that one cohort, the pattern is consistent and it is the opposite of protecting them:

Siblings whom nobody talked to about what to expect as their brother or sister was dying reported higher anxiety two to nine years later (Lövgren and colleagues, Psycho-Oncology, 2016). 70% reported witnessing their sibling suffering in the last hours of life. Many who were not present during the illness or at the death expressed regret about it.

The predictors of unresolved grief (Lövgren and colleagues, Journal of Palliative Medicine, 2018) were the sibling's perception that it was not a peaceful death, limited information in the last month of life, and being told about the impending death only the day before. We are giving you the direction and not the odds ratios, because one of those confidence intervals spans a factor of seventeen — the direction is solid, the size is not.

And Eilertsen and colleagues (Journal of Pediatric Oncology Nursing, 2013) found that an unsatisfied need for social support was associated with later anxiety at every point they measured — the last month before the death, the time after it, and at follow-up. Including this: siblings who did not perceive that their parents cared for them after the death had nearly three times the risk of later anxiety.

We want to be careful about how that last one lands, because it is the cruellest sentence in this guide if read wrongly. It is perception, in a self-report study, years later. It does not mean anybody failed. What it means is that a sibling needs to be told they are loved out loud, at a time when every adult in the house has nothing left — and that saying it, even badly, even late, is not a small thing.

What the siblings themselves advised, asked years later (Wallin and colleagues, Death Studies, 2020): about half of what they said was be with your sibling, and cherish the time. The rest was about communicating, letting go of guilt, and living life as usual.

One honest gap. We searched for research on bereaved siblings and school — re-entry, what teachers should do, academic trajectories — and found four records in total. There is nothing citable. If your child is going back into a classroom, nobody has studied it, and we are not going to invent guidance.

Bereaved fathers

The literature is overwhelmingly about mothers, and the size of the gap is itself the finding.

McNeil and colleagues (Pediatrics, 2021) ran a systematic review of grief in fathers after a child's death. They screened 1,848 titles to find 21 articles, covering 2007 to 2019, and most were descriptive. What they synthesized: fathers “often avoided discussing their grief with others, returned to work earlier, and used goal-oriented tasks as coping strategies,” and their conclusion was that “many fathers navigate loss through stoicism, self-isolation, and hard work” and that for some, those may be inadequate.

And now the study that points the other way, which we are going to leave unresolved rather than harmonize. Pohlkamp, Kreicbergs and Sveen (Psycho-Oncology, 2019) surveyed 133 mothers and 92 fathers across Sweden who had lost a child to cancer one to five years earlier. Symptoms of prolonged grief, depression, post-traumatic stress and insomnia were elevated in every one-year subsample regardless of how much time had passed — and mothers showed higher symptom levels of prolonged grief, depression and post-traumatic stress than fathers. A US study of 79 parents (Seecharan and colleagues, 2004) found no difference at all.

So: “fathers suffer worse in silence” is not established, and neither is its opposite. What is best established is subtler and comes from the strongest study design in this whole field. Li and colleagues (The Lancet, 2003) followed 21,062 bereaved Danish parents against 293,745 matched controls for up to eighteen years. Bereaved mothers' excess mortality from natural causes appeared only in years ten to eighteen. Fathers showed only an early excess, from unnatural causes.

That is not one parent suffering more. It is two people breaking in different ways at different times — which, if you are the couple in question, may be the single most useful thing on this page, because it explains why you are not in the same place in the same month and why that is not a sign that one of you has stopped caring.

What we will not do is dress any of this up in the “instrumental versus intuitive griever” typology you may have been handed. It is a teaching heuristic with no validation study behind it that we could find. The measured differences above are real findings. The typology is not one, and using the first to prop up the second is how a convention gets mistaken for evidence.

Work, and what your employer actually owes you

Start here: there is no federal bereavement leave entitlement in the United States.

The Family and Medical Leave Act does not cover it. The six qualifying reasons at 29 CFR § 825.112(a) are birth, adoption or foster placement, caring for a spouse, child or parent with a serious health condition, your own serious health condition, a military exigency, and caring for a covered servicemember. A death in the family is not among them, and the Department of Labor's own FMLA guidance does not mention bereavement, funerals or grief anywhere.

One route that is not bereavement leave and should not be confused with it: FMLA does cover your own serious health condition. A bereaved parent who develops a diagnosed condition requiring treatment may be eligible on that basis. That is our reading of the regulation's plain text, not a Department of Labor ruling, and we have no data on how often it is granted.

Below is what we verified against state agency and legislature sources. We checked eight jurisdictions and we are not going to imply we checked fifty.

This is published law, not advice about your situation. Read your own state labor department's page and your own employer's written policy, and look for how the policy defines “immediate family” — that is where the answer usually lives.

On returning to work itself, we have almost nothing honest to give you. The only measured finding we found is the synthesized observation that fathers returned earlier. We could not find a single primary study of how long US bereaved parents actually stay away, what share lose their job, or what share of US employers offer bereavement leave at all. The employer-survey percentages that circulate come from sources we could not verify, so we are not repeating them.

Their room, and their things

There is no research on this. None. We want to be unusually plain about it because it is one of the most-asked questions in this subject and the internet is full of confident answers.

We ran four targeted searches of the medical literature for work on what bereaved parents do with a child's belongings, possessions or bedroom. Zero relevant records. A broader search returned twenty, of which three touch the question and none is about a parent deciding what to do with a dead child's things.

So all of the following are convention, not evidence: that there is a right time to clear a room; that keeping it unchanged is harmful; that clearing it quickly is harmful; that keeping clothes helps; that “you'll know when you're ready”; that a memory box helps; that any order of keep, donate and discard leads to better adjustment. Nobody has tested any of it.

The nearest thing to evidence cuts in an unexpected direction. A 2026 study of 250 bereaved adults found that 83.2% kept an object connected to the person who died, and that the keepers scored higher on grief symptoms while the non-keepers scored higher on avoidant attachment. It is cross-sectional, so nobody knows which way it runs — keeping things may be a marker of intense grief rather than a cause of it, and the avoidant-attachment finding among non-keepers points that way. We are not turning that into advice in either direction, and neither should anybody else.

There is one sound study near this, and it is about something else: Foster and colleagues (Journal of Pediatric Oncology Nursing, 2009) interviewed 36 mothers, 27 fathers and 40 siblings from 40 US families about ten months after a child's death from cancer. Many of the children had done deliberate things to be remembered — making crafts, willing away belongings, writing letters. And some, particularly those who were very ill or died unexpectedly, did nothing intentional at all. We include that second half because every piece of writing about legacy implies your child left you one, and for a lot of families that is simply not what happened, and it is not a failure of theirs or yours.

The anniversary, and why we are not going to tell you to brace for it

A predictable anniversary effect is not established, and for a parent on the anniversary of a child's death there is no study at all. We looked specifically.

The best-designed evidence that exists is a self-controlled analysis of Danish national register data covering everyone bereaved by the suicide of a close relative between 1980 and 2016, where each person serves as their own control so that fixed differences between people drop out. Its finding: “We found no evidence of an elevated risk of suicidal behaviour during periods around anniversaries of a death or the deceased's birthdays.” It is also the only direct test of the dead person's birthday that we found, and it is null.

Other studies do find effects, but only in some sexes, some age bands and some narrow windows, with wide confidence intervals and few events — and one case-crossover study of sibling-loss anniversaries found risk running in opposite directions for men and women and concluded that overall, anniversaries were not associated with elevated mortality. Every one of these studies is about losing a parent or a sibling. Not one is about losing a child.

Which is why we think the common advice here may do harm. Telling a parent to expect a collapse on a date might simply install the expectation. Some people find certain days hard; nobody has shown it is predictable. The researchers' own conclusion is the better guidance: bereavement support should be available at all points after a loss, because needs differ across the whole trajectory — not rationed to the calendar.

Where to actually go — we checked these ourselves

Three traps in this sector, because the names read right.

The Sibling Support Project is live, active and well regarded — and it is a disability organization, for siblings of people with developmental, health and mental health conditions. It is not a bereavement organization. This is the easiest mistake to make here.

What's Your Grief is a well-regarded professional content business, not a nonprofit — its own About page states it receives no funding or grants and that it is between fiscal sponsors, so donations are not tax-deductible. Good material; do not list it as a charity.

Alex's Lemonade Stand Foundation's SuperSibs is widely recommended, but the program as its own current page describes it is a two-year mailing program for siblings during a brother or sister's treatment, with no bereavement content, and supersibs.org is now only a redirect. Other organizations still describe it as offering grief support and scholarships. We could not tell which description is current. Ask them directly before you count on it.

What we have found to help

Ours, not findings. Take it or leave it.

Tell the siblings what is happening, in plain words, before it happens. If there is one thing in this guide with real evidence behind it, it is this one, and the evidence runs against every instinct to protect them.

Say to the surviving children, out loud, that you love them — on the days you have nothing left. Badly is fine. Late is fine. The Swedish data on siblings who did not perceive that their parents cared for them afterwards is the one finding we would most want every bereaved parent to have seen, not as a reproach but as information: it is a perception, and perceptions respond to being told.

Expect to be in different places from your partner, in different months, and do not read it as a verdict. Eighteen years of Danish register data says the two of you may break at different times. That is the shape of it, not a failure of the marriage.

And then, eventually, do something for somebody else. It is what our family did and it is why this charity exists. We are not going to tell you it is what grief does — the evidence on growth after loss is far weaker than the internet claims, and we set out where it thins in the main guide. We are telling you what we did.

Common questions

Is cancer the leading cause of death in children?

No. In CDC's 2024 data, unintentional injury is the leading cause in every childhood age band, and cancer ranks third at ages 1 to 4, second at 5 to 9, third at 10 to 14 and fourth at 15 to 24. The accurate statement, and the one the American Cancer Society and the National Cancer Institute both use, is that cancer is the leading cause of death from disease among children and adolescents.

Do most marriages end after a child dies?

No, and the claim that 90% do is false. It traces to a 1977 book that said something different, that many couples face serious marital difficulty, and cited nothing. A Norwegian study of 977,928 married couples, including 4,590 parenting a child with cancer, found no increased risk of parental divorce overall. Marital strain is real and widely reported. The marriages mostly hold.

How are bereaved siblings affected long term?

Two things are true at once. In a study of 4,558 adult siblings in the Childhood Cancer Survivor Study, the 733 bereaved siblings reported excellent absolute psychosocial health, with depressive symptoms in 6.5%, somatization in 4.4% and anxiety in 3.5%. But relative to non-bereaved siblings their risk was genuinely raised, including a relative risk of 1.53 for depression, reduced social quality of life, lower educational attainment and a greater burden of chronic health conditions.

Should we tell a child's brother or sister that they are dying?

The evidence says yes, and it runs against the instinct to protect them. In a Swedish nationwide cohort of 174 bereaved siblings surveyed two to nine years later, those whom nobody talked to about what to expect reported higher anxiety years afterwards, and being told only the day before was among the predictors of unresolved grief. Many siblings who were not present during the illness or at the death later expressed regret about it.

Do fathers grieve differently from mothers?

Not in the way the popular version claims. A systematic review screened 1,848 titles to find 21 mostly descriptive studies and synthesized that fathers often avoid discussing grief, return to work earlier and use task-focused coping. But a Swedish study of 133 mothers and 92 fathers found mothers had higher symptom levels of prolonged grief, depression and post-traumatic stress, and a US study of 79 parents found no difference. The best-designed evidence, from eighteen years of Danish register data, suggests not that one parent suffers more but that mothers and fathers are affected in different ways at different times.

When should we clear out our child's room?

There is no research on this at all. We ran four targeted searches of the medical literature and found nothing on what bereaved parents do with a child's belongings or bedroom, or whether any approach leads to better adjustment. That there is a right time, that keeping a room unchanged is harmful, that clearing it quickly is harmful, that a memory box helps, that you will know when you are ready, are all convention rather than evidence. Anyone giving you a timeline is giving you their opinion.

Does my employer have to give me bereavement leave?

Not under federal law. There is no federal bereavement leave entitlement in the United States and the Family and Medical Leave Act does not cover bereavement among its six qualifying reasons. Some states mandate it, such as California with five unpaid days per death and Oregon with two weeks per family member. Illinois gives two weeks for most deaths and twelve weeks only where a child died by suicide or homicide, so a cancer death gets two. Washington's paid leave program states that it does not cover bereavement leave at all. New York, despite several published roundups saying otherwise, is not a bereavement leave state.

Last reviewed 27 October 2026. Written by Austin's family. We are not clinicians, and nothing here is medical, legal or financial advice — the leave section describes published law and is not advice about your situation. Nothing on this page is behind a paywall, no organization named here pays us, and there are no affiliate links. See our editorial and sourcing policy.