Association of Child Maltreatment with Risk of Death during Childhood in South Australia

Leonie Segal, James Doidge, Jason M. Armfield, Emmanuel S. Gnanamanickam, David B. Preen, Derek S. Brown, Ha Nguyen

Research output: Contribution to journalArticlepeer-review

9 Citations (Scopus)


Importance: Child maltreatment is a prominent public health concern affecting 20% to 50% of children worldwide. Consequences for mental and physical health have been reported, but population-level estimates of risk of death during childhood that are adjusted for confounders have not been published to date. Objective: To estimate the association of documented child protection concerns regarding maltreatment with risk of death from infancy to 16 years of age. Design, Setting, and Participants: This case-control study was nested in a population birth cohort of 608547 persons born in South Australia, Australia. Case children were those who died between 1 month and 16 years of age (with the death registered by May 31, 2019). Control children were randomly selected individuals from the same population who were alive at the age at which the case child died, matched 5:1 for age, sex, and Aboriginal status. Data were analyzed from January 2019 to March 2021. Exposure: Children were assigned to 1 of 4 child protection concern categories (child protection system notification[s] only, investigation[s] [not substantiated], substantiated maltreatment, and ever placed in out-of-home care) based on administrative data from the South Australia Department for Child Protection or were classified as unexposed. Main Outcomes and Measures: Mortality rate ratios for death before 16 years of age, by child protection concern category, were estimated using conditional logistic regression, adjusted for birth outcomes, maternal attributes, and area-based socioeconomic status. Patterns of cause of death were compared for children with vs without child protection concerns. Results: Of 606665 children included in the study, 1635 were case children (57.9% male [when sex was known]; mean [SD] age, 3.59 [4.56] years) and 8175 were control children (57.7% male; mean [SD] age, 3.59 [4.56] years [age censored at the time of death of the matched control child]). Compared with children with no child protection system contact, adjusted mortality rate ratios among children who died before 16 years of age were 2.69 (95% CI, 2.05-3.54) for children with child protection system notification(s) only; 3.16 (95% CI, 2.25-4.43) for children with investigation(s) (not substantiated); 2.93 (95% CI, 1.95-4.40) with substantiated maltreatment; and 3.79 (95% CI, 2.46-5.85) for children ever placed in out-of-home care. External causes represented 136 of 314 deaths (43.3%) among children with a documented child protection concern and 288 of 1306 deaths (22.1%) among other children. Deaths from assault or self-harm were most overrepresented, accounting for 11.1% of deaths in children with child protection concerns but just 0.8% of deaths among other children. Conclusions and Relevance: In this case-control study, children with documented child protection concerns, who were known to child protection agencies and were typically seen by clinicians and other service providers, had a higher risk of death compared with children with no child protection service contact. These findings suggest the need for a more comprehensive service response for children with protection concerns..

Original languageEnglish
Article numbere202113221
JournalJAMA network open
Issue number6
Publication statusPublished or Issued - 10 Jun 2021
Externally publishedYes

ASJC Scopus subject areas

  • Medicine(all)

Cite this