Design
Retrospective case series, records review.

Setting
A regional children's hospital and the King County (Washington) Medical Examiner's office.

Patients
Children younger than 19 years who sustained submersion injury between 1983 and 1991 and were hospitalized or autopsied.

Outcome Measures
Two pediatricians, using preestablished criteria, categorized abstracted case scenarios as either inflicted or unintentional events. The two groups were compared.

Results
Of 205 submersions, 16 (8%) were judged to have been inflicted. Objective physical signs of abuse and incompatibilities between the history and the child's stage of development or physical findings were common (69% and 50%, respectively). Inflicted submersion victims were likely to be young (median age, 2.1 years). They tended to be the youngest sibling in a large (three or more children) household. Social and demographic attributes of inflicted and unintentional submersion victims did not differ significantly. Bathtubs were the most common site for inflicted submersions (9/16 [56%]), and submersions in bathtubs were frequently inflicted (9/34 [26%]). Compared with unintentional submersion victims, children who were inflicted submersion victims were less likely to be revived by bystanders (relative risk, 0.10; confidence interval, 0.01 to 0.72) and were more likely to die (relative risk, 4.32; confidence interval, 1.40 to 21.43).

Conclusions
Enough childhood submersions are inflicted to warrant careful case evaluation. Bathtub submersion victims and children with physical and historical findings common to other forms of abuse are most likely to be the victims of inflicted submersion.

(Arch Pediatr Adolesc Med. 1996;150:298-303)