Distinguishing suspected inflicted from accidental falls from a caregiver's arms: a pattern-based analysis of extremity injuries in children.

Caregiver-reported drop/fall from caregiver hold is a common history in children presenting with extremity injuries, but distinguishing accidental injury from clinician-suspected inflicted injury remains challenging. This study examined clinical and radiographic injury patterns associated with clinician classification as suspected inflicted versus accidental at the index presentation. This single-center cross-sectional study included consecutive children younger than 12 years presenting to the orthopedic emergency room after a reported drop/fall from caregiver hold between 16 June and 31 December 2025. Cases were classified at presentation as clinician-suspected inflicted drop/fall or accidental drop/fall based on structured medico-legal assessment. Injury-pattern variables were analyzed as study variables. Group comparisons used chi-square/Fisher's exact tests and the Mann-Whitney U test. Associations with clinician-suspected inflicted classification were examined using univariable logistic regression and a pre-specified exploratory multivariable Firth penalized logistic regression model. Fifty-eight children were included: 18/58 (31.0%) were classified as clinician-suspected inflicted drop/fall and 40/58 (69.0%) as accidental drop/fall. Clinician-suspected inflicted cases were older and more frequently had father reported as the caregiver. Bruising, displaced fractures, and multiple injuries/fractures were more frequent in clinician-suspected inflicted cases than in accidental cases. The combined pattern of bruising plus displaced fracture was observed in 12/18 (66.7%) clinician-suspected inflicted cases and 0/40 accidental cases in this sample, a zero-cell finding requiring cautious interpretation and external validation.In the pre-specified exploratory Firth penalized logistic regression model, father as the reported caregiver, bruising, and multiple injuries/fractures remained associated with clinician-suspected inflicted classification, rather than with independently confirmed abuse. In this single-center cohort, selected clinical and radiographic injury patterns were associated with clinician-suspected inflicted classification after caregiver-reported drop/fall from caregiver hold. Because abuse was not independently confirmed, these findings should be interpreted as risk indicators to support early safeguarding triage and careful documentation, not as stand-alone evidence of abuse or intent. Because clinician classification was not independent of routinely available clinical and radiographic findings, incorporation bias may have affected the observed associations. Bidirectional misclassification between accidental and clinician-suspected inflicted classifications may also have occurred. The absence of a full skeletal survey and uniformly documented complete examination may have led to missed occult injuries.