Sensational headlines describing police entanglements with young children have prompted questions about how often these incidents occur and why. The authors of this cross-sectional study examined the factors associated with police versus nonpolice arrivals to the psychiatric emergency room and those predicting subsequent police arrivals. Electronic medical records of children ages 5.0-12.9 years brought to a comprehensive psychiatric emergency program (CPEP) at a university hospital were reviewed to determine whether a child was brought by police ("police arrival") in response to a 911 call by school personnel, a mental health or other medical professional, or a caregiver. Extracted data included the child's age, sex, race-ethnicity, family makeup, insurance status, arrival status, referral source, diagnosis, disposition, treatment, number of CPEP and police encounters, and occurrences of aggression and suicidality. Multilevel and ordinary logistic regression models were used to identify factors associated with a first and subsequent police arrival. Of 339 children with CPEP encounters from September 2017 to April 2018, 103 (30%) had had at least one police arrival. Children brought by police were more likely than peers brought by caregivers to be Black or Latinx, have Medicaid, come from families without two parents, and have aggressive outbursts or suicidal behavior. Results from multilevel logistic regression indicated that aggressive outbursts and suicidality were most significantly and consistently associated with experiencing both a first and subsequent police arrival. Clinical and sociodemographic differences in police arrivals highlight the need for a comprehensive systems approach for children, especially marginalized youths, who need psychiatric emergency care.