Who comes back? Characteristics and predictors of return to emergency department services for pediatric mental health care

Acad Emerg Med. 2010 Feb;17(2):177-86. doi: 10.1111/j.1553-2712.2009.00633.x.


Objectives: The objective of this study was to investigate predictors of emergency department (ED) return visits for pediatric mental health care. The authors hypothesized that through the identification of clinical and health system variables that predict return ED visits, which children and adolescents would benefit from targeted interventions for persistent mental health needs could be determined.

Methods: Data on 16,154 presentations by 12,589 pediatric patients (<or=17 years old) were examined from 2002 to 2006, using the Ambulatory Care Classification System (ACCS), a provincewide database for Alberta, Canada. Multivariable logistic regressions identified predictors, while survival analyses estimated time to ED return.

Results: In the multivariable analysis, there were four patient factors significantly associated with ED return. Male sex (odds ratio [OR] = 0.78; 99% confidence interval [CI] = 0.69 to 0.89) was associated with a lower rate of return, as was child age. The likelihood of ED return increased with age. Children <or=5 years (OR = 0.26; 99% CI = 0.14 to 0.46) and between ages 6 and 12 (OR = 0.64; 99% CI = 0.51 to 0.79) were less likely to return, compared to 13- to 17-year-olds. Patients with families receiving full assistance for covering government health care premiums were more likely to return compared to those with no assistance (OR = 1.59; 99% CI = 1.33 to 1.91). Patients were more likely to return if their initial presentation was for a mood disorder (OR = 1.72; 99% CI = 1.46 to 2.01) or psychotic-related illness (OR = 2.53; 99% CI = 1.80 to 3.56). There were two modest health care system predictors in the model. The likelihood of return decreased for patients triaged as nonurgent (OR = 0.62; 99% CI = 0.45 to 0.87) versus those triaged as urgent (level 3 acuity) and increased for patients with visits to general (vs. pediatric) EDs (OR = 1.25; 99% CI = 1.03 to 1.52). ED region (urban vs. rural) did not predict return. Within 72 hours of discharge, 6.1 and 8.7% of patients diagnosed with a mood disorder and psychotic-related illness, respectively, returned to the ED. Throughout the study period, 28.5 and 36.6% of these diagnostic populations, respectively, returned to the ED.

Conclusions: Among children and adolescents who accessed the ED for mental health concerns, being female, older in age, in receipt of social assistance, and having an initial visit for a mood disorder or psychotic-related illness were associated with return for further care. How patient presentations were triaged and whether visits were made to a pediatric or general ED also affected the likelihood of return.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adolescent
  • Alberta
  • Child
  • Child, Preschool
  • Emergency Service, Hospital / statistics & numerical data*
  • Female
  • Hospitals, General / statistics & numerical data
  • Hospitals, Pediatric
  • Humans
  • Infant
  • Logistic Models
  • Male
  • Mental Health Services / organization & administration
  • Mental Health Services / statistics & numerical data*
  • Mood Disorders
  • Multivariate Analysis
  • Retrospective Studies
  • Triage