Magnitude and Correlates of Symptomatic, Global Illness, and Social Functioning Response; Out-of-Home Placement; and Length of Stay in 1,189 Consecutively Hospitalized Children and Adolescents With Mental Disorders.

Objective: To identify outcome predictors in hospitalized youth with mental disorders. Methods: This retrospective analysis of systematically recorded clinical parameters in youth hospitalized for psychiatric treatment in 2004-2015 assessed magnitude and correlates of symptom response (SR), global illness response (GIR), social functioning (SF), out-of-home placement (OOHP), and length of stay (LOS). Backward elimination regression analyses were performed to identify independent baseline correlates of each of the 5 outcomes, with R2 representing the variance explained by the independent correlates retained in the final model. Results: Across 1,189 youth (median age = 14.4 years; interquartile range = 11.6,16.1 years; range, 5-19 years; females = 61.5%), frequencies of coprimary outcomes were as follows: SR = 57.5% (statistically significant correlates = 13, R2 = 0.154), GIR = 30.0% (correlates = 5, R2 = 0.078), SF = 19.0% (correlates = 8, R2 = 0.207), OOHP recommendation = 35.2% (correlates = 13, R2 = 0.275), and mean ± SD LOS = 65.0 ± 37.5 days (correlates = 11, R2 = 0.219). In multivariable analyses, 11 factors were statistically significantly (P < .05) associated with > 1 poor outcome: 4 with 4 outcomes (disturbed social interaction, substance abuse/dependence symptoms; sole exception for both = LOS; disturbed drive/attention/impulse control, sole exception = OOHP; higher admission BMI percentile [but shorter LOS], sole exception = GIR), 3 with 3 outcomes (higher admission age [but good SF and shorter LOS], more abnormal psychosocial circumstances, more mental health diagnoses), and 4 with 2 outcomes (intelligence level [IQ] < 85, obsessive-compulsive disorder symptoms, disturbed social behavior, somatic findings). Additionally, 17 correlates were statistically significantly (P < .05) associated with 1 outcome, ie, SR = 6, OOHP = 5, LOS = 5, SF = 1. Conclusions: Higher admission BMI percentile, disturbed social interaction, disturbed drive/attention/impulse control, and substance abuse/dependence symptoms were independently associated with multiple poor outcomes in mentally ill youth requiring inpatient care. Knowledge of global and specific correlates of poor inpatient treatment outcomes may help inform treatment decisions.