Comparison of Behavioral Outcomes and Crisis Service Utilization Across the Six Specialized Inpatient Units in Phase I of the Autism Inpatient Collection (AIC)

Background:

Children and adolescents with Autism Spectrum Disorder (ASD) are psychiatrically hospitalized at a much higher rate than non-ASD children  due to serious emotional and behavioral problems including aggression, self-injury and tantrum-like behaviors. There is no comparative information, however, examining phenotypic differences, changes in serious problem behaviors, and use of crisis services, across specialized child psychiatry units.

Objectives:

Test for differences in demographic and clinical factors, patient characteristics, and site characteristics between six specialized child psychiatric units. Examine changes in child problem behavior over time (admission, discharge, and two month follow-up) both across the full sample and between sites. Examine changes in patients’ use of crisis services two months prior to admission versus two months post-discharge.

Methods:

350 youth aged 4-20 years, with an ADOS-2 confirmed ASD diagnosis, admitted to specialized inpatient psychiatry units, were prospectively enrolled in the Autism Inpatient Collection (AIC) study, a six-site consortium investigating phenotypes, genotypes and behavioral outcomes among children. Parents completed measures at admission, discharge and two months post-discharge, and reported their child’s demographics, problem behaviors (Aberrant Behavior Checklist – Irritability subscale, ABC-I), expressive communication and adaptive behavior (Vineland – 2), and crisis service utilization for the two months before and after hospitalization. Analysis of variance and chi-square tests were used to quantify site differences between subjects’ demographic and clinical variables as continuous and categorical variables respectively. Paired t-tests were calculated to examine change in crisis services utilization over time. Changes in problem behavior over time were examined using multilevel model repeated measures analysis of covariance.

Results:

Site differences were found for children’s ethnicity, race, non-verbal IQ, expressive communication, and hospital length of stay (see Table 1). There was a significant decrease in child problem behavior between admission and discharge for all sites, and slight increases in problem behaviors post-discharge, which varied by site. Improvement in problem behaviors, however, was not uniform across sites, even after controlling for significant site differences in our sample. There were also notable differences in the two subdomains of the ABC-I, tantrum-like behaviors (TLB) and self-injurious behaviors (SIB). SIB and TLB decreased while children were hospitalized, but we found an increase in TLB after discharge. Length of stay was the only statistically significant covariate in all models. Lower expressive communication scores predicted SIB but not TLB. We found an overall reduction in post-hospitalization utilization of crisis services for all sites combined. The decrease in emergency department visits and police contacts reached statistical significance at only three sites.

Conclusions:

Hospitalization in specialized child psychiatry units appeared to be effective in reducing the severity of child behavior problems from admission to 2-months post discharge for children and adolescents with ASD. Understanding the factors that contribute to variations in behavioral outcomes and crisis service utilization may enhance treatment for this population. Examining the comparative effectiveness of hospitalization in specialized units and general child psychiatric units, while matching for ASD symptom and behavioral severity in future studies would further inform public policy decisions on investing for services for this growing, high-need population.