Evidence of autism in a psychiatrically hospitalized sample

Background: The similarity of the symptoms of ASD with other psychiatric disorders, and the fact that misdiagnosis may lead to inappropriate treatment, has led to interest in the prevalence of ASD in psychiatric populations. The four studies in this area have estimated the prevalence of ASD in adult psychiatric samples to be between 0.6% and 5.3%.

Objectives: To determine the potential prevalence of ASD among psychiatric inpatients and characteristics that discriminate between adults likely to have ASD and other psychiatric disorders.

Methods: The sample included 350 out of 396 patients in one state psychiatric hospital in Pennsylvania. Nursing staff completed the Social Responsiveness Scale (SRS) for each subject. Chart reviews were conducted to examine functioning and medical history. T-tests and chi-square tests were used to examine differences in clinical presentation, putative diagnoses, and medical history among patients scoring above 100 on the SRS (a score highly specific for autistic disorder in the general population) and patients scoring below 100.

Results: Twenty-one percent of patients received an SRS score over 100. They were significantly more likely than other patients to be diagnosed with undifferentiated schizophrenia (30% vs. 22%) and have indication in their charts of childhood onset or a “long history” of psychiatric problems (68% vs. 50%), not starting high school (20% vs. 8%), abnormal movements (20% vs. 10%), gastro-intestinal problems (34% vs. 23%), and mental retardation (15% vs. 5%). Analyses of differences in medication use and self-injurious behaviors are ongoing.

Conclusions: While not conclusive regarding the prevalence of ASD in a psychiatric inpatient sample, these findings are provocative and suggest the need for further research. We currently are conducting patient and family interviews to augment existing data. Improved diagnostic assessment for adults with ASD, especially those that discriminate ASD from the negative symptoms of schizophrenia, may have important treatment implications.