Prevalence of Autism in a Psychiatric Inpatient Population

Background: Previous studies suggest that autism spectrum disorders (ASD) are over-represented and under-diagnosed in adult psychiatric populations.  Adults in state psychiatric hospitals, who received their putative diagnoses prior to the changing conceptualization and increased awareness of ASD, may have undiagnosed ASD. In psychiatrically hospitalized adults, ASD symptoms may appear similar to the negative symptoms of schizophrenia. Four studies have estimated the prevalence of ASD in adult psychiatric populations (inpatient and outpatient) to be between 0.6% and 5.3%; no more than 10% of subjects in these studies who were found to have ASD were previously diagnosed as such. They were most commonly diagnosed with schizophrenia instead.
Objectives: The objective of this study is three-fold: to determine the prevalence of ASD among psychiatric inpatients; evaluate the use of the Social Responsiveness Scale (SRS) as a screening instrument in this sample; and to identify characteristics that discriminate between severely impaired adults with ASD and other psychiatric disorders.

Methods: The sample included 263 civilly-committed patients in one state psychiatric hospital in Pennsylvania.  Nursing staff completed the SRS for each patient as part of standard of care. All patients with scores ≥ 100 on the SRS and a stratified random sample of those with lower scores were consented to conduct in-depth chart reviews and contact family members to conduct the Autism Diagnostic Interview-Revised (ADI-R). Chart reviews focused on developmental history, paying particular attention to age of onset and clinical features indicative of ASD.  Patients also completed clinical interviews and a neurobehavioral battery to assess the presence and severity of psychotic symptoms and cognitive impairment.  Case conferences with two psychiatrists and the team of assessing psychologists will be held for all patients who meet ASD criteria on the ADI-R and a sub-sample of other patients.

Results: Twenty percent of patients received a SRS score ≥100. ADI-R administration and case conferences are ongoing. To date, 31% of those with SRS scores ≥100 met cutoff criteria for ASD on the ADI-R. Based on chart review and clinical observation (including the ADI-R), a third of subjects scoring positive on the ADI-R met criteria for ASD after case conferences (10% of the total sample).

Conclusions: A larger proportion of previously undiagnosed adults in this state psychiatric hospital met criteria for ASD than has been found in previous studies. Previously validated screening and diagnostic instruments, however, including the Social Responsiveness Scale (SRS) and the Autism Diagnostic Interview-Revised (ADI-R), did not demonstrate the same accuracy in this sample as they have in the general population. Improved screening and diagnostic assessments for more severely impaired adults with ASD, especially those that discriminate ASD from other psychiatric disorders, may have important implications for their treatment and supports.  The results of this study will aid in efforts to determine the prevalence of ASD among institutionalized adults, increase understanding of the prevalence of ASD in adults, and guide policy and practice regarding diagnostic practices and service delivery to adults with ASD.