Psychosis in Autism: How to Assess Risk

New research helps us understand how to assess psychosis in autism.

Posted Dec 24, 2020

As discussed in other posts on this page, autism spectrum disorder (ASD) and schizophrenia share a diagnostic history. ASD was originally thought of as a symptom of schizophrenia, prior to becoming its own unique diagnosis. Psychosis, which is defined as a state in which someone has trouble distinguishing between what is real and what is not, is a hallmark of schizophrenia. Even though we now recognize that ASD and schizophrenia are different diagnoses, research suggests that there is an increased risk of psychosis in individuals with ASD.

There is a growing body of research suggesting that early identification and intervention in psychosis can improve future outcomes. Essentially, the earlier we can identify that someone is at risk for (or experiencing early signs of) psychosis, and the sooner we can get them help, the better their outcomes are likely to be. Given the above, it is important to accurately identify early symptoms of psychosis in individuals with ASD. One problem with this, however, is a lack of research on whether instruments designed to assess risk for psychosis work accurately with individuals with ASD.

A paper recently published in the journal Autism measured the accuracy of the Structured Interview for Psychosis-Risk (SIPS) in adolescents with and without ASD. The researchers administered the SIPS to 21 adolescents (ages 12-18) with ASD and 22 adolescents without ASD who had cognitive and linguistic abilities in the average range. The researchers also asked the adolescents to rate their satisfaction with the SIPS (i.e. asked if the questions were easy to follow, understandable, etc). The research team thought that adolescents with ASD might have trouble responding to questions that required high levels of language abstraction. For example, one question on the SIPS asks, "Do you ever see things that others can't or don't seem to see?" Answering this question accurately within the context of the SIPS requires the participant to understand that the examiner is not asking about having highly accurate vision (i.e. being able to see things that are far away), but rather is asking about visual hallucinations. As individuals with ASD often have trouble with non-literal language and/or abstraction, this type of question might be prone to errors for this population.

Contrary to expectations, the study found that adolescents with ASD did not differ from their neurotypical peers in understanding questions on the SIPS. Also, adolescents with and without ASD did not differ in their ratings of satisfaction with the SIPS. 

Interestingly, across both groups, the ability to understand ambiguous language predicted response errors. That is, in both groups of adolescents, scores on a standardized test of ambiguous language was the best predictor of whether a participant would make response errors. The test of ambiguous language is from a larger language assessment and asks participants to generate multiple meanings of a sentence. For example, "Did you see that fly?" could either mean, "Did you see that insect?" or "Did you see that flying object?" Regardless of diagnostic status, adolescents who had more difficulties with ambiguous language were more likely to make errors in responding to SIPS questions. 

Overall, these findings suggest that using the SIPS for adolescents with ASD is valid. 

What's Next

It is critical to make sure that standardized measures for assessing psychosis are accurate for individuals with ASD. This study is important because it suggests that the SIPS is accurate for individuals with ASD.

This study also suggests that for everyone — whether or not a diagnosis of ASD is present — the ability to use abstract language correlates with more accurate SIPS responses. Essentially, people who are better at understanding multiple meanings of words/concepts will be more accurate when responding to questions on the SIPS. 

More research is needed to understand how to accurately assess psychosis in those who have language difficulties. One helpful tool might be to use "understanding checks" throughout the measure. For example, after asking whether someone often "sees things that others cannot," a clinician should follow up to ensure that the participant understands that the question is asking about hallucinations rather than vision accuracy. 

Overall, this study is exciting because it suggests that the SIPS is equally accurate in adolescents with and without ASD who have average cognitive and language abilities. 


Wilson, C. S., Anthony, L., Kenworthy, L., Fleischman, R., Demro, C., Andorko, N., Chelsea Armour, A., & Schiffman, J. (2020). Feasibility of psychosis risk assessment for adolescents diagnosed with autism. Autism, 24(4), 834–850.