The Value of the AQ in Outpatient Mental Health Service

Background:

In clinical practice, the Autism Spectrum Quotient (AQ; Baron-Cohen et al., 2001) is often used to measure the degree to which adults exhibit autistic traits and to decide whether further ASD assessment is required (case identification).

The AQ is a self-report questionnaire. Since self-insight might be diminished in individuals with ASD, the perspective of an informant is important. Therefore, an AQ Dutch spouse-version was developed (Blijd-Hoogewys, 2014).

There is debate on whether there are gender differences concerning the expression of autistic traits, and whether these differences influence case identification.

Objectives:

The aim was to study the value of the AQ as an instrument for case identification, comparing self-report and spouse-report, and gender differences.

Methods:

The sample consisted of Dutch patients with a suspected ASD and an average intelligence. They were referred to a specialized ASD outpatient mental health service.

In total, there were 194 AQ self-reports (n = 106 men, n = 88 women) and 49 AQ spouse-reports (n = 26 men, n = 23 women). AQ-scores were analyzed, using a clinical cut-off point of 26 (dichotomous scoring).

Results:

There was no significant difference in AQ-score between self-report and spouse-report (M = 28.08, SD = 9.14; M = 27.02, SD = 8.03 respectively; t = .61, df = 96, p = .54).  Concerning the classification of the AQ-score as clinical (≥ 26) or non-clinical, there was a moderate agreement between self-report and spouse-report (κ = .58, p < .001).

The AQ-scores of men and woman did not differ significantly on the self-report (M = 25.06, SD = 9.28; M = 25.51, SD = 9.11 respectively; t =-.34, df = 192, p = .73). They did differ significantly on the spouse-report (men: M = 29.60, SD = 6.57; women: M = 24.13, SD = 8.67; t = 2.50, df = 47, p =. 02).

Results of the self-reports showed a relation between AQ-score above/below cut-off and the presence/absence of an ASD-diagnosis (X² = 28.15, df = 1, p < .001). Overall, 76% of participants with clinically elevated AQ-scores were diagnosed with ASD (versus 38% of participants with non-clinical AQ-scores). The predictive values of the AQ differed for men and women. For men, sensitivity was 58% and specificity was 66% (PPV = 0.73, NPV = 0.50). For women, sensitivity was 73% and specificity was 82% (PPV = 0.80, NPV = 0.75). The same pattern was seen in spouse-reports.

Conclusions:

The AQ is moderately valuable for case-identification in an outpatient mental health service. The AQ self-report and AQ spouse-report give roughly similar information about autistic traits. There are gender differences though; female spouses report their men to exhibit more ASD traits than reported by the men themselves. Also, the AQ self-report of women seems to be more predictive for an ASD diagnosis than those of men.

Note that the AQ is not intended to be diagnostic. If there are clinically significant levels of autistic traits, a comprehensive diagnostic evaluation is warranted.