Autism-spectrum quotient Japanese version measures mental health problems other than autistic traits

The purpose of the present paper was to examine the extent to which the Autism Spectrum Quotient Japanese version (AQ-J) measures mental health problems other than autistic traits, with the 12-item General Health Questionnaire Japanese version (GHQ-12) as a criterion. A questionnaire involving AQ-J and GHQ-12 was sent to 2000 adults aged 20-39 randomly selected from the general population, and intact data were obtained from 215 (mean age, 30.4 years; 86 male). The AQ-J score was significantly associated with the GHQ-12 score (r=0.518) and was significantly higher in 111 scoring >or=4 on the GHQ-12 (mean=24.8+/-6.5) than in 104 scoring <4 (mean=19.4+/-5.5). At a cut-off 22, the AQ-J had modest sensitivity (0.64) and specificity (0.66) for predicting mental health problems. The 21-item mental health AQ-J (AQJ-21MH; range, 0-21; cut-off, 9; sensitivity, 0.69; specificity, 0.76), items of which had a significant odds ratio (OR) for GHQ-12>or=4, and the AQ-J-4MH (range, 0-4; cut-off, 2; sensitivity, 0.68; specificity, 0.74) consisting of four items with a significant OR adjusted for collinearity selected from the 21 items by multiple logistic regression, were more efficient than the AQ-J. Because the AQ-J and its short forms measure mental health problems other than autistic traits, it is important to consider such problems in interpreting AQ-J scores to identify persons who may need professional help when screening normally intelligent adolescents and adults with pervasive developmental disorders.