ࡱ > bjbjss A I4 \ + + + 8 c 4 + F " 7 9 9 9 9 9 9 $ <# ` ] 9 ] * * * 7 * 7 * * 0 + # 0 , # ~ # # $ * ] ] * # : Contents of the Supplementary material
A) Supplementary Method pages 13
B) Supplementary Results pages 45
C) Supplementary Figs S1 and S2 pages 67
D) Supplementary Table S1 page 8
E) Supplementary Table S2 pages 910
Supplementary Method
Assessment Procedure and Missing Items
The AQ50 questionnaire (from which AQ50 and AQ10 scores were derived) was obtained for all participants, as part of the questionnaire pack sent to all patients in the BGC clinic. In the majority of cases the AQ50 was completed before the ADOS-G/ADI-R interviews took place. The AQ, ADOS-G, and ADI-R were scored independently, to ensure that scores on one measure would not influence the scoring of the other instruments.
We only calculated AQ10 scores when all ten of the required items were completed, considering it impossible to accurately correct for missing values for such a small pool of items. In the case of the AQ50, questionnaires with 10% or more blank or improperly completed items (5+ items) were discarded. AQ50 questionnaires with 1 - 4 missing items were corrected as follows: corrected score=raw score x (50 / (50 missing items). This is equivalent to imputing missing responses as the mean of completed responses. 85 of the 456 AQ50s (19% of the total) were corrected in this way.
Exclusion Criteria
All individuals who completed the AQ50 as part of the questionnaire pack, and who were seen at the BGC, were included in the study, unless they: i) did not attend enough of the assessment process to provide both AQ scores and to receive a clinical expert diagnosis; or if they ii) declined consent to have their anonymised data used in research. Consent for research was obtained at the beginning of the assessment process. Individuals who did consent were given a further opportunity to opt out at a later date; or if they iii) Were aged below 18 years of age at the time of the assessment.
Measures of Psychiatric Disorders (Comorbidities)
Comorbid mental health problems are common in adults with ASD. To investigate and control for the presence of these disorders, we examined records from the comprehensive psychiatric interview (based on the DSM-IV) which forms part of the assessment process for all patients seen at the Behavioural Genetics Clinic. These assessments were conducted by an experienced psychiatrist.
For the purposes of this study we examined the following six disorders i) major depressive disorder (MDD), ii) panic disorder and/or agoraphobia, and iii) generalised anxiety disorder (GAD) iv) social anxiety disorder and/or social phobia v) attention-deficit hyperactivity disorder (ADHD) vi) obsessive compulsive disorder (OCD). Other disorders such as schizophrenia, substance abuse and personality disorders were diagnosed in only a small number of participants (10 or fewer), precluding statistical treatment, so we did not include them in this analysis.
We coded a disorder as 'present' if the patient was given a current definite diagnosis of the disorder at the assessment. Each of the six disorders was coded as a separate binary variable where 0=absent, 1=present. No hierarchy rules were applied.
Data Analysis and Statistics
The data was analysed using SPSS 21. To examine the performance of the screening measures (see above), we tested the performance of the AQ10 and AQ50 screening measures as a predictor of clinical diagnosis. To determine the statistical significance of the AQ10 and AQ50, we used chi squared tests with scoring above threshold vs. receiving a clinical diagnosis, both as binary variables. We further examined test performance using the following metrics: accuracy (% of predictions that were correct), sensitivity (correct positives / total positives), specificity (correct negatives / total negatives), positive predictive value (PPV), negative predictive value (NPV), and Youdens J ADDIN EN.CITE Youden1950588[23]58858817Youden, W. J.Index for rating diagnostic testsCancerCancerCancerCancerCancerCancer32-5311950/01/01*Neoplasms1950Jan0008-543X (Print)
0008-543X (Linking)15405679http://www.ncbi.nlm.nih.gov/pubmed/15405679eng(Youden 1950), a measure of the information content of the test, equal to: Sensitivity + Specificity 1. J ranges from -1 to 1, with 0 indicating that the test provides no useful information. The values of accuracy, sensitivity, specificity were interpreted as follows: values <0.70 are poor; 0.70-0.79 fair; 0.80-0.89 good; 0.90-1.00 excellent ADDIN EN.CITE Cicchetti1995608[25]60860817Cicchetti, D.Volkmar, F.Klin, A.Showalter, D. R.Diagnosing autism using ICD-10 criteria: A comparison of neural networks and standard multivariate procedures.Child NeuropsychologyChild Neuropsychology26-3711995(Cicchetti et al. 1995). These guidelines have been used previously when evaluating the properties of ASD screening tools ADDIN EN.CITE ADDIN EN.CITE.DATA (Oosterling et al. 2009).
To further investigate the psychometric properties of the AQ at a range of cut-offs, we calculated Receiver Operating Characteristic (ROC) curves for i) the AQ10 and ii) the AQ50 as predictors of clinical judgement. The outcome was the Area Under the Curve (AUC), which is 0.5 under the null hypothesis of no significant predictive power at any cut-off. If the AUC was significantly different from 0.5, the optimal cut-off and the Youdens J at that cut-off were calculated. For the presence of other psychiatric diagnoses, as these data were nominal, chi-squared tests were used in order to determine whether rates of each comorbidity differed across different participant groups (e.g. those receiving a clinical diagnosis of ASD vs. those not receiving one.)
Supplementary Results
Comparison of the ADOS-G and ADI-R As Definitions of Caseness
To determine whether AQ scores predict scoring above the ASD threshold on the ADOS-G and ADI-R criteria, we repeated our primary analyses using ADOS-G and ADI-R scores to define caseness in place of clinical consensus diagnosis. We applied the standard, v a l i d a t e d c u t - o f f s ( A D O S - G : c o m m u n i c a t i o n e" 2 a n d s o c i a l e"4 a n d s o c i a l + c o m m u n i c a t i o n e" 7 . A D I - R : s o c i a l e" 1 0 a n d c o m m u n i c a t i o n e" 8 a n d r e p e t i t i v e b e h a v i o u r s e" 3 . ) S e e S u p p l e m e n t a r y T a b l e S 2 . W i t h t h e A D O S - G , n o v a r i a n t o f t h e A Q p r e d i c t e d c a s e n e s s b e t t e r t h a n c h a n c e ( a l l 2 <