Comparison of DSM-IV-TR and DSM-5 for ASD

The study will involve the following aims and hypotheses:

Aim 1: To prospectively compare DSM-IV-TR and DSM-5 criteria for Autism Spectrum Disorder in a sample of 250 children and adolescents who meet DSM-IV-TR criteria for ASD.

Aim 2: To examine the impact of IQ, comorbid behavior problems, age and ASD symptom severity on the specificity and sensitivity of DSM-5 diagnostic criteria.

Hypothesis 1: DSM-5 diagnostic criteria will have good specificity but poor sensitivity when comparing results with DSM-IV-TR ASD diagnostic criteria.

Hypothesis 2: Lower IQ will be associated with higher DSM-IV/DSM-5 concordance.

METHODS

Families and children will be recruited from seven sites affiliated with the Autism Treatment Network. An IRB submission and consent form will be developed for each of the participating programs. Parents (along with child assent and/or consent, depending upon the study site) will sign informed consent to participate in the study. Participants must be between the ages of 2.0 and 17.11 years who undergo an ATN assessment at one of the seven sites. Both subjects who are found to meet criteria of ASD/PDD and those who do not meet criteria (after completing the ATN assessment) will be enrolled. Therefore, the only inclusion criteria will be that the child be between 2.0 and 17.11 years of age and be referred for an ATN diagnostic evaluation. The only exclusion criteria will be when the ADOS is determined by the clinician to be "invalid." Examples may include a child who is extremely non-compliant, severely depressed, or unresponsive. The following measures will be obtained:

ADOS:

All participants will be administered the Autism Diagnostic Observation Schedule 2 (Lord et al., 2012). The ADOS2 is a standardized, interactive protocol for direct observation (45 minutes) of social and communicative behavior associated with ASD, and consists of structured and semi-structured methods for interaction. This assessment tool is designed to assess a wide range of children and adolescents, from nonverbal/barely verbal children to adolescents and adults with fluent speech.

Assessment of IQ:

All participants will have an updated assessment of full IQ using any one of the following tools: the Stanford-Binet V (Roid, 2003), Wechsler Intelligence Scales for Children IV (Wechsler, 2003), the Wechsler Preschool and Primary Scale of Intelligence - Third Edition (Wechsler, 2002), Differential Ability Scales, 2nd Edition, (Elliot, 2007), Wechsler Abbreviated Scale of Intelligence 2 (Wechsler, 2011) or Mullen Scales of Early Learning AGS Edition (Mullen, 1989). If a child has had a full-scale IQ assessment conducted within one year, this will be accepted in place of an updated evaluation.

Assessment of Behavior:

Child Behavior Checklist (CBCL): The CBCL (Achenbach 2001; Achenbach \& Rescorla, 2000) is a parent completed questionnaire comprising of 100 items (scored on a 3-point likert scale from "not true" to "very true or often true"). There is a preschool version (1.5-5 years) and a school version (6-18 years). A range of subscales are derived, including externalizing and internalizing disorders as well as more specific co-morbidities such as anxiety, ADHD and PDD.

Aberrant Behavior Checklist (ABC). The ABC (Aman \& Singh, 1986) is a standardized scale comprising 58 items, for assessing problem behavior in subjects with developmental disabilities. The ABC was empirically derived from ratings on approximately 1,000 subjects, and the items resolve onto five subscales: (I). Irritability (15 items), (II) Lethargy/Social Withdrawal (16 items), (III) Stereotypic Behavior (7 items), (IV) Hyperactivity (16 items), and (V) Inappropriate Speech (4 items). The primary caregiver will serve as the rater.

Additional Information:

Demographic Questionnaire: Parent will complete the ATN demographic questionnaire, which inquires about family make-up and education, medical history, etc.

Supplemental Information: At all sites, clinicians will conduct clinical interviews and informal observations sufficient to gather enough information to inform DSM-IV and DSM-5 criteria determination. Sites may also use supplemental information that is available to them. For example, prior evaluations may have been conducted. In addition, sites also may use a variety of additional questionnaires, such as the Social Communication Questionnaire (ref) or the Social Responsiveness Scale (SRS). Questionnaires or reports from a child's school are also frequently obtained as part of the assessment process. Such information can be used by the evaluation team to determine the final diagnosis.

Diagnostic Determination:

At the conclusion of the evaluation, the assessor (along with any other team members) will use the information obtained from the above sources to complete the DSM-IV checklist. If more than one clinician was involved in the assessment, a consensus will be obtained. After the DSM-IV checklist has been completed, the DSM-5 checklist will be completed using the same information (including the consensus scores on the DSM-IV checklist).

DSM-IV Checklist: The DSM-IV checklist (APA, 2002) is a clinician-completed questionnaire that lists the X symptoms for autistic disorder. Each is scored as being "present/absent." Finally, the clinician notes which, if any, PDD diagnosis was made (based upon the results of the DSM-IV checklist).

DSM-5 Checklist: The DSM-5 checklist is a clinician-completed questionnaire that lists the seven symptoms for autism spectrum disorder from DSM5 (APA 2013). Each is scored as being "present/absent." In addition, there is a column for noting if the symptom is currently present or was only reported in the past ("present by history"). Finally, the clinician notes if an ASD diagnosis was made (based upon the results of the DSM-5 checklist).

Final Diagnosis:

The following information regarding each subject's final diagnosis will be entered onto the Diagnostic Form:

ADOS Results: Autism, Autism Spectrum, Nonspectrum

DSM-IV Diagnosis: Autism, PDD NOS, Asperger's Disorder, Nonspectrum

DSM-5 Diagnosis: ASD, Social Communication Disorder, Nonspectrum

Consensus Diagnosis Old: Autism, PDD NOS, Asperger's, Nonspectrum

Consensus Diagnosis New: Autism Spectrum Disorder (Level 1, 2 or 3); Nonspectrum

Item-level and algorithim scores: Should be included in the database for the ADOS-2 and both DSM-IV and DSM-5 checklists.

Additional Variables:

The following information will also be noted:

• List of additional measures used to make the diagnostic decision (e.g., SRS School, SRS Home, SCQ, Vineland)
• If child is Nonspectrum, indicate other primary diagnoses (e.g., OCD, Anxiety Disorder, ADHD, RAD)

Statistical Analysis

Hypothesis 1: DSM-5 diagnostic criteria will have good specificity but poor sensitivity when comparing results with DSM-IV-TR ASD diagnostic criteria. A Chi-Square test will be used to compare the rate of ASD diagnosis (autistic disorder, PDD-NOS, Asperger's disorder) using DSM-IV-TR versus DSM-5 criteria. Based upon the DSM-IV-TR diagnosis (which will serve as the gold standard), both specificity and sensitivity for DSM-5 will be calculated. Sensitivity will be calculated as the proportion of individuals who meet DSM-IV-TR criteria for an ASD (including autistic disorder, PDD-NOS, and Asperger's disorder) who also meet DSM-IV criteria for an ASD (i.e., true positives). Specificity will be calculated as the proportion of individuals who do not meet DSM-IV-TR criteria who also do not meet DSM-5 criteria (i.e., true negatives). Sensitivity and specificity will also be calculated for each of the three DSM-IV subtypes (autistic disorder, PDD NOS, and Asperger's disorder).

Hypothesis 2: Lower IQ will be associated with greater DSM-IV/DSM-5 concordance. Concordance will be defined as agreement between DSM-IV and DSM-5 regarding the presence or absence of an ASD diagnosis; discordance will be defined as disagreement (either DSM-IV positive and DSM-5 negative, or DSM-IV negative and DSM-5 positive). Logistic regression will be used to determine whether Full Scale IQ score predicts concordance (versus discordance).