Extension of the PDD Behavior Inventory to Adolescents

Background:  

The PDD Behavior Inventory (PDDBI) is a reliable and valid informant-based rating scale for ASD designed for assessment, diagnosis, and treatment planning.  Studies have indicated that scores from the PDDBI are good indicators of response to intervention, diagnosis, and outcome, and are also associated with certain genetic factors.  The PDDBI was designed to cover ages ranging from 18 months to 12 years, five months.  However, this age range restricts clinicians and researchers to younger children and prevents use of the PDDBI for longer term clinical follow-up as well as research studies of adolescents with ASD. 

Objectives:  

In a large cohort of adolescents, examine the cross-sectional effects of age on PDDBI raw cluster scores and total scores; compare internal consistency and inter-rater reliability with established measures for the standardization sample; and examine the construct and criterion-related validity of the PDDBI for this older group.

Methods:  

PDDBIs were completed by 138 parents and 100 teachers of adolescents with ASD, ages 12 years, 6 months to 18 years, 5 months.  The sample was divided into three age groups (years:months): 12:6 to 14:5; 14:6 to 16:5; and 16:6 to 18:5.  Raw cluster and raw total scores were compared across age groups using ANOVAs.  Internal consistencies were evaluated with alpha coefficients and factorial validity with principal components analyses (PCA).  Criterion-related validity correlations for the PDDBI scores were examined against ADI-R domain scores and against the Aberrant Behavior Checklist (ABC) in a subset (n=18) of parents.  

Results:  

Across age group, significant attrition (>20%) in completion of the Gestural Approach, Social Play, and Imaginative Play item clusters for the Social Approach (SOCAPP) domain was noted.  Regression techniques were then used to estimate missing cluster scores based on the remaining clusters.  ANOVAs revealed no significant difference in the revised SOCAPP mean score, as well as all other domain mean scores, across age groups.  Internal consistencies were high, ranging from 0.77 to 0.98.  Parent-teacher agreement, within a 6 month period, was good for most of the domain scores but poor for the RITUAL, SOCPP, SEMPP, and LMRL domains.  Factorial validity replicated previous findings of two behavioral dimensions: Approach/Withdrawal Problems and Receptive/Expressive Social Communication Abilities.  Agreement between Approach/Withdrawal Problems domain scores and corresponding ABC factor scores was very good and agreement between the PDDBI and ADI-R was remarkably good given that the ADI-R scores were historical in nature.

Conclusions:  

These analyses confirm both the reliability and validity of the PDDBI for adolescents with ASD given minor modification in scoring of the SOCAPP domain.