Poster #45 - Reliability of the Vineland Adaptive Behavior Scales, Third Edition in a Heterogeneous Clinical Research Sample

• In Event: 3-206 - Poster Session 13
  In Poster Session: PS 13 Section - Developmental Disabilities

Integrative Statement

Introduction: A DSM-5 diagnosis of Intellectual Disability is conditioned on impairments in cognitive functioning and adaptive behavior. The Vineland Scales of Adaptive Behavior (VABS) is a parent interview commonly used to assess adaptive behavior. The recently published third edition includes several updates; expanded items to capture a broader developmental range may be particularly relevant to the use of the VABS-3 for diagnosing Intellectual Disability. No data independent of the publisher’s reports are yet available on the comparability of the VABS-3 and the VABS-II. We aimed to: (1) evaluate the reliability of the VABS-3 with the VABS-II, with an emphasis on understanding how differences between scales may impact individuals in the range of intellectual disability, and 2) compare each form’s construct validity, operationalized as discrepancy with full scale IQ (FSIQ). We expected VABS-3:VABS-II correlations to be strong, but given the additional items at lower levels of ability, we hypothesized that VABS-3 scores would be lower than VABS-II. Relatedly, we expected the VABS-3 composite score to be more closely aligned to FSIQ than VABS-II.
Methods: Both versions of the VABS were administered contemporaneously to 97 parents of individuals in two clinical research settings that assess individuals with neurodevelopmental disorders. Participants in this convenience sample ranged from 1.3-50 years (mean age=13.3±10.8) and the majority had impaired cognitive functioning (mean FSIQ=64.6±28.9, range=6.16-131). The target variables were the domain standard scores (Communication, Socialization, Daily Living Skills, Motor) and the Adaptive Behavior Composite (ABC), which have a population mean of 100 and SD of 15. Given the potential for floor effects in this sample, we used Spearman correlations to evaluate the reliability between the two Vineland versions, and a paired t-test was used to evaluate differences in scores.
Results: Correlations between VABS versions were very strong; VABS-II scores tended to be higher than VABS-3 with mean differences ranging from ½ to ¾ of a standard deviation (see Table 1 and Figure 1A). Seven of 37 participants with VABS-II scores >70 had scores <70 (in the ID range) on the VABS-3. While the VABS-II ABC was significantly higher than the FSIQ (95% CI for mean difference: 0.7 – 9.3, p=.02), VABS-3 ABC did not differ from FSIQ (95% CI for mean difference: -4.9 – 2.9, p=.61; see Figure 1B). Further analysis of the subdomain scaled scores and age equivalents will be presented.
Discussion: This study suggests that while VABS-II and VABS-3 scores are very strongly correlated, the VABS-3 tends to yield lower scores than the VABS-II. However, these scores were more consistent with participants’ directly assessed cognitive function (FSIQ), which suggests improved construct validity with respect to the diagnosis of intellectual disability. Although this study is based on a convenience sample, these data are among the first to compare the VABS-3 to the VABS-II. A major strength of this study is the high proportion of individuals with low levels of adaptive behavior and impaired cognitive functioning. Results will directly inform the validity of the VABS-3 for one of its most common uses: the diagnosis of Intellectual Disability.

Authors

• Cristan Farmer, NIMH
  Non-Presenting Author

• Dee Adedipe, NIMH
  Presenting Author

• Vanessa Bal, Rutgers University
  Non-Presenting Author

• Megan Fok, Rutgers University
  Non-Presenting Author

• Somer Bishop, University of California, San Francisco
  Non-Presenting Author

• Audrey Thurm, NIMH
  Non-Presenting Author