Premium
Development and validation of a streamlined autism case confirmation approach for use in epidemiologic risk factor research in prospective cohorts
Author(s) -
Newschaffer Craig J.,
Schriver Emily,
Berrigan Lindsay,
Landa Rebecca,
Stone Wendy L.,
Bishop Somer,
Burkom Diane,
Golden Anne,
Ibanez Lisa,
Kuo Alice,
Lakes Kimberly D.,
Messinger Daniel S.,
Paterson Sarah,
Warren Zachary E.
Publication year - 2017
Publication title -
autism research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.656
H-Index - 66
eISSN - 1939-3806
pISSN - 1939-3792
DOI - 10.1002/aur.1659
Subject(s) - autism , context (archaeology) , autism spectrum disorder , observational study , test validity , psychology , overfitting , predictive validity , convergent validity , medicine , psychometrics , clinical psychology , developmental psychology , machine learning , paleontology , artificial neural network , computer science , internal consistency , biology
The cost associated with incorporating standardized observational assessments and diagnostic interviews in large‐scale epidemiologic studies of autism spectrum disorders (ASD) risk factors can be substantial. Streamlined approaches for confirming ASD case status would benefit these studies. We conducted a multi‐site, cross‐sectional criterion validity study in a convenience sample of 382 three‐year olds scheduled for neurodevelopmental evaluation. ASD case classification as determined by three novel assessment instruments (the Early Video‐guided Autism Screener E‐VAS; the Autism Symptom Interview, ASI; the Screening Tool for Autism in Toddlers Expanded, STAT‐E) each designed to be administered in less than 30 minutes by lay staff, was compared to ADOS scores and DSM‐based diagnostic assessment from a qualified clinician. Sensitivity and specificity of each instrument alone and in combination were estimated. Alternative cutpoints were identified under different criteria and two‐stage cross validation was used to avoid overfitting. Findings were interpreted in the context of a large, prospective pregnancy cohort study utilizing a two‐stage approach to case identification. Under initial cutpoints, sensitivity ranged from 0.63 to 0.92 and specificity from 0.35 to 0.70. Cutpoints giving equal weight to sensitivity and specificity resulted in sensitivity estimates ranging from 0.45 to 0.83 and specificity ranging from 0.49 to 0.86. Several strategies were well‐suited for application as a second‐stage case‐confirmation. These included the STAT‐E alone and the parallel administration of both the E‐VAS and the ASI. Use of more streamlined methods of case‐confirmation in large‐scale prospective cohort epidemiologic investigations of ASD risk factors appears feasible. Autism Res 2017, 10: 485–501 . © 2016 International Society for Autism Research, Wiley Periodicals, Inc.