Factors Related to False-Positive Diagnoses by Community Sources

Background: As genetic research continues to attempt to amass large numbers of children and families for analysis, more and more children that have received diagnoses from community sources are being recruited for potential participation.  Phenotypic assessment of these children utilizing gold standard assessment tools such as the Autism Diagnostic Observation Schedule (ADOS), Autism Diagnostic Interview-Revised (ADI-R), as well as ultimate clinical diagnosis (CBE), commonly find that a large percentage of children referred by community sources do not, in fact, meet criteria for an ASD.  

Objectives: In the current report we analyzed factors commonly associated with community referrals of false-positive ASD cases.

Methods: The family cases analyzed were participants in the Vanderbilt site of the Simons Simplex Collection. Each family had one child with either a previous, community-based diagnosis of ASD or an elevated score on the SCQ (i.e., >= 9). The family then completed a battery of questionnaires, a thorough medical history interview, and a series of cognitive and diagnostic assessments, including an ADOS and an ADI-R. Each case was then reviewed by at least one licensed clinician and clinical best estimate diagnosis was issued.  Files of cases not meeting diagnostic criteria for ASD were reviewed and variability associated with 1) administered measures and/or 2) important medical and environmental factors was explored.

Results: Of the 85 families completing our assessment protocol, 12 recruited children did not receive an ASD diagnosis (14.11% of the sample). The age range for these children was 4 to 12 years. 41.6% of these subjects were female. 33.3% were home-schooled; 66.7% had conflicting cursory diagnoses by medical professionals using measures other than the ADOS/ADI-R; 50% met criteria on the ADI-R and other parent report measures, but not on face-to-face measures; 25% met criteria on both the ADI-R and ADOS but did not warrant a CBE of ASD; and 16.6% expressed significant external pressures (i.e. state financial support or internet self-diagnosis) to secure an ASD diagnosis.

Conclusions: Several factors were prominent in children diagnosed with ASD in the community or meeting thresholds solely on a parent referral checklist who ultimately did not meet diagnostic criteria of ASD using gold-standard assessment practices.  Specifically, families with children with limited exposure to traditional school settings, receiving a diagnosis of ASD from an M.D., and children with parents who generated self-diagnoses were common in our sample.  These findings highlight the importance of standardized assessment tools in combination with clinical expertise in order to control for potential false-positive cases in large scale recruitment endeavors such as genetic repositories. The trends reported here also may have implications for recruitment and screening methods chosen by researchers attempting to access participants with ASD from broader community settings.