Final DSM-5 under-identifies mild Autism Spectrum Disorder: Agreement between the DSM-5, CARS, CASD, and clinical diagnoses

Abstract

Agreement between the final DSM-5 ASD criteria, Childhood Autism Rating Scale (CARS), and Checklist for Autism Spectrum Disorder (CASD) was assessed in 143 children with ASD and other disorders (e.g., ADHD, intellectual disability, and oppositional defiant disorder). Diagnostic agreement between the CARS and CASD was high (94%), but their agreement with the DSM-5 was lower (84% and 88%). Agreement between the DSM-5 and both the CARS and CASD increased to 94% and diagnostic accuracy increased from 92% to 96% when one less DSM-5 social communication and interaction symptom was required for a diagnosis. Children with ASD not meeting DSM-5 criteria most often did not have criterion A2 (deficits in nonverbal social communication). Total scores on the DSM-5, CASD, and CARS were far higher for children with mild ASD (formerly PDDNOS) than no ASD, indicating that these children are clearly on the autism spectrum and are quite different from children with other disorders. However, only one child with mild ASD was identified by the DSM-5. This study and 11 others show that the DSM-5 under-identifies children with ASD, particularly children at the mild end of the spectrum. This can be rectified by requiring one less social communication and interaction symptom for a diagnosis.

Introduction

According to 11 studies, the proposed 2012 DSM-5 Autism Spectrum Disorder (ASD) criteria under-identifies children with ASD, particularly those with mild ASD or pervasive developmental disorder not otherwise specified (Barton et al., 2013, Frazier et al., 2012, Gibbs et al., 2012, Huerta et al., 2012, Matson et al., 2012a, Matson et al., 2012b, Matson et al., 2012c, Mayes et al., 2013, McPartland et al., 2012, Worley and Matson, 2012, Young and Rodi, 2013). Together, these studies included individuals of all ages (toddlers through adults), levels of functioning (severe intellectual disability to gifted), and degree of ASD (PDDNOS to severe autism). An important finding was that individuals who had a clinical diagnosis of ASD but did not meet the proposed DSM-5 criteria had significant autism symptoms on autism severity measures and when compared with clinical and normal controls (Matson et al., 2012a, Matson et al., 2012b, Matson et al., 2012c, Mayes et al., 2013, Worley and Matson, 2012). The majority consensus was that the proposed DSM-5 ASD criteria should be relaxed. Three studies (Frazier et al., 2012, Matson et al., 2012b, Mayes et al., 2013) showed that by simply requiring one less symptom for a DSM-5 diagnosis, fewer individuals with ASD were missed (increasing sensitivity by 12%, 14% and 16%, respectively), specificity was minimally decreased (by 2%, 0%, and 3%, respectively), and diagnostic agreement between the proposed DSM-5 and the DSM-IV improved.

The final DSM-5 ASD criteria (American Psychiatric Association, 2013, pp. 50–54) differ from the initial 2012 proposed criteria, potentially increasing the likelihood of an ASD diagnosis, as well as agreement with established autism tests. Major changes are: (1) specifically stating at the beginning of the diagnostic criteria that the symptoms are manifested “currently or by history” so that symptoms are scored based on life-time occurrence, (2) stating that examples in the criteria “are illustrative, not exhaustive” and providing additional examples not included in the initial criteria, (3) replacing “and” with “or” in a few places so that all symptoms are not required (e.g., as in “difficulties in sharing imaginative play or in making friends”), (4) relaxing the wording of some symptoms (e.g., “total lack of initiation of social interaction” is replaced with “failure to initiate or respond to social interactions”), (5) emphasizing the importance of peer relationships (replacing “absence of interest in people” with “absence of interest in peers”), (6) adding to Criterion C that symptoms “may be masked by learned strategies later in life,” and (7) providing several additional examples of symptoms. These include: (A1) “deficits in…the ability to engage with others and share thoughts and feelings” and “difficulties processing and responding to complex social cues (e.g., when and how to join a conversation, what not to say),” (A2) “absent, reduced, or atypical use of …facial expressions, body orientation, or speech intonation,” (A3) “inappropriate approaches that seem aggressive or disruptive,” “insistence on playing by very fixed rules,” and “preference for…interacting with much younger or older people,” (B1) “use of ‘you’ when referring to self,” (B2) “difficulties with transitions,” “rigid thinking patterns,” “insistence on adherence to rules,” “repetitive questioning,” and “pacing a perimeter,” and (B4) “excessive food restrictions.”