Discriminating Reactive Attachment Disorder From Autism Spectrum Disorders: Key Symptoms and Clinical Characteristics

Background:

Differentiation between children with Autism Spectrum Disorder (ASD) and Reactive Attachment Disorders (RAD) is problematic. Both have similar social and communication difficulties. There is a need for scientific studies and guidelines to assist clinical decision-making.

Objectives:

The main objectives of the study were to answer the following questions: i) Do the social and communication difficulties, associated with ASD, affect the quality of parent-child relationships? ii) Are there qualitative differences between the relationship and attachment difficulties experienced by children with ASD and RAD?  

Methods:

We used the Relationship Problem Questionnaire (RPQ) to measure parent-reports of symptoms characteristic of RAD. ~The RPQ is a 10-item parent-rated questionnaire, with two subscales, measuring socially inhibited and disinhibited behaviours respectively. A cross sectional design was used to examine the discriminative value of the RPQ in a sample of 224 participants in age range 3 to 16 years (clinical diagnoses RAD (20%), ASD (35%) and population controls (45%). RAD diagnosis was made by a modified Child and Adolescent Psychiatric Assessment interview (CAPA), and all clinical participants had standardized ASD interviews and observations (3Di/ADOS), as well as extensive cognitive and other behavioural and observational measures. ASD/RAD families were recruited from specialist clinics in London and Glasgow, and controls were a representative sample of Scottish families.

Results:

There were highly significant differences between all three groups on our measure of relationship and attachment difficulties, the RPQ (p<0.001).  Children with ASD had more attachment relationship problems than those in the general population, but those with RAD obtained more abnormal scores than either of the other groups (p<0.001).  A logistic regression was used to identify items that distinguished children who had been clinically diagnosed with RAD from ASD. The optimal model found a limited set of items correctly classified over 80% of cases. RAD was characterized by indiscriminate and insincere affection, self-directed aggression, and frozen fear. In contrast, ASD was characterized by the use of personal, intrusive questioning. RAD inhibition did not distinguish these clinical groups, but disinhibition was significantly greater in the RAD sample. Neither age nor IQ had any significant impact on RPQ scores, in any of the 3 samples. 34% with a clinical diagnosis of ASD scored above the RPQ threshold for RAD, of whom 26% (9% of total ASD) had callous and unemotional (CU) traits. No CU traits were reported in ASD children who scored below the RAD threshold. On basis of ADI-R algorithm scores, the ASD group scored significantly higher than either RAD or general population, but 46% RAD scored above threshold for reciprocal social interaction problems; 63% had communication problems; 20% had repetitive and stereotyped behaviours.

Conclusions:

Whilst RAD and ASD can look superficially similar to the clinician, and share autistic symptoms, there are characteristics of both conditions that assist in their discrimination. The RPQ provides a useful means of distinguishing between RAD from ASD, and just five items contribute to the greater part of the variance.