Stability of the ADI-R
Stability of the Autism Diagnostic Interview - Revised from pre-school to elementary school age
in children with autism spectrum disorders
The issue of diagnostic stability in the field of autism spectrum disorders (ASD) has become an
increasing focus of research and is of particular relevance for studies of prevalence and intervention.
Several studies confirm that pre-school diagnoses of autism made by expert clinicians remain relatively
stable 2-22 years after initial diagnosis (Billstedt, Gillberg & Gillberg, 2005; Cox, et al., 1999; Eaves &
Ho, 2004; McGovern & Sigman, 2005; Moore & Goodson, 2003; Lord, 1995; Stone et al., 1999).
Other studies have explored the stability of autism/ASD diagnoses based on specific diagnostic
instruments. Although the ADI-R is among the best validated of diagnostic interview measures, follow-
up studies indicate that ADI-R based diagnoses made early in life may be less stable than clinical
judgement (see Charman & Baird, 2002; Chawarski, Klin, Paul & Volkmar, 2007). For example,
follow-up studies by Lord and colleagues (2006) reported that clinical judgement at age 2 was a more
accurate predictor of diagnosis at age 9 than an ADI-R at 2 years. Lord et al. (2006) found the ADI-R
to be initially more inclusive than clinical judgement for both autism and ASD criteria based on Risi et
al.’s ASD criteria (2006). Lord et al. (2006) noted that the use of standardised assessments such as the
ADI-R alongside clinical judgement was likely to result in more stable diagnosis than either ADI-R or
clinical judgement alone. Moreover, Charman et al. (2005) in a study of 26 children monitored the
numbers meeting ADI-R criteria for autism at 4 time points from 2 to 7 years. Six participants changed