Validity for Clinical Global Impression-Improvement in Minimally Verbal Children with Autism Spectrum Disorder

Poster Presentation
Thursday, May 2, 2019: 11:30 AM-1:30 PM
Room: 710 (Palais des congres de Montreal)
A. Holbrook1, C. K. Toolan2, A. J. Schlink3, N. Brady4, D. Almirall5 and C. Kasari2, (1)University of California Los Angeles, Los Angeles, CA, (2)University of California, Los Angeles, Los Angeles, CA, (3)UCLA Semel Institute for Neuroscience and Human Behavior, Los Angeles, CA, (4)University of Kansas, Lawrence, KS, (5)University of Michigan, Ann Arbor, MI

There are few valid and reliable ASD-specific assessments available for minimally verbal (MV) children (Kasari et al., 2013). Even with appropriate tests, MV children may not fully display their communicative abilities in unfamiliar contexts, complicating assessment of change over time. Observational and naturalistic measures may be informative, but coding and scoring are often time-consuming. One promising approach is the Clinical Global Impression-Improvement (CGI-I) scale (Guy, 1976), a brief, well-established tool for assessing global improvement in specific constructs of psychiatric disorders (Busner & Targum, 2007). It typically uses an observational approach, considering information from naturalistic contexts, and is completed by blinded raters. However, for quickly assessing intervention progress in real-time (e.g., in a SMART study), interventionists can also apply the CGI-I. While the CGI-I has been validated for measuring symptomatology of complex disorders (e.g., schizophrenia), it has not yet been validated as a measure of social communication change in MV children with ASD.


Examine validity of the CGI-I as a measure of social communication change in MV children with ASD.


Participants included 54 MV (<20 words) school-aged children with ASD (Mage=6.05 years, SD=1.18). Children received intervention targeting social communication skills, focused on improving joint attention (JA).

Children received ratings of their severity of social communication prior to intervention and their improvement in social communication after 6 weeks of intervention. Primary interventionists rated CGI-I scores for each participant. CGI-Is were also randomly rated by blinded researchers for reliability. Children with CGI-I scores of 1 (“very much improved”) and 2 (“much improved”) were considered “fast responders;” children with scores of 3 (“minimally improved”) or higher were considered “slow responders.”

The early social communication scales (ESCS) was administered at entry and after 6 weeks of intervention. ESCS was coded by blinded coders using the Communication Complexity Scale (CCS; Brady et al., 2012), which examines communication in individuals who primarily use pre-linguistic or non-symbolic communication. The CCS is a valid and sensitive measure of social communication change (Brady et al., under review). From this we calculated ΔOptimal score, ΔBehavior Regulation (BR) score, and ΔJA score. Mann-Whitney U tests were conducted to determine differences between fast and slow responders.


37% of children were rated as “fast responders” on the CGI-I. There was no difference between fast and slow responders in distributions of ΔOptimal or ΔBR. However, the distribution of ΔJA was significantly different across fast and slow responders (p=.039), such that fast responders improved more in their ΔJA scores.


This study examines the validity of the CGI-I as a measure of social communication change after a short intervention period in a SMART study. The CGI-I significantly corresponded to changes in JA, indicating that the CGI-I validly measures change in social communication during treatment for MV children with ASD. These findings are practically significant. The CGI-I is relatively easy to administer compared to other standardized assessments. It does not require any additional materials and can be assessed in real-time. It is clinician-based, allowing children to demonstrate small but clinically meaningful changes in social communication skills.