Minimal Clinically Important Difference on the Social Responsiveness Scale in Autistic Children Aged 4 to 10 Years

The Social Responsiveness Scale (SRS) is widely used in autism clinical trials to assess treatment effects on core symptoms. However, its minimal clinically important difference (MCID) remains undefined, limiting meaningful interpretation of treatment efficacy. This study pooled data from four clinical trials (542 observations from 319 autistic children, aged 4-10 years) to estimate an anchor-based MCID. Using the Clinical Global Impression-Improvement (CGI-I) scale as an anchor, we employed linear mixed-effects models to estimate MCID, defined as the between-group difference between participants reporting minimal improvement versus no change. We tested interactions among baseline SRS score and clinical characteristics. To complement the anchor-based approach, we also incorporated distribution-based MCID estimates using data from 1,405 autistic children (aged 4-10 years) drawn from an autism registry. We found that MCID depended on baseline SRS score, described by the formula MCID = 10.33 - 0.18 × (baseline SRS score). This translates to MCIDs of 5.87 and 9.47 for baseline SRS scores of 90 and 110, respectively. No significant interactions were found with sex, age, or co-occurring intellectual disability/developmental delay. Distribution-based estimates aligned closely with anchor-based findings. These estimates offer a crucial benchmark for interpreting treatment outcomes in autism trials.Lay AbstractThe Social Responsiveness Scale (SRS) is widely used to assess core symptoms in autistic individuals; however, its minimal clinically important difference remains undefined, limiting clinicians' ability to distinguish meaningful changes from random variations. By analyzing data from over 300 autistic children aged 4 to 10 years across multiple clinical trials and over 1,000 from an autism registry, we showed that the threshold for meaningful clinical improvement varies systematically with baseline SRS score. These findings provide critical benchmarks for interpreting treatment outcomes in clinical practice and autism research.