Authors :
Presenting Author: Andrea Miele, PhD – Children’s Hospital Colorado
Morgan Jolliffe, MA – Children's Hospital Colorado
Megan Stringfellow, BS – Children's Hospital Colorado
Dana Bennink, DPT – Children's Hospital Colorado
Andrea Gerk, OTD – Children's Hospital Colorado
Ann Reynolds, MD – Children's Hospital Colorado
Kaitlyn Kennedy, CRA – Children's Hospital Colorado
Tristen Dinkel, BSN – Children's Hospital Colorado
Katie Angione, MS, CGC – Children's Hospital Colorado
Margarita Saenz, MD – Children's Hospital Colorado
Hsiao-Tuan Chao, MD, PhD – Baylor College of Medicine
Zachary Grinspan, MD, MS – Weill Cornell Medicine
James Goss, PhD – STXBP1 Foundation
Charlene Son Rigby, BA, MBA – STXBP1 Foundation
Scott Demarest, MD – Children's Hospital Colorado
Rationale: The STXBP1-Related Disorders (STXBP1-RD) are a developmental epileptic encephalopathy characterized by developmental delay, intellectual disability, seizures
, and behavioral difficulties. Given the high developmental burden, assessment tool selection is limited. Here, we report growth scale value (GSV) data to highlight feasibility and psychometric properties from the Vineland Adaptive Behavior Scale – Third Edition (Vineland-3) and the Bayley Scales of Infant Development, Fourth Edition (Bayley-4). GSV scores have been used since the 1970s and represent “absolute performance.” Each GSV scale is unique to the specific test; GSV scores are created from raw scores that have been converted to an equal-interval scale following a Rasch analysis of the test
1.
1) Daniel, M. H., & Vannier, L-C. (2022).
Growth scale value (GSV): Theory, development, and applications [GSV Technical Report 1]. NCS Pearson.
Methods: All individuals completed comprehensive, clinical neuropsychological evaluations at Children’s Hospital Colorado by the first author. Participants consented for their data to be used in research on a protocol approved by COMIRB, and data was collected as part of the STXBP1 STARR Natural History Study. Distributions of GSV scores were compiled for all available domains on the Vineland-3 and Bayley-4.
Results: From the Vineland-3, data from a total of 28 individuals with STXBP1-RD (mean age=8.6 years, SD=5.69), range from 6-months to 22.8 years) were collected. Descriptive statistics revealed similar standard deviations for several domains (see Table 1). Boxplots revealed GSV score distributions were broader than the raw score distribution for all domains except for play and leisure (see Figures 1 and 2). While several “minimum” scores were obtained (i.e., raw score of 0, 1, or 2), these scores equated to a minimum GSV scores of 10, 21/22, and 28. Of note, these lowest raw scores were not obtained by the youngest members of the cohort. GSV scores are not available for certain domains based on age.
On the Bayley-4, data from a total of 17 individuals with STXBP1-RD (mean age=7.37 years, SD= 5.45) were collected. Sample sizes are different as this measure was not developmentally appropriate for certain individuals. Descriptive statistics again revealed similar standard deviations between several domains (see Table 2). Boxplots revealed GSV score distributions were larger compared to the raw scores for the receptive and expressive language and fine motor domains, but smaller for the cognitive and gross motor domains (see Figures 3 and 4).
Conclusions: This study reveals GSV scores have a wider distribution for 10 of 11 domains on the Vineland-3, and 3 of 5 domains on the Bayley-4. The reduced ranges on the Bayley-4 may be explained by smaller sample size resulting in a truncated range of cognitive and gross motor abilities. Overall, GSV scores seem well poised to meaningfully detect smaller units of change compared to raw scores, thus making the Vineland-3 and Bayley-4 strong contenders as outcome measures for clinical trials. Future research will investigate GSV changes across time in the same individuals.
Funding: This work was supported by the STXBP1 Foundation.