A-96 Stability of ImPACT's Latent Structure from Baseline to Post Concussion Assessment. (30th August 2021)

• Record Type: Journal Article
• Authors: Goodwin, Grace J; Maietta, Julia E; Ahmed, Anthony O; Hopkins, Nia A; Moore, Sara A; Rodrigues, Jessica; Pascual, Melanie S; Kuwabara, Hana C; Kinsora, Thomas F; Ross, Staci R; Allen, Daniel N
• Is Part Of: Archives of clinical neuropsychology. Volume 36:Number 6(2021)
• Page Start: 1143
• Page End: 1144
• Publication Date: 2021-08-30
• Subjects: Clinical neuropsychology -- Periodicals; 616.805
• DOI: 10.1093/arclin/acab062.114
• Languages: English
• ISSNs: 0887-6177

Abstract

Abstract: Objective: ImPACT is commonly used for sport-concussion management. Baseline and post-concussion tests serve as within-athlete comparisons for return-to-play decision-making. Previous literature has questioned whether ImPACT's five composites accurately represent the internal structure of its cognitive scores. A recent alternative four-factor structure has strong confirmatory evidence for baseline scores (Maietta et al., doi:10.1037/pas0001014 ). The present study examined the stability of these constructs post-concussion. Method: The current study utilized a case-matched design (age, sex, sport category) to select a sample of 3560 high school athletes' baseline (n = 1780) and post-concussion (n = 1780) assessments. Multi-group CFA of first-order, hierarchical, and bifactor models was conducted to assess measurement invariance (configural, metric, scalar, and residual invariance) between baseline and post-concussion samples. Change in comparative fit indices was interpreted as the primary indicator of model invariance. Results: ImPACT's five composite structure, as well as the hierarchical and bifactor models, exhibited inadequate fit to the baseline and post-concussion data. The four-factor model demonstrated superior fit in the baseline sample and good fit in the post-concussion sample. The four-factor structure demonstrated invariance across injury status (baseline to post-concussion). Conclusion: Given that ImPACT's scores are used for return-to-play decision-making, it is important that they are psychometrically sound. Recent literature suggests that ImPACT's five composites are not an adequate representation of the cognitive constructs. Findings support validity of the four-factor structure despite injury status, suggesting these cognitive constructs are assessable at both pre- and post-concussion. Additional research is needed to determine implications of these findings for tracking cognitive change following sport-related concussion and making return-to-play decisions.