Internal (Factorial) validity of the ANAM using a cohort of woman high-school soccer players

Abstract

Introduction: Computerized neuropsychological testing is a cornerstone of sport-related concussion assessment. Female soccer players are at an increased risk for concussion as well as exposures to repetitive head impacts from heading a soccer ball. Our primary aim was to examine factorial validity of the Automated Neuropsychological Assessment Metrics (ANAM) neuropsychological test battery in computing the multiple neurocognitive constructs it purports to measure in a large cohort of interscholastic female soccer players. Methods: Study participants included 218 interscholastic female soccer players (age = 17.0±0.7 year; mass = 55.5±6.8 kg; height = 164.7±6.6 cm) drawn from a large (850+) prospective database examining purposeful heading from four area high schools over a 10-year period. The ANAM-2001 measured neurocognitive performance. Three methods were used to identify integral constructs underlying the ANAM: (a) exploratory factor analysis (EFA), (b) first-order confirmatory factor analysis (CFA), and (c) hierarchical CFA. Results: Neuropsychological phenomena measured by the ANAM-2001 were best reproduced by a hierarchical CFA organization, composed of two lower level factors (Simple Reaction Time, Mental Efficiency) and a single, general composite. Although the ANAM was multidimensional, only the composite was found to possess sufficient construct dimensionality and reliability for clinical score interpretation. Findings failed to uphold suppositions that the ANAM measures seven distinct constructs, or that any of its seven tests provide unique information independent of other constructs, or the composite, to support individual interpretation. Conclusions: Outcomes infer the ANAM possesses factorial-validity evidence, but only scores from the composite appear to sufficiently internally valid, and reliable, to support applied use by practitioners.