The contribution of neurocognitive functions to academic and psychological outcomes in pediatric cancer: A latent profile analysis

• Published in: Child neuropsychology 2020-10, Vol.26 (7), p.881-899
• Main Authors: Karlson, Cynthia W., Sarver, Dustin E., Raiker, Joseph S., Espil, Flint M., Cox, Amanda D., Elkin, T. David, Annett, Robert D.
• Format: Article
• Language: English
• Subjects: Adolescent; Child; Child, Preschool; Childhood cancer; Cognition Disorders - etiology; Cognition Disorders - psychology; cognitive; Emotions; Executive Function - physiology; Female; Humans; Latent Class Analysis; Male; Neoplasms - complications; neuropsychological; Neuropsychological Tests; Quality of Life - psychology; risk factors; survivors; Survivors - psychology
• ISSN: 0929-7049; 1744-4136
• DOI: 10.1080/09297049.2020.1734553

This study aimed to identify the impact of neurocognitive functioning on academic and psychological domains using a novel person-centered latent profile analysis approach. We further examined the contribution of identified risk factors (e.g., age at diagnosis, treatment) on latent class membership. 101 pediatric oncology patients and survivors (M age = 11.2, 35.6% female; 47.5% African American; M time since diagnosis = 3.4 years) completed neuropsychological evaluations at a university medical center between February 2004 and June 2017. Neurocognitive, academic, and emotional-behavioral functioning were examined using validated measures. Discreet, homogenous neurocognitive subgroups (latent classes) were identified using latent profile analysis. Demographic and medical factors were evaluated as predictors of latent class. A 3-class model indicated excellent class separation (range: .00-.04) and homogeneity (range: .94-.99). Classes were distinguished by differential cognitive patterns. Class 2 (52%) and Class 3 (25%) displayed overall normative functioning; however, Class 3 displayed significantly poorer attention than the other two classes. Class 1 (23%) demonstrated Borderline neurocognitive, low average academic, and poorer emotional-behavioral and inhibition/executive control functioning. Class membership was predicted by race and whole brain radiation dose. Latent profile analysis identified discrete groups in neurocognitive functioning in this heterogeneous pediatric cancer population. Class membership was predicted by race, whole brain radiation dose, and referral source. Other medical variables (e.g., diagnosis, age at diagnosis) were not significant predictors of neurocognitive function in our sample.