| Poster | Poster Session 09 Program Schedule
02/16/2024
03:30 pm - 04:45 pm
Room: Majestic Complex (Posters 61-120)
Poster Session 09: Epilepsy | Oncology | MS | Infectious Disease
Final Abstract #72
Predictors of Reading Comprehension Among Children with Epilepsy
Kelly Macdonald, Children's National Hospital, Washington, United States
Chloe Hooker, Children's National Hospital, Washington, United States
Leigh Sepeta, Children's National Hospital, Washington, United States
William Gaillard, Children's National Hospital, Washington, United States
Madison Berl, Children's National Hospital, Washington, United States
Category: Epilepsy/Seizures
Keyword 1: academic skills
Keyword 2: executive functions
Keyword 3: epilepsy / seizure disorders
Objective:
Children with epilepsy (CWE) are at higher risk of learning difficulties than their neurotypical peers; however, academic difficulties are understudied in CWE relative to other cognitive impairments (e.g., memory, attention). Struggles with reading comprehension (RC) have the potential to negatively impact student performance across subjects given that students are gradually expected to learn academic content through reading. This study integrates theory from the well-established reading disabilities literature (Gough & Tunmer, 1986) with prior work in epilepsy (Fastenau et al., 2003) to test a predictive model of RC that integrates broad language skills (vocabulary), pre-literacy skills (rapid naming, RAN), sight word reading, working memory (WM), processing speed (PS), and epilepsy characteristics (age of onset, number of anti-seizure medications, seizure type, and seizure frequency). We hypothesized that language, pre-literacy, seizure characteristics, and executive functioning skills in particular, would predict RC performance.
Participants and Methods:
Participants (n = 150) were recruited through multiple clinical research protocols with different primary aims but included common measures. We included CWE who completed a measure of RC and had IQ > 70 to rule out global impairment as the reason for their reading problems. Bivariate relationships between predictors and RC were assessed with correlations and one-way ANOVAs. The contributions of language and literacy skills (Step 1), WM and PS (Step 2), and seizure characteristics (Step 3) were tested using stepwise multiple regression.
Results:
Mean RC performance ranged from 55 to 130 (M = 90.3, SD = 14.9). Among the predictors, vocabulary was the strongest skill (M = 95.9), while RAN was the lowest (M = 87.9); significantly worse than vocabulary. RC was significantly correlated with sight word reading (r = .63), vocabulary (r = .61), RAN (r = .27), WM performance (r = .44), WM ratings (r = -.27), and PS (r = .31). RC performance differed by seizure type (generalized<focal, F = 7.75, p = .006). In Step 1, vocabulary (β = .36, p < .001) and sight word reading (β = .47, p < .001), but not RAN (β = .10, p = .179) were predictors accounting for 51% of the variance. In Step 2, WM performance (β = .17, p = .015), contributed another 6% of explained variance. In Step 3, seizure type did not contribute uniquely beyond language, literacy, and working memory (β = -.11, p = .095).
Conclusions:
Consistent with classic theoretical models of RC, fundamental skills (vocabulary and sight word reading) account for variability in RC in CWE. Beyond these skills, WM also uniquely explains variability in RC in CWE. The relationship between WM and reading was higher in our sample (r = .44) than found in the general population (meta r = .29; Peng et al., 2018). It is possible that RAN is not a significant predictor in CWE as slow processing is common, often due to anti-seizure medications. Even though RAN likely contributes to RC difficulties, it does not differentiate performance on RC within CWE. Our results inform how evidence-based interventions for RC in CWE might be tailored for their unique profile.
|
|---|
