Abstracts

Rationale: Various social factors prevent patients with epilepsy from achieving seizure freedom even after access to specialized epilepsy care. Early identification of patients who are more vulnerable to poor outcomes could guide targeted efforts to improve epilepsy care quality and equity. We investigate whether patients' social vulnerability index (SVI), a census tract-specific measure estimated by the Centers for Disease Control and Prevention indicating 15 different social determinant variables, can predict patients' change in seizure frequency after 6-9 months of epilepsy treatment.

Methods: In a prospective cohort registry (part of a larger quality improvement network) we collected address, standardized physician-recorded seizure type and frequency data of adult patients who were followed for 6-9 months after their baseline visit at two tertiary epilepsy centers from 1/2019 to 3/2021. Each patient's census tract was identified at baseline and mapped to an SVI value, ranging from 0 to 1 and increasing as the patient's vulnerability increased. There were 7 different categories of seizure frequency, from mild to severe. We quantified the change in seizure frequency categories from baseline to last visit within 6-9 months post-baseline visit. If a patient had multiple seizure types, we only kept the seizure type for which the change in seizure frequency was least favorable. We compared SVI distribution among patients whose outcomes improved, remained the same, or worsened after treatment. With a logistic regression model, we examined whether SVI values could predict the likelihood of maintenance or positive change in seizure frequency from baseline to 6-9 months follow-up. We stratified patients by their baseline seizure activity and evaluated each stratum separately.

Results: Our study cohort included 191 patients meeting eligibility criteria and whose characteristics are described in Table 1. Among those with low and high SVI (i.e., below 25th percentile and above 75th percentile, respectively) in the active seizure stratum, 83% of (30/36 and 29/35) patients improved or remained with the same seizure frequency after 6-9 months of follow-up. The SVI median (mean ± std) was: 0.25 (0.32 ± 0.23), 0.29 (0.36 ± 0.29) and 0.65 (0.55 ± 0.27) for patients whose seizure outcome improved, remained the same and worsened after treatment, respectively. These results indicate that patients whose seizure outcome worsened after treatment have higher SVI than those with more favorable outcomes. All patients with inactive seizure at baseline maintained their outcome after treatment. The coefficient of SVI in the logistic regression for the active seizure stratum was -2.17 (95% CI -4.70 - 0.23, p-value 0.08), suggesting that as SVI increases, the probability of maintaining or improving seizure outcome after treatment decreases.

Conclusions: A baseline measure of SVI could be informative in identifying opportunities for system’s improvement in the care of patients with epilepsy. More research is needed to determine SVI’s predictive power of seizure outcomes.

Funding: Please list any funding that was received in support of this abstract.: none.