Abstracts

Rationale:
Individuals with epilepsy who do not adequately respond to at least two different regimens of anti-seizure medications (ASMs) are considered to have DRE. VNS is a neurostimulator for use in patients with DRE that has been shown to control seizures and improve quality of life. We previously found VNS to be associated with significantly lower levels of hospitalizations and emergency department (ED) visits, and significantly lower costs of care (all vs. expected values), during the two year period following implantation among commercially (i.e., private) insured patients with DRE. However, the degree to which this reflects the experience of Medicare enrollees with DRE who receive VNS is not well understood.

Methods:
We used healthcare claims obtained from the US Center for Medicare and Medicaid Services CMS to identify Medicare fee-for-service enrollees with DRE who underwent VNS implantation between 2002 and 2019. The earliest date during this period on which implantation was noted was deemed the index date. Patients without an epilepsy diagnosis on the index date were excluded, as were those: (1) without ≥ one ASM claims in the year prior to implantation; and (2) not continuously enrolled during the two year period prior to index (“pre-index”). Data on use and cost of healthcare resources and pharmacotherapy during pre-index was used to develop multivariate regression models to estimate expected values for the 24-month post-index period (“follow-up”). Observed outcomes during follow-up were contrasted to those expected, and were assessed alternatively using an all-cause and epilepsy-related basis; the latter was defined as all medical care claims with diagnoses of epilepsy and all ASM dispenses.

Results:
A total of 3,155 patients met all selection criteria. After twelve months post-index, observed hospitalizations were 6.1 per 100 person-months (PMs) lower than expected values, and emergency department (ED) visits were 8.7 per 100 PMs lower; corresponding values at 24 months post-implantation were 9.3 per 100 PMs lower and 15.7 per 100 PMs lower, respectively (p< 0.0001 for all comparisons). Twenty-four months after VNS implantation, observed all-cause monthly costs were $2,113 less than expected values; observed epilepsy related monthly costs were $2,355 less (p< 0.0001 for both comparisons) (Figure 1). Differences between cumulative observed versus expected costs ceased to be statistically significant at month 19.