3283 Comparative Effectiveness of Levetiracetam versus Phenobarbital for Infants Aged One Month to One Year, using US Medicaid Claims

Abstract

OBJECTIVES/SPECIFIC AIMS: Using national Medicaid claims, this retrospective cohort study aims to compare the outcomes of levetiracetam (LEV) versus phenobarbital (PHB) as initial monotherapy in infants with epilepsy aged one month to one year. We primarily analyzed health services outcomes, as follows: (1) Emergency Department (ED) visits: proportion of those with at least 1 ED visit, time to first ED visit, total number of ED visits, proportion leading to inpatient admission (2) Inpatient Admissions: proportion of those with at least 1 admission, time to first admission, total number of inpatient admissions, total length of stay (3) Treatment discontinuation: proportion of those who discontinued medication, time to discontinuation, proportion of those with prescription of second antiepileptic drug (AED) (4) Cost: total cost at 1 year from first prescription. METHODS/STUDY POPULATION: The project is a retrospective cohort study utilizing Medicaid claims from 2009-2012 from all states and DC. Inclusion criteria were: (1) Epilepsy diagnosis (ICD9 345.x) at age 1 month to 1 year (2) 1 year of continuous Medicaid enrollment Exclusion criteria were: (1) Neither LEV nor PHB monotherapy as initial treatment (2) Neither LEV nor PHB prescription within 45 days of epilepsy diagnosis (3) Less than 1 year of follow-up after first LEV or PHB prescription (4) Infantile Spasms diagnosis (5) Brain Surgery (6) Death within 1 year of follow-up Using R for statistical analysis, we analyzed outcomes including ED visits, inpatient admissions, treatment discontinuation, and total cost. In this abstract, we present our preliminary bivariate analysis. RESULTS/ANTICIPATED RESULTS: Demographics and Etiology: Compared to infants prescribed PHB (n = 1954), infants prescribed LEV (n = 1248) were older (median 6 months [IQR 3-9] vs 3[2-6]; p < 0.001), more likely to be white (44.2% vs 38.9%; p < 0.05) and not Hispanic (63.5% vs 58.9%; p<0.05). There were also important differences in epilepsy etiologies (p < 0.05). For example, infants prescribed LEV were more likely to have a diagnosis of tuberous sclerosis (1% vs 0.15%) or traumatic brain injury (12.8% vs 0.56%). Health Services Outcomes: After 1 year, infants prescribed LEV had more ED visits (2 [0-4] vs 1 [0-3]; p < 0.001) but shorter inpatient length of stay when admitted (3 days [2-5] vs 3 [2-6]; p < 0.001). They were less likely to discontinue the medication (46.6% vs 64.3%; p<0.001) but more likely to have a second AED prescription (53.3% vs 43.4%; p < 0.001). Other outcomes, including total cost, were similar. DISCUSSION/SIGNIFICANCE OF IMPACT: This preliminary analysis suggests that the healthcare trajectory of infants treated with LEV and PHB differ in complex ways. In ongoing work, we are conducting a multivariable comparative effectiveness analysis of LEV versus PHB using propensity score weighting to account for observable selection bias and multiple imputation to account for missing data.