Abstracts

Rationale: Infantile spasms (West Syndrome) is an early life epilepsy characterized by epileptic spasms that occur in clusters. Timely and effective treatment can lead to epilepsy remission and improved neurodevelopmental outcomes. Guidelines, quality measures, and consensus statements consistently recommended use of ACTH, vigabatrin, or oral steroids for first line treatment. The relationship between medication selection and health services outcomes (cost, ED visits, and inpatient admissions) is uncertain. These outcomes are important to understand given the difference in cost for the three medications. A course of oral steroids may cost less than 100 USD, while a course of ACTH may cost more than 100,000 USD. Methods: We performed a retrospective cohort study to examine the association of medication selection on health services outcomes for children with infantile spasms using Medicaid claims data for all 50 US states from 2009 – 2012.  We defined a case of infantile spasms as any child who had at least one claim with an ICD-9 diagnosis of 345.6x (infantile spasms) between 2 months and 9 months of age.  We considered ICD-9 codes up to 90 days after the diagnosis of 345.6x to determine etiology and comorbidity burden. We examined cost, ED visits, and inpatient visits at 1 and 2 years after the first diagnosis.  We defined cost as any money paid by Medicaid for health services plus any money paid by the family (as indicated in the claims data).  We examined four different treatments groups: ACTH, vigabatrin, oral steroids, and topiramate.   Results: We identified 874 patients from 47 states with a diagnosis of infantile spasms (345.6x) between age 2 and 9 months.  126 were prescribed ACTH, 70 prednisolone, 24 vigabatrin, and 79 topiramate. There was evidence of selection bias: prednisolone was more likely given to girls, vigabatrin more likely given to infants with tuberous sclerosis. Hormonal therapy (ACTH and prednisolone) was more likely given to individuals without one of seven common etiologies, and less likely given to infants with a history of neonatal ICU care. (Table 1) Two years of follow up were available for only a subset of patients: 99 in the ACTH group, 51 prednisolone, 59 topiramate, and 17 vigabatrin.  The 1^st year non-pharmacy costs ranged from a median of 14k [IQR 12k – 24k] USD in the vigabatrin group to 22k [13k – 43k] USD in the ACTH group, though these differences were not statistically significant.  The 1^st year pharmacy costs were significantly higher in the ACTH group (95k [71k – 124k]) compared to the other groups (prednisolone 3 [1 – 17], topiramate 6 [2 – 45], vigabatrin 27 [23 – 37]; p < 0.001).  [PA1] There were also differences in use of health services which were complex: children with ACTH and vigabatrin were somewhat less likely to use the ED in years 1 and 2 compared to children with prednisolone or topiramate. Children with ACTH were nearly all admitted in the first year, but were the least likely to be admitted in the second year.   Conclusions: Analysis of Medicaid claims data highlights the known clinical practice of preferential selection of medications for infantile spasms in different populations - hormonal therapy for unknown etiologies and vigabatrin for tuberous sclerosis. There is a large cost difference in the first year of care for children with IS treated with ACTH compared with other common selections, on the order of 75k - 100k USD, which is attributable to the cost of the medication. Use of health services also differ by treatment selection, but interpretation of these unadjusted findings is complex. In ongoing work, we are conducting additional analyses to evaluate cost and health services outcomes with adjustments for observable selection bias. Funding: Not applicable