Introduction:
Endovascular therapy (EVT) and tissue plasminogen activator (tPA) are highly effective treatments but little is published on their impact on health care utilization and payments. We analyzed of data from the Michigan Value Collaborative (MVC) registry to compare 90-day episode payments for EVT and tPA treatments acute ischemic stroke patients.
Methods:
The MVC registry uses claims data from Michigan residents insured by Medicare fee-for-service and Blue Cross Blue Shield, to construct risk-adjusted and price-standardized payments for 90-day episodes of care. Total payments are categorized into index hospitalization, professional services, readmissions, and post-acute care. All ischemic stroke admissions between January 2014 and June 2019 were identified and then categorized using a combination of diagnosis and procedure codes into 3 mutually exclusive treatment groups: EVT (with and without tPA), IV or IA tPA only, and all other ischemic strokes (non EVT-tPA). We compared 90-day episode payments between the 3 groups using analysis of variance and Tukey’s pairwise comparisons.
Results:
We identified 47,235 acute ischemic stroke patients admitted to 83 Michigan hospitals, 1,775 (3.8%), 4,889 (10.3%), and 40,571 (85.9%) received EVT, tPA only, or neither, respectively. Statistically significant differences were observed across the 3 treatment groups in all payment categories. EVT treated patients had the highest 90-day mean payments in all categories, but particularly for readmissions and post-acute care (Table). tPA treated patients had significantly higher costs than the non EVT-tPA group for total payments, professional and hospitalization fees.
Conclusions:
EVT and tPA treatments significantly affect 90-day episode payments - primarily due to higher hospitalization and professional payments. Higher readmission and post-acute cost in EVT treated patients reflect higher readmission rates and use of in-patient rehabilitation.
Post–Acute Care Data for Predicting Readmission After Ischemic Stroke: A Nationwide Cohort Analysis Using the Minimum Data Set