OBJECTIVES: Comprehensive stroke centers (CSCs) are facilities that can diagnose and treat stroke patients with intensive and advanced care. CSC capabilities can be assessed using CSC scores that are calculated based on a hospital’s fulfillment of 25 recommended items. This study aimed to investigate the relationships between CSC scores and medical and long-term care (LTC) expenditures in ischemic stroke patients after hospital discharge.

METHODS: Medical and LTC expenditures were compared between two groups of ischemic stroke patients using a LIFE Study database from April 4, 2017 to March 31, 2020. One group comprised patients admitted to hospitals with high CSC scores, and the other group comprised patients admitted to hospitals with low CSC scores. Observation of each patient began from the month after hospital discharge, and continued for up to 12 months. Covariates were adjusted using an inverse probability weighting method. Monthly medical expenditures were analyzed using a weighted Gamma regression model, and monthly LTC expenditures were analyzed using a weighted Gamma regression model combined with a hurdle model.

RESULTS: A total of 372 ischemic stroke patients were admitted to hospitals with recorded CSC scores. The high-CSC group comprised 204 patients and the low-CSC group comprised 168 patients. For the first and second observation months, the high-CSC/low-CSC ratios of the weighted average medical expenditures were significant (1.36, P=0.02 and 1.49, P=0.01, respectively). In contrast, the high-CSC/low-CSC odds ratios of the proportions of patients incurring LTC expenditures were first significant in the sixth observation month (1.45, P=0.02) and in approximately half of the following months.

CONCLUSIONS: Patients in high-CSC hospitals tended to incur higher medical expenditures soon after discharge, but incurred higher LTC expenditures at a later period. Further investigations into the underlying causes of these differences may help to facilitate measures to reduce medical and LTC expenditures.