Objective:

We hypothesized that LOS mediates the association between insurance status and hospital costs for acute stroke patients. 

Background:

Length of stay (LOS) is a key quality indicator for stroke hospitalization, impacting costs, resource allocation, and transitions of care.  Insurance status often affects stroke LOS through delayed authorization for rehabilitation services.

Design/Methods:

We queried data from a single comprehensive stroke center for all hospital admissions of adults (≥ 18 years) with any stroke (ICD10 I60-I69) from 2017 to 2021. We performed a mediation analysis using linear regression models to understand how much LOS carries the effect of various insurance plans on total acute stroke care cost using Medicare as the referent with age and race/ethnicity as covariates. We calculated non-parametric mediation metrics and bias-corrected-accelerated 95% confidence intervals from 10,000 bootstraps (p <0.05).

Results:

The cohort included 3,604 hospitalized stroke patients: 52.16% men, median (IQR) age 68 (57 - 77) years. Median (IQR) LOS was 3 (1 - 7) days with a median (IQR) total admission cost of $67,426.10 ($37,610 - $132,077). Per patient estimated hospital costs compared to Medicare were lower for Medicaid (-$7,538.14; 95%CI [-$18,303, $3,226.72]; p 0.17) and self-pay (-$22,026.75; 95%CI [-$37,628.14, -$6,425.36]; p <0.01), but higher for commercial plans ($11,139.60; 95%CI [$2,878.89, $19,400.30]; p <0.01). LOS mediated 100% of the effect of Medicaid on hospital costs (p < 0.01), but did not mediate costs for self-pay (p = 0.95) and commercial insurance (p = 0.32).

Conclusions:

In this single center analysis, length of stay fully mediated the effect of Medicaid on acute stroke hospital costs. Hospitals could possibly reduce stroke costs by improving Medicaid authorization and processing times. Further research is warranted to understand higher costs with commercial carriers compared to Medicare.