Abstract 171: A Population-based Comparison of Total Costs: The Economic Burden of Atrial Fibrillation-Associated Stroke

Abstract

Introduction Accurate health economics data is essential for service planning and clinical guideline development. Stroke associated with atrial fibrillation (AF) is characteristically more severe and disabling than non-AF stroke. Few data exist on the economic impact of AF-stroke compared to non-AF-stroke in population-based samples. Methods The North Dublin Population Stroke Study is a population-based prospective study of incident stroke in Dublin, Ireland. Both direct (healthcare-related) and indirect costs were calculated over a 2-year post-stroke period for individual patients, using data for survival and disability, discharge destination, outpatient and family practitioner visits, community supports, and ongoing treatment. Acute inpatient care was costed using the casemix approach, measuring the resource use per hospitalization (index stroke plus post-stroke admissions). Length of stay was used to cost rehabilitation and nursing home admissions. Indirect costs of illness were calculated using in-hospital length of stay to determine loss of productivity. Total costs were compared for AF-stroke and non-AF stroke, calculated at 2007 prices and converted to US$ (2007 rates). Results Of 568 ischemic and hemorrhagic incident stroke patients, 31% (177) had AF-associated stroke. 2-year fatality was higher in AF-stroke patients (50.3% versus 35.1%, p=0.001). Total 2-year median cost was $34,982 for AF-stroke patients (25-75% IQR $14,213-$88,283) compared to median cost of $17,736 (25-75% IQR $10,803-45,838] in non-AF stroke patients (p<0.001). Inpatient care contributed 47% of total costs: cost per AF-stroke patient (index stroke) was median $14,257 (IQR $10,286-$40,989) compared to median $10,624 per non-AF stroke patient (IQR $9,482-26,241, p=0.002). Long-term institutional care was the second highest cost contributor with greater proportions of living AF stroke patients in nursing homes (2-year mean cost $23,997(±standard deviation [SD] $68,585) versus non-AF-stroke (mean $12,988 (±SD $51,561, p=0.002 for cost comparison). Community supports (2-year mean $5,487 ±SD $13,283 versus $3,808 ±SD $11,844 (p=0.1), post-stroke specialized equipment (mean $2,729±$8,927 vs $2,122±$8,267, p=0.2), and repeat hospital admission (mean $3,580±$12,124 vs $1,951±$8,446, p=0.4) costs were non-significantly higher in patients with AF-stroke. In linear regression models, AF was an independent predictor of total costs after adjusting for age (p=0.02, β 0.1) but the association was lost on adding stroke severity (72-hour Rankin score) (p=0.1, β 0.07). Conclusion In this population study, AF-associated stroke was associated with substantially higher costs for hospitalisation and community support after discharge. With ageing populations and increasing stroke burden, AF-stroke is likely to account for an increasing proportion of overall cost to health services.