gms | German Medical Science

Introduction: Between 1% and 2% of the German population suffer from atrial fibrillation (AF) with a high number of patients remaining undiscovered because of the frequently asymptomatic nature of AF [1]. The prevalence is estimated to double in the next 50 years as the population ages. The presence of AF independently increases the risk of stroke and thromboembolic events by approximately 5-fold across all age groups [2]. These fatal or disabling complications lead to impaired quality of life and higher medical care costs in patients with AF. Thus, the aim of the study was to assess stroke related health care utilisation and respective costs in patients with AF.

Methods: The analysis was based on data of an 18.75% random sample of all members of a large local German statutory health insurance provider, "AOK" in the federal state of Hesse. The study population consisted of all patients 18 years and older with AF diagnosis in the year 2009 (n=8145). AF patients suffering from valvular diseases were excluded. The case population comprised all patients with a following hospital discharge diagnosis of stroke in 2009 (n=243). To assess stroke related health care utilisation and respective costs in the year of stroke and the three following years, a matched cohort design was applied. Controls were selected from the study population but without any stroke after AF diagnosis until end of 2012 using propensity score matching (1:2) with a nearest neighbour approach (n=486). Costs attributed to stroke were estimated as the difference in costs of AF patients with and without stroke. To improve precision of cost estimation cases and respective controls from two years (2008, 2009) were pooled in the cost analysis (n=527). Costs will be presented as 10% winsorised costs.

Results: Differences in health care utilisation between patients with AF and stroke and matched AF patients without stroke were especially seen in inpatient care. The mean number of hospital stays in the year of stroke was 3.2 in cases and 1.2 in controls, while differences were less pronounced in the following years (2010: 1.9 vs. 0.9, 2011: 1.1 vs. 0.8, 2012: 1.5 vs. 0.9, in cases and controls, respectively). Moreover, in the year of stroke cases were more than three times longer in hospital than controls (47 days vs. 14 days), whereas differences where less pronounced in the following years (2010: 22 vs. 10, 2011: 13 vs. 9, 2012: 15 vs. 9, for case and controls, respectively). Utilisation of nursing care was about 61% among the cases and 38% among the controls in the year of stroke with no significant differences over time for both cohorts.

In outpatient health care utilisation differences were seen with regard to therapeutic remedies, therapeutic aids and number of different prescriptions. In the year of stroke and in the following year 42% and 44% of cases received therapeutic remedies (controls: 24% in both years) and 74% and 73% utilised therapeutic aids (controls: 58% and 54%), respectively. Cases received on average in the year of stroke 13 different prescriptions, decreasing to 12, 11.5 and 11 in the three following years, whereas controls without stroke had a mean number of 10 different prescriptions on average in all years. In the year of stroke the proportion with antidepressants (25% vs. 16%), drugs for the cardiovascular system (94% vs 89%), vitamin-K-antagonists (47% vs. 37%) and platelet aggregation inhibitors (35% vs. 24%) were higher in cases than compared to controls, respectively.

The direct incremental overall health care costs attributed to stroke (including both costs reimbursed by health insurance and by nursing care insurance) were €14,998, €6616, €5291, €4043 per person year in the year of stroke and the three following years, respectively. Of that amount €1236 were spent for nursing care in the year of stroke and €3436, €3245, €2806 in the three following years, respectively. The highest costs attributed to stroke were observed in inpatient care with €11,930, €1467, €762 and €359 per person year in the year of stroke and the three following years, respectively.

Conclusion: The medical and economic burden of stroke in patients with AF are high. Besides the hospital stay for and the rehabilitation after the acute event, the utilisation of therapeutic remedies, therapeutic aids and of nursing care is clearly increased compared to AF patients without stroke. In patients with AF stroke may cause twice as high direct medical and nursing care costs in the first four years compared to AF patients without stroke. The analyses strengthen the need for guideline recommended therapy to prevent thromboembolic events.